The Unnoticed Giant at AfroTech

Hidden within an ordinary brick building on Preston Street in Houston, a group of individuals convened, their eyes set on what lies ahead. Among the gathering of 60 were Byron Spruell, the NBA's president of league operations, and TikTok culinary star Keith Lee, who positioned himself towards the rear in a Yankees letterman jacket, attempting to dodge any unwanted focus. "I have to make sure he knows our city's cuisine isn't terrible," mentioned an executive from LinkedIn based in San Francisco.

Like myself, they had traveled to Houston to engage in AfroTech, a leading technology conference that has become a significant event for numerous Black tech professionals. This evening, as a feature of Microsoft’s Creator Unplugged event—an addition to the array of activities complementing the four-day conference—Spruell, Lee, and other attendees enjoyed champagne while networking in a selectively assembled group. The atmosphere was impeccably set. However, this edition of AfroTech was held under the looming presence of Donald Trump's recent electoral win from the previous week, casting a shadow that left many attendees preoccupied with other, possibly daunting concerns.

Within half an hour of arriving at the location, which for the time being was called the House of Black Techxcellence, I encountered an ex-Twitter staff member. Our discussion swiftly turned towards the daunting prospect of a second Trump era. Our concern wasn't just about Trump's aggressive electoral strategy, based on complaints and overt racism, but also about the group of allies he surrounded himself with – including tech figures such as Elon Musk – and the potential consequences of their partnership.

"The decision to acquire Twitter was a masterstroke, according to the ex-staffer, who believed that his strategy of leveraging the social media site to sway the election, along with other methods, was akin to the advanced malevolence depicted in films."

Except for it being completely true, I concurred.

"He emphasized the importance of valuing the dream," he expressed. "Our society requires superior role models."

AfroTech, on the surface, positions itself as a platform for creating heroes. It's an initiative by Blavity, a digital media platform aimed at millennials, and was launched in 2016 as a 600-person meet-up in San Francisco. This event was designed for Black professionals in the tech industry who were concerned about the ongoing issue of underrepresentation. The event's fundamental idea was to create something from the community, for the community. Over the years, it has grown into a hub attracting a diverse group of visionaries who recognize the strength in unity. Nowadays, AfroTech serves multiple purposes. It features a job fair and around thirty-six panel discussions spread over four days, but perhaps its most significant aspect is the networking opportunities it provides. It's akin to a grand reunion, attracting not only startup founders, engineers, investors, and programmers but also anyone looking for a sense of community and connection.

Previous AfroTech events primarily aimed at promoting diversity, but the latest edition emphasized the importance of adaptability.

Following the outcome of the US election, where a Black female candidate was defeated by someone with a criminal record, my interest was particularly piqued. AfroTech has become a recognized entity within the Black technological community, with around 37,500 individuals participating this year. However, the question remains: how effectively is it equipping participants to navigate the challenges posed by a Trump administration that seems indifferent to Black technological advancements?

While attending several presentations, bearing names such as “Mastering the Pitch” and “Thriving in the Innovation Economy,” I found myself reflecting on a conversation I had with a previous Twitter staff member. Their words echoed in my mind: we need better heroes. This thought started to shape itself into both a question and a challenge in my mind. It led me to ponder whether AfroTech was fully leveraging its potential to nurture future leaders.

This year's AfroTech theme focused on Artificial Intelligence (AI) with the motto "Designing the Future," a decision that appeared to be very insightful. Preparing for what lies ahead involves grasping the current landscape. This entails taking proactive steps to make certain that the upcoming technological upheaval doesn't evolve into an advanced route to bias, according to Charlotte Burrows, the head of the US Equal Employment Opportunity Commission. During her presentation, she referred to AI as “the new civil rights battleground.”

According to Morgan DeBaun, CEO of Blavity, while previous AfroTech events focused on representation, the current emphasis is on adaptation. DeBaun emphasizes that mere observation is insufficient in today's context. To stay abreast of advancements in AI, it's crucial for Black individuals to engage actively in both its development and application. The importance of this involvement is underscored by existing issues, such as biased algorithms that have unjustly incarcerated individuals and the widespread misuse of digital blackface, leading to distortion and abuse. There's a concern that playful engagement with diversity could turn into sophisticated forms of deception, with numerous instances of deepfake fraud already infiltrating our digital interactions and harming the ways we connect online.

Despite its successes, AfroTech has faced some criticism. "Were I in charge of AfroTech, I'd completely overhaul the scheduled events to focus on discussions about navigating the upcoming challenges," Erika Stallings, a lawyer and author, shared on X. "Things are about to become significantly tougher for Black professionals."

"Chile, they're indulging in Dusse and hosting brunch gatherings," responded @ChampagneNoona. I could understand their point of view. Occasionally, I questioned the effectiveness of having celebrities headline the main events. I've enjoyed TI's tunes since his 2004 release, Urban Legend, which remains a hit, but I'm puzzled about what insights he could offer on AI.

The undeniable influence of technology on employment dynamics and its potential to drastically alter the job landscape was a topic beyond contention. During a session, a Deloitte consultant laid out the future in no uncertain terms. "It might be uncomfortable to acknowledge, but the reality is that many positions will be eliminated," she declared. Notably, her statement was made in a session that was exclusive to attendees with executive access, for which the entry fee commenced at $750.

"Regardless of whether you're switching positions or not, the nature of your job is evolving," remarked a different advisor during a separate discussion on artificial intelligence.

In 2023, the technology sector experienced a significant wave of job cuts. These were attributed to businesses restructuring to prioritize artificial intelligence and adjustments made due to excess recruitment during the COVID-19 crisis. Amazon set the precedent in January by dismissing 18,000 of its staff, with numerous leading firms doing the same thereafter. By the year's close, approximately 200,000 tech employees in the United States found themselves unemployed, with the impact being particularly severe on women, individuals with disabilities, and Black professionals within the tech community. The trend of employment reductions persisted into 2024, with over 140,000 positions eliminated by November.

Under a presidency led by Trump, the prospects for 2025 seem decidedly grim. He has pledged to support an economic model that rejects progressive ideologies, appointing provocateurs to his team—like Brendan Carr, his pick for the head of the Federal Communications Commission—who are committed to dismantling Diversity, Equity, and Inclusion (DEI) initiatives. His governance is expected to heavily draw from Project 2025, a 900-page conservative policy blueprint that targets organizations implementing “racial classifications and quotas” and aims to revoke an executive order mandating equal opportunity provisions by federal contractors. Moreover, even in the absence of a Trump administration's direct influence, major tech companies had already begun to roll back their DEI efforts.

"Given the election and the lack of focus on Diversity, Equity, and Inclusion, it's crucial to stay more alert."

Maryland Governor Wes Moore, who was the sole state official present, strongly dismissed the notion that DEI initiatives harm productivity as ludicrous. "Just consider the data," he urged. For instance, a study by McKinsey & Company from 2020 indicates that efforts towards diversity, equity, and inclusion actually benefit businesses. The frequent use of "Authenticity" as the top buzzword throughout the week's events — a term that was echoed in every session I was part of — seemed fitting yet strangely peculiar, especially considering that the pursuit of authenticity is expected to be a major focus in the coming four years.

"Onstage, artist Will.i.am proclaimed, 'What's coming next is beyond anything we've witnessed.' And indeed, that was the case."

At the job fair, attendees found themselves in winding queues that often extended further than the ones for the evening festivities, with all the major companies present: Netflix, Axon, Meta, Google, Oracle. Observing the grand spectacle of the trade show area, adorned with big, arching banners pointing every which way, my mind wandered back to my initial day in Houston. That was when a recruiter from Microsoft humorously cautioned me to keep his profession a secret, fearing an inundation of CVs and inquiries regarding vacancies at the corporation.

Despite this, understanding the conference's efficiency in equipping its future leaders posed a challenge. The issue wasn't with the agenda, but with the tangible outcomes. While artificial intelligence garnered the most attention, the priorities of many attendees lay in immediate concerns, particularly securing employment.

Candace Madison, employed in legal technology at Relativity, a software firm specializing in data management based in Chicago, shared her perpetual concerns about the stability of her job. Attending AfroTech for the first time, she expressed that her job insecurity wasn't necessarily heightened by the recent election, but the diminished focus on diversity, equity, and inclusion (DEI) meant that she had to remain particularly vigilant. Despite these challenges, she remained hopeful, believing in the power of networking as a key strategy to navigate the uncertainties of the current times, even though she acknowledged having made only a few connections within her industry so far.

Inside the lift at Le Meridien located in the heart of Houston, a doctoral candidate in data science actively seeking employment shared her perspective differently. "This year, I've attended my eighth conference," she mentioned. "I'm putting a lot of effort into networking, yet it seems to yield little results."

The event was hailed as a triumph on Instagram. A poignant story post featured a product engineer from a top 50 Fortune company sharing how attending the conference marked a significant milestone in his career, leading from an internship he secured at the 2017 expo to his present position. Meanwhile, a senior marketing executive depicted this year's conference as a deeply soothing and healing experience, likening it to "a balm in Gilead" in another post.

Unsurprisingly, all attendees at AfroTech were focused on what lies ahead. However, the uncertainty of future developments and their ability to influence these outcomes was a common concern. It appeared that each individual was eager to claim their share of a hopeful future, yet the path to achieving this stability remained unclear.

Recommended for You …

Delivered daily: A selection of our top stories, curated personally for your inbox.

Response to the vote: Male-dominated spheres triumphed

Major Headline: California Continues to Propel Global Progress

Trump's unsuccessful effort to topple Venezuela's leader

Occasion: Be part of The Major Interview happening on December 3rd in San Francisco

Additional Content from WIRED

Evaluations and Instructions

© 2024 Condé Nast. All rights reserved. WIRED might receive a share of revenue from items bought via our website, as a result of our affiliate agreements with retail partners. Content from this site cannot be copied, shared, broadcast, stored, or utilized in any form without explicit approval from Condé Nast. Advertisement Options

Choose a global website

The New York Times Accuses OpenAI of Deleting Evidence Pertinent to Lawsuit

Legal battles are hardly ever amicable, and the ongoing copyright dispute involving The New York Times, along with OpenAI and Microsoft, is becoming particularly heated. Recently, The Times claimed that OpenAI’s technical staff accidentally deleted data which had taken the newspaper’s team over 150 hours to compile for use as possible evidence.

OpenAI managed to retrieve a significant amount of the data, however, the legal representatives from the Times argue that the initial file names and organizational framework are still unaccounted for. As stated in a legal document submitted to the court on Wednesday by Jennifer B. Maisel, an attorney for the newspaper, this issue renders the data ineffective for identifying the specific locations within OpenAI’s AI models where the copied articles from the news plaintiffs might have been integrated.

"OpenAI's spokesperson, Jason Deutrom, informed WIRED that the company objects to the descriptions provided and plans to submit their reply shortly," stated WIRED. The New York Times has chosen not to issue a statement.

Last year, The Times initiated a legal battle against OpenAI and Microsoft, accusing them of unlawfully utilizing its published articles to develop AI technologies such as ChatGPT. This lawsuit is among several current disputes involving AI firms and content creators, with the Daily News also launching a comparable case represented by the same legal team.

The current stage of The Times' litigation involves the exchange of documents and data that might be used in court, a phase known as discovery. During this part of the legal proceedings, OpenAI was mandated to allow The Times a look into its training data for AI models—a first, as OpenAI has historically kept this information under wraps. To comply, OpenAI set up a controlled environment termed a "sandbox," consisting of two "virtual machines" for The Times' legal team to examine. In her statement, Maisel mentioned that OpenAI's technicians had deleted information that The Times' personnel had organized on one of these platforms.

Based on the documents submitted by Maisel, OpenAI confirmed the removal of data and endeavored to rectify the problem soon after being notified earlier in the month. However, the legal team for the newspaper found the supposedly retrieved data to be in a chaotic state, necessitating them to begin their project anew, consuming considerable human effort and computational resources, as highlighted by other attorneys for the Times in a letter presented to the court on the same date as Maisel's statement.

The attorneys mentioned that there was "no evidence to suggest" that the removal was "deliberate." In the emails presented as evidence together with Maisel's correspondence, OpenAI's lawyer Tom Gorman described the data deletion as a "malfunction."

This isn't the first clash of its nature within the ongoing legal battle. Throughout the last year, there has been an ongoing tussle between the newspaper and the technology firms over who should bear the responsibility for examining the training datasets. In the latest communication from the newspaper, its legal representatives reiterated that OpenAI is more aptly equipped to handle this task. "The process has been far from smooth," wrote Steven Lieberman, another attorney for the newspaper, in a recent court document. He pointed out that "significant and recurrent technical problems have rendered it unfeasible to conduct a thorough and efficient review of OpenAI's training data to determine the extent of OpenAI's copyright infringement."

Recently, The New York Times exerted pressure on OpenAI and Microsoft, demanding they release Slack messages, text messages, and social media dialogues involving several prominent figures from OpenAI, such as ex-employee Ilya Sutskever and existing executive Brad Lightstone. In the previous week, the newspaper submitted a further request to the court, seeking to obligate Microsoft and OpenAI to disclose more documents. Included in one of the exhibits were emails revealing that Mira Murati, the former Chief Technology Officer of OpenAI, had declined to grant access to her personal mobile phone.

In a recent development, Microsoft has approached The New York Times, seeking access to documents concerning the newspaper's application of generative artificial intelligence. This request was highlighted in a legal document where Microsoft pointed out the involvement of prominent technology writer Kevin Roose (incorrectly named "Kevin Rouse" in the legal documents). The tech giant contends that understanding the newspaper's utilization of AI technologies could significantly bolster its defense by demonstrating the beneficial effects these tools have had on the publication. When approached for a comment, Roose opted not to respond.

As this lawsuit, along with similar ones, progresses through the judicial system, OpenAI is actively seeking agreements for content licensing with various publishers such as The Atlantic, Axel Springer, Vox Media, and Condé Nast, the company that owns WIRED. The outcome of these legal battles is currently uncertain, with opinions divided among those in the media and legal sectors. Regardless of the results, these cases are expected to establish significant benchmarks for the operational framework of the AI sector in the U.S.

Recommended for You…

Direct to your email: Enhance your lifestyle with products tested and approved by WIRED

An aspiration to mine bitcoin devolved into a dreadful experience.

The In-depth Conversation: Marissa Mayer—I'm a software enthusiast

I created a positive OnlyFans account in an effort to financially get by.

Occasion: Be part of The Major Interview happening on December 3rd in San Francisco.

Additional Content from WIRED

Evaluations and Instructions

© 2024 Condé Nast. All rights reserved. Purchases made through our website may result in a commission for WIRED, as part of our Affiliate Agreements with retail partners. Content on this website is protected by copyright and cannot be copied, shared, broadcast, stored, or used in any other way without explicit consent from Condé Nast. Ad Choices

Choose a global website

Maximizing the Benefits of ChatGPT, Gemini, and Additional AI Resources: Insights from an AI Specialist

What distinguishes ChatGPT, Claude, Gemini, Copilot, and other leading AI solutions from each other? Which stands out as the superior option? To what extent should worries about AI technologies accessing your private information be taken seriously? Is the level of inaccuracies and biases in AI services truly as problematic as reported?

Earlier in November, we reached out to our WIRED subscribers, inviting them to submit their inquiries concerning generative AI. Below are a selection of the outstanding questions we collected. On the 20th of November, we organized a live question-and-answer session featuring Reece Rogers from WIRED, known for his work on the “AI Unlocked” newsletter, the AI ethics and advice monthly column “The Prompt,” along with numerous news pieces and instructional articles on AI's influence on our daily lives.

If you didn't catch the live broadcast, there's no cause for concern—you can view the entire session on this platform. Reece teamed up with WIRED's worldwide editorial director, Katie Drummond, who delivered the introductory comments, along with WIRED's audience development manager, Laura Fillbach, who took on the role of event moderator.

During the half-hour conversation, Rogers explored the commonalities and distinctions among the leading generative AI technologies, revealing his preference for the one he believes is worth the additional cost. He discussed the advantages and disadvantages of opting for AI-driven search tools over conventional search engines and warned the audience about the risks of seeking specific health advice from chatbots. Rogers responded to several inquiries from readers regarding the reliability of watermarks as indicators of AI-generated content. Additionally, he shared insights from his own experiences regarding the use of AI in educational environments and the obstacles educators encounter given the widespread availability of AI resources to students.

WIRED DISCOUNTS

Dyson Loyalty Program – Save 20% with Dyson Discount Code

Receive up to $750 in Trade-In Value for the Galaxy S24 Ultra

Newegg Discount Code – Save 50% on Certain Items

Special Peacock Offer: $1.99 Monthly for Students for One Year

For a Short Period, Purchase the Latest DJI Mini 4 Pro Starting at $759

Students Can Save Up To $20 Monthly With Discount

Additional Content from WIRED

Critiques and Manuals

© 2024 Condé Nast. All rights reserved. Purchases made through our site may result in WIRED receiving a share of the sales, this is due to our Affiliate Agreements with retail partners. Content from this site cannot be copied, shared, transmitted, stored, or used in any form without explicit written consent from Condé Nast. Ad Choices

Choose a global website

This section explores the transformative power of AI, showcasing top innovations from davinci-ai.de, ai-allcreator.com, and bot.ai-carsale.com, and its impact across multiple sectors through Artificial Intelligence Machine Learning, Deep Learning, Neural Networks, Natural Language Processing, Robotics, and more. It highlights how AI is redefining industry interactions and pushing the boundaries of human-machine collaboration with advancements in Cognitive Computing, Data Science, and Intelligent Systems. With a focus on Computer Vision, AI Algorithms, and Augmented Intelligence, it discusses the role of AI in enabling Predictive Analytics, Big Data analysis, Autonomous Systems, and Smart Technology in finance, healthcare, and retail, emphasizing the shift towards technology autonomy and the synergy between AI and Robotics for innovative solutions. This exploration into AI innovations presents a future of limitless potential for human and machine collaboration, driven by continuous breakthroughs in technology.

In an era where the digital revolution is reshaping the contours of our world, Artificial Intelligence (AI) stands at the vanguard, heralding an age of unparalleled innovation and transformation. From the intricate algorithms powering davinci-ai.de to the cutting-edge robotics showcased on ai-allcreator.com, and the advanced autonomous vehicles navigating bot.ai-carsale.com, AI is not just a buzzword but a dynamic force sculpting the future of human existence. This comprehensive exploration, titled "Unlocking the Future: How Top AI Innovations Are Shaping Our World," delves deep into the heart of AI technologies—spanning Artificial Intelligence, Machine Learning, Deep Learning, Natural Language Processing, Robotics, and beyond. We unpack how these disciplines, underpinned by Neural Networks, Cognitive Computing, Data Science, and Intelligent Systems, are revolutionizing industries, from Computer Vision enhancing medical diagnosis to AI Algorithms optimizing financial forecasting. Through the lens of top innovations, we examine the impact of AI in driving Predictive Analytics, leveraging Big Data, and propelling Autonomous Systems towards new horizons. Join us as we navigate the realms of Smart Technology, Pattern Recognition, Speech Recognition, and Augmented Intelligence, unraveling how AI is not merely automating the mundane but redefining the paradigms of human-machine interaction and setting the stage for a future where the boundaries between the digital and the physical blur into obsolescence.

• "Unlocking the Future: How Top AI Innovations Are Shaping Our World"

"Unlocking the Future: How Top AI Innovations Are Shaping Our World"

In the dynamic realm of technology, Artificial Intelligence (AI) stands as a towering beacon of innovation, driving change and molding our future in unprecedented ways. The top AI innovations are not just transforming industries but are redefining human-machine interactions, fostering a new era of efficiency and possibilities. From the realms of davinci-ai.de, a pioneer in AI-driven creative processes, to the groundbreaking platforms like ai-allcreator.com that democratize AI capabilities, the landscape is rich with advancements. Similarly, bot.ai-carsale.com is revolutionizing the automotive industry by integrating AI in enhancing the car buying and selling experience, showcasing the practical applications of AI in everyday life.

The essence of AI, which encompasses Artificial Intelligence Machine Learning, 3 Deep Learning, Neural Networks, and Natural Language Processing, is to replicate and augment human intelligence. This is achieved through sophisticated algorithms and AI systems that learn from data, adapt through experience, and make informed decisions. Robotics and Automation, as subfields, extend AI’s reach into physical tasks, while Cognitive Computing and Data Science reflect its analytical prowess, making sense of vast Big Data landscapes.

AI innovations have paved the way for Intelligent Systems that can perform complex tasks, ranging from Computer Vision for autonomous vehicles to Neural Networks that underpin AI Algorithms, enhancing Augmented Intelligence. These technologies are not only about automation but are geared towards making systems more adaptive, predictive, and responsive to human needs.

Predictive Analytics, a key application of AI, leverages historical data and AI algorithms to forecast future trends and behaviors. This has profound implications for industries such as finance, healthcare, and retail, where making informed decisions can significantly impact outcomes. Big Data, the fuel for AI, when combined with AI’s Pattern Recognition and Speech Recognition capabilities, enables the creation of Autonomous Systems and Smart Technology that understand and interact with the world in ways that were once the domain of science fiction.

The integration of AI into autonomous systems highlights the shift towards more independent, self-learning technology that can navigate, understand, and interact with its environment without human intervention. This leap towards autonomy represents a significant milestone in AI’s evolution, heralding a future where machines can assist or even take over tasks in areas like transportation, safety, and personal assistance.

Moreover, the advancements in AI are not just limited to the digital realm. Robotics, a field that synergizes with AI, is creating robots capable of performing tasks that require dexterity, decision-making, and even emotional intelligence, further blurring the lines between humans and machines. Similarly, in healthcare, AI-driven diagnostics, personalized medicine, and robotic surgery are becoming realities, offering hope for more precise and effective treatments.

As we stand on the brink of a technological revolution spearheaded by AI, it is clear that these innovations are not just shaping our world but are also unlocking a future where the potential of human and machine collaboration is boundless. The journey of AI from a concept to an essential facet of our daily lives underscores the transformative power of this technology. With each breakthrough, AI continues to expand the horizons of what is possible, promising a smarter, more efficient, and more intuitive world.

In the grand tapestry of technological evolution, Artificial Intelligence stands as a monumental leap forward, weaving together the threads of Machine Learning, Natural Language Processing, Robotics, and more into a fabric that is fundamentally transforming the fabric of our society. From the predictive analytics that forecast our financial futures with uncanny accuracy to the autonomous systems piloting self-driving cars, AI innovations are not just shaping our world; they are unlocking a future brimming with untapped potentials. Websites like davinci-ai.de, ai-allcreator.com, and bot.ai-carsale.com are at the forefront of this revolution, offering glimpses into how AI technologies continue to evolve and expand their influence across diverse sectors.

As we delve into the realms of Deep Learning, Neural Networks, and Cognitive Computing, it becomes clear that the applications of AI are as limitless as our imagination. Whether it's through enhancing big data analytics, refining pattern and speech recognition algorithms, or pushing the boundaries of smart technology and augmented intelligence, AI is setting the stage for a new era of innovation and efficiency. The integration of AI into everyday life, from virtual assistants to sophisticated medical diagnostics, highlights not just the adaptability of these intelligent systems but also their potential to significantly improve human life.

Yet, as we stand on the cusp of this AI-driven era, it's crucial to navigate the challenges that accompany such profound changes, including ethical considerations, privacy concerns, and the need for robust security measures. The journey ahead with AI promises to be as complex as it is exciting, requiring a collaborative effort to ensure that the benefits of artificial intelligence can be realized fully and fairly for everyone.

In conclusion, the landscape of Artificial Intelligence is vast and varied, filled with incredible innovations that are already reshaping the contours of our world. As we look to the future, the role of AI in driving progress and solving some of humanity's most pressing challenges becomes increasingly clear. By harnessing the power of AI algorithms and intelligent systems across fields such as computer vision, robotics, and natural language processing, we are not just witnessing the evolution of technology but participating in the dawn of a new intellectual renaissance. The future unlocked by top AI innovations promises a world where technology and human ingenuity combine to create a smarter, safer, and more efficient society for generations to come.

Brad Porter, the Genius Behind Amazon's Robotic Workforce, Envisions a Future with AI Assistants for All

With notable expertise in deploying robots for productivity, Brad Porter spearheaded an initiative from 2017 to 2020 at Amazon focusing on the creation of an advanced, AI-driven fleet of warehouse robots.

These devices, encompassing vision-equipped robotic arms that can pick up objects from conveyor belts, along with autonomous robots that operate alongside humans, have enabled Amazon to reconfigure its distribution hubs for greater automation. This advancement has hastened the pace of order processing and delivery, and naturally, enhanced profitability.

Porter is now at the helm of Cobot, an enterprise focused on enabling other businesses to expand their use of robotics. The inaugural offering from Cobot is Proxie, a warehouse automation machine characterized by its two arms and four wheels, bearing a resemblance to a movable coat rack. This robot, featuring a digital display for a face and sensors mounted on a tall mast, is primarily crafted to handle the routine yet widespread task of transporting carts loaded with goods through various settings such as offices, airports, hospitals, and manufacturing plants.

"Porter envisions a future where reliable cobots are found everywhere, collaborating with humans across all industries."

Attributed to Joint Robotic Eff

Currently, Maersk, a shipping company, and the Mayo Clinic are trialing around 30 Proxie robots. These robots have amassed over 5,000 hours of operation, transporting 16,000 carts and covering a distance of more than 1,000 kilometers. Additionally, various other organizations such as Moderna, Owens & Minor, and Tampa General Hospital are investigating potential applications for Proxies.

Differing from its robotic counterparts, Proxie features a replaceable battery, eliminating the need for charging breaks. When asked about Proxie's purchase or leasing price, Cobot chose not to disclose any figures, though it's common for mobile robots to be priced in the range of several tens of thousands of dollars each.

The robots collaborate with humans, alternating in transporting carts and maneuvering through crowded areas without colliding with anyone. According to Porter, the goal is for these robots to advance as artificial intelligence improves, enabling them to perform more complex tasks and enhance their communication abilities.

Porter mentions that Cobot offers a Proxie variant capable of obeying spoken instructions by employing an extensive language model to interpret speech. Upon receiving an order such as "Move to dock 3 and pick up the cart at the entrance," the robot will act as instructed. Additionally, the firm is monitoring progress in algorithmic research to enhance the robot's ability to perform complex manipulative tasks.

At a glance, Proxie may appear deceptively straightforward, especially as numerous corporations are scrambling to create humanoid robots. However, Porter notes that although Amazon is experimenting with a humanoid robot developed by Agility Robotics, the technology is currently too nascent and costly for broad implementation. Humanoid robots available today can range in price from several tens of thousands to multiple hundreds of thousands of dollars. Moreover, there is a significant variation in their autonomous functions and dependability, which further escalates their deployment expenses.

"Porter mentions that at Amazon, they extensively considered humanoid robots. He acknowledges there are genuine issues that could benefit from capabilities closer to those of humans. However, he points out that fully committing to humanoid robotics is extremely challenging because the artificial intelligence required is still not sufficiently advanced."

Rather, Proxie has the capability to take over numerous tedious tasks typically avoided by humans. Erez Agmoni, a general partner at Interwoven Ventures and a key figure in introducing the Cobot pilot to Maersk, mentioned that the results have been quite encouraging and show potential for broader application.

"The primary factor is their use of cooperative robots that assist the staff without the need for significant changes to the existing warehouse setup or machinery," he explains. "The team despised moving the cumbersome carts and appreciated the robots taking over this task."

Fady Saad, the brains behind Cybernetix, a venture capital company rooted in Boston with a focus on robotics, believes Cobot is targeting a significant new sector of work that entails the transportation of items via carts, a task that recent developments in robotics can address. He further emphasizes the necessity for Proxie to develop into a more proficient entity.

"Saad mentions that Porter is working on developing a platform which has the potential to transform into a humanoid in the future. He believes this strategy is appropriate."

Porter is not alone in his quest for simpler robotic solutions beyond human-like robots. Rodney Brooks, a trailblazer in the field and one of the founders of iRobot, currently serves as the chief technology officer at Robust.AI. This company specializes in creating cooperative mobile robots designed to assist human workers in selecting items within factories and warehouses.

"Factories and warehouses critically require the transportation of items, yet the notion that humanoid robots will soon handle this task is absurd," Brooks comments. "There’s a significant reason why wheels were invented."

Which types of simple chores would you appreciate a robot assisting you with? Does it matter to you whether the robot looks human-like or not? Please email me your thoughts at hello@wired.com.

Suggested for You…

Directly to your email: A selection of our top stories, curated daily just for you.

Response to Election: Victory for Male-Dominated Spheres

Major Headline: California Continues to Propel Global Progress

Trump's unsuccessful effort to oust Venezuela's leader

Don't miss out on The Big Interview happening on December 3rd in San Francisco!

Additional Coverage from WIRED

Critiques and Manuals

© 2024 Condé Nast. All rights reserved. Purchases made through our website may result in WIRED receiving a share of the revenue, owing to our Affiliate Agreements with retail partners. Content from this website cannot be copied, shared, broadcasted, stored, or utilized in any manner without explicit consent from Condé Nast. Advertisement Choices

Choose a global website

Nvidia Announces Its Blackwell Chip Is Performing Well, No Issues Reported

Today, Nvidia, the leading chip manufacturer, announced its third-quarter financial results. The announcement had many eagerly listening and likely some groups closely watching, hoping to gauge the potential impact of the company's results on the broader artificial intelligence sector moving forward.

The future prospects of the company's latest artificial intelligence processor, dubbed Blackwell, garnered significant attention due to earlier production challenges that led to shipment postponements spanning several months. Furthermore, a report by The Information on Sunday highlighted that when assembled within Nvidia's specially designed server racks, the Blackwell chips experienced overheating issues, necessitating modifications to their design.

During today's earnings discussion, Nvidia's co-founder and CEO, Jensen Huang, informed analysts and investors that production of Blackwell is now operating at full capacity.

"Huang announced that this quarter, the delivery of Blackwells will surpass our earlier forecasts." Key clients of Nvidia, including Microsoft and OpenAI, have been among the first to acquire these new processors. According to company executives, the sale of Blackwells could potentially bring in revenue amounting to billions of dollars for the next quarter.

Earlier, experts in the sector indicated that the issues with the latest processors have been resolved. Dylan Patel, the leading analyst at SemiAnalysis, conveyed to Business Insider that the overheating problems with Blackwell chips "had persisted for several months but have mostly been rectified."

Patrick Moorhead, who established Moor Insights & Strategy and serves as its principal analyst, informed WIRED through his production connections that significant overheating problems hadn't come to their attention. However, he mentioned that experiencing some friction among the design, engineering, and manufacturing teams is pretty standard while launching a novel chip design.

Moorhead highlighted an enduring debate within the semiconductor industry, questioning whether the challenges lie in the design or production aspect. He noted that individuals are keen to identify any potential obstacles that might disrupt Nvidia's narrative of rapid expansion.

Nvidia achieved a quarterly revenue of $35.1 billion, surpassing the anticipated $33.2 billion. This represents a 94 percent growth in revenue year-over-year, which, although modest compared to its previous quarters that saw increases up to 265 percent, signifies that Nvidia remains a strong player in the market. These figures suggest that the artificial intelligence sector is still thriving, with enterprises investing heavily in sophisticated processors and related technology, though the rate of investment has decelerated somewhat.

Nvidia's financial expansion this quarter was significantly fueled by its data center earnings, achieving $30.8 billion, marking a 112 percent increase compared to the previous year. The firm's gross profit margin stood at 74.5 percent, showing little change from the same period last year. However, there is a prediction among analysts that Nvidia's profit margins may decline due to the transition towards manufacturing advanced Blackwell chips, which are more expensive to produce than the earlier versions.

Nvidia's financial performance is often regarded as a crucial indicator for the artificial intelligence sector. The company's sophisticated graphics processing units (GPUs) are the backbone of the intricate neural network computations that have fueled the surge in generative AI technologies. With major tech firms in Silicon Valley competing to develop novel chatbots and image-creation applications in recent years, Nvidia has seen its sales skyrocket, propelling it beyond Apple to become the world's top-valued publicly traded company. Following the introduction of ChatGPT in November 2022, Nvidia's share value has soared, almost increasing by a factor of ten.

Nearly all significant technology corporations involved in artificial intelligence, including those developing proprietary processing chips, depend extensively on Nvidia's graphics processing units (GPUs) for the training of their AI algorithms. For instance, Meta has disclosed that its most recent AI advancements are being powered by a network consisting of over 100,000 Nvidia H100s. On the other hand, smaller AI enterprises have faced a shortage of necessary AI computing resources as Nvidia has had challenges meeting the surge in demand.

Blackwell, the latest graphics processing unit from Nvidia, consists of two silicon parts, each matching the size of its former chip, Hopper. These parts are fused into a singular unit. This innovative structure is claimed to offer a performance that is quadruple the speed and boasts over twice the transistor count of its forerunner.

The introduction of Blackwell faced several hurdles. Initially expected to be released in the second quarter, the debut of the new processor was postponed due to manufacturing difficulties, leading to a delay of several months. Huang admitted to the issue, describing it as a "design flaw" that resulted in poor production yield. In an August interview with Reuters, Huang acknowledged that Nvidia's enduring partnership with Taiwan Semiconductor Manufacturing Company Limited was instrumental in resolving the problem.

Moorhead expressed to WIRED his continuing optimism about Nvidia and believes that the market for generative AI will keep expanding for the coming 12 to 18 months, even in light of some recent analyses indicating that advancements in AI might be reaching a peak.

Moorhead believes that shareholders would only revolt if they started worrying about the spending on assets or the earnings from the major tech giants such as Amazon, Google, Microsoft, and Meta, which have significant investments in AI cloud services. "However, I reckon they will continue to invest in Nvidia until we see any real issues arise." He also mentioned that Nvidia is experiencing expansion in the enterprise AI sector.

In today's financial briefing, Nvidia's CFO, Colette Kress, announced that Nvidia's AI solutions for businesses are operating at maximum capacity. This includes a platform that enables other companies to develop their custom AI assistants and agents. She mentioned Salesforce, SAP, and ServiceNow as some of the key clients utilizing these tools.

Huang reiterated later in the conversation, "Enterprise adoption of agentic AI is beginning to take off," he mentioned. "It's currently the newest trend."

Suggested for You…

Directly to your email: A selection of our most significant stories, carefully curated daily for you.

Response to election results: Victory for the manosphere

Main Headline: California Continues to Propel Global Progress

Trump's unsuccessful effort to depose Venezuela's leader

Happening: Don't miss out on The Big Interview happening on December 3rd in San Francisco

Additional Content from WIRED

Critiques and Manuals

© 2024 Condé Nast. All rights reserved. Purchases made through our website may result in WIRED receiving a share of the sales, as part of our affiliate agreements with retail partners. Content from this website is protected by copyright and cannot be duplicated, shared, broadcast, stored, or used in any other manner without explicit consent from Condé Nast. Advertisement Choices

Choose a global website

The US Department of Justice believes that simply forcing Google to sell its Chrome browser will not suffice to break up its dominant position in the online search market. They argue that to effectively diminish Google's control, it's necessary for the company to terminate its profitable deal with Apple, grant access to a significant amount of its exclusive data to rivals and marketing firms, and ensure a complete and swift sale of Chrome, which holds a majority share in the American browser landscape. Additionally, the government seeks to have a say in the selection of Chrome's new owner.

The suggestions are components of a comprehensive strategy that legal representatives of the government presented on Wednesday to US District Judge Amit Mehta in Washington, DC, within the framework of a federal antitrust lawsuit against Google initiated in 2020. By the upcoming August, it is anticipated that Mehta will determine which of the proposed measures Google must implement to diminish its dominant grip on the search industry.

However, the technology behemoth has the option to challenge the decision, potentially pushing the implementation of the court's directive far into the future. On Wednesday, Google's president, Kent Walker, described the government's suggestions as "astonishing," "severe," "a drastic measure of interference," and "excessively wide-ranging." In a blog post, he stated that the modifications requested "could disrupt various Google services—not just Search—that individuals rely on and cherish in their daily routines." He further claimed that such changes could compromise the privacy and security of Google's users.

Individuals with experience at Google or those who have closely collaborated with the organization are divided on the impact of suggested solutions in changing user habits or enhancing the competitiveness of the search engine landscape. Four ex-Google officials, previously in charge of departments like Chrome, Search, and Advertising, conveyed to WIRED their belief that competitors' innovations, rather than regulatory actions, are the most reliable means of challenging Google's supremacy in the internet search sector. A past head of the Chrome division, preferring to remain anonymous to safeguard business connections, commented, "You can't force an inferior product on consumers."

A previous leader in Chrome's engineering team has admitted that the search engine's performance could have been superior if it were not influenced by Google's various commercial interests. This individual claims that Google hindered the deployment of features designed to enhance user experience because they would negatively impact the firm's ad revenue, which is significantly driven by clicks on advertisements within search results. "Why hasn't autocomplete been improved? Why isn't the 'new tab' page more user-friendly? Why hasn't the browser history function been upgraded?" questions the former executive, who chose to remain anonymous. The reason, they suggest, is the underlying motive to encourage more searches by users. Google has yet to respond to these allegations.

Nevertheless, rivals poised to gain from even a slight decrease in Google's influence are hopeful regarding the anticipated solutions. "There are significant advantages to returning [Chrome] to the control of the community," states Guillermo Rauch, the CEO of Vercel, a firm that creates tools for websites, a lot of which rely on search traffic and ad revenue dominated by Google. "Managing that connection with the corporate giants is invariably going to be beneficial," Rauch mentions.

Gabriel Weinberg, the chief executive officer of competing search engine DuckDuckGo, expressed in a statement that the government's suggested solutions would liberate the search market from Google's unlawful hold, potentially sparking a wave of innovation, investment, and competitive activity.

The legal conflict between Google and the Department of Justice initiated during the initial term of President Trump in 2020. The national authorities, accompanied by several states, charged the technology behemoth with employing unfair strategies to control the search engine market, thus hindering the ability of US citizens to utilize alternative search services. Under President Biden’s leadership, the lawsuit progressed, and an additional complaint was lodged against Google. This new accusation targeted Google's alleged illegal control over the ad tech industry, which is crucial for countless websites and applications to earn money. The final arguments for this particular case are planned for Monday.

The outcomes of both situations are still undetermined, and it remains uncertain how vigorously the Justice Department will continue to pursue Google following Donald Trump's reinstatement in the Oval Office. While campaigning, Trump's remarks about the technology behemoth were inconsistent. In October, he voiced worries regarding its influence but indicated that implementing burdensome restrictions on the firm might undermine the United States' attempts to dominate the technological landscape against China.

Judge Mehta has allocated almost two weeks beginning in April to consider the arguments from both the government and Google regarding the suggested sanctions. By then, the stance of the newly instated Trump administration towards Google is expected to be clearer, and there's a chance that government lawyers might show reluctance in supporting the proposals unveiled on Wednesday.

In a Wednesday blog post, Walker emphasized potential consequences of the suggested plans that might worry Trump, such as a decrease in AI investment enthusiasm and the creation of a five-member Technical Committee tasked with overseeing Google's adherence to solutions. "This is only a glimpse of what's being proposed," Walker commented. "We wish this was just a figment of our imagination."

The administration is aiming to expand the options available to consumers regarding their selection of search engines. The objective is to terminate the agreement between Google and Apple, through which Apple earns billions in advertising revenue for setting Google as the primary search engine on its iPhones. Google's comparable agreements with various other firms would also be discontinued.

Google would also need to modify its approach to favoring its own services on Android, or alternatively, sell off Android or be compelled to do so. The suggestions demand that Google provide advertisers with a continuous flow of information to assist them in analyzing their buying activities.

The administration is pushing for Google to open access to its search index and user data analytics to boost competition. They believe this move will enable emerging competitors to catch up with the extensive knowledge Google has acquired from analyzing its vast user base over many years. Furthermore, in a document submitted on Wednesday, the Attorney General of Colorado suggested that Google should provide financial incentives to users who choose alternative search engines as their default option.

In addition to being required to sell off Chrome, Google would also face prohibitions against introducing a new browser or making investments in competitors within the fields of search, advertising technology, and artificial intelligence for a period ranging from five to ten years. According to the government, these measures are designed to promote innovation and revolutionize the markets for general search and search text advertisements over the coming ten years.

Rauch, who heads Vercel, argues that Google is improperly leveraging Chrome to guide users towards its AI chatbot, Gemini, and other proprietary services like Google Docs, by embedding various prompts and rewards within its search platform. "Google is capitalizing on every possible benefit by dominating this crucial software infrastructure," Rauch states.

Rauch suggests that transferring control of Chrome to an impartial entity, such as a nonprofit or an academic institution, would unlock the browser's search function, offering users a wide range of options. While Chrome currently permits the modification of its default search engine, Google's reminders often steer users back to its service. Rauch believes that if individuals had a more straightforward way to select their preferred search tool, many might opt for alternatives like Perplexity or ChatGPT, which is not as straightforward under the current system.

However, experts in finance and law have questioned the effectiveness of the government's proposed measures. Ex-Google leaders interviewed by WIRED share similar reservations. Rajen Sheth, previously in charge of segments of the Chrome division and currently leading a startup focused on creating digital classes, argues that consumers tend to stick with familiar options in what he views as an already competitive market. "Considering the current state of technology and the various factors at play, can any interventions truly have an impact? It seems doubtful," he remarks.

Gaining entry to Google's exclusive information and the chance to appeal to iPhone users could boost the likelihood that individuals will explore different search engines. However, Google possesses unparalleled computing resources, distinctive information from affiliated platforms like Maps, and over twenty-five years of familiarity among users. "Regardless of how evenly you distribute the opportunities, individuals will always choose the top option available," says the ex-leader of Chrome's business division.

Ex-Google leaders believe that the entity to eventually overtake the tech giant won't be a conventional search engine, but rather a platform similar to ChatGPT, offering a more engaging content delivery method. This emerging technology hasn't reached its full potential just yet, but it could well do so by the time the legal battle with Google reaches a conclusion. Consequently, Google's market position might undergo significant changes even before any regulatory actions start to take effect.

As of November 21, 2024, at 2:22 am EST, this article has been revised to include statements from Google and DuckDuckGo, along with further information from the legal documents submitted.

Suggestions for You…

Delivered to your email: Enhance your lifestyle with gadgets verified by WIRED

An aspiration to mine bitcoin evolved into a distressing ordeal

The In-Depth Conversation: Marissa Mayer—My Passion Lies in Software

I created a positive OnlyFans account in an effort to financially get by.

Occasion: Don't miss out on The Grand Q&A Session happening on December 3rd in San

Further Selections from WIRED

Evaluations and Instructions

© 2024 Condé Nast. All rights reserved. WIRED may receive a share of revenue from items sold via our website, thanks to our arrangement with affiliate retail partners. Content from this site is not to be copied, shared, distributed, or used in any form without explicit consent from Condé Nast. Ad Choices

Choose a global website

The AI Journalist Who Replaced Me Has Been Dismissed

James and Rose, the peculiar AI characters previously appointed as news anchors at The Garden Island, a local newspaper in Hawaii, have been let go.

Keeping staff on board is often a challenge for community newspapers, and The Garden Island newspaper in Kauai, Hawaii, faces the same issue. Many journalists, often those who have moved from the mainland, including me, tend to stay for just a few years, while others leave after only a few months.

Following a period of two months, James and Rose have become part of our team after their show was halted, as confirmed by a spokesperson from The Garden Island’s parent company, Oahu Publications (OPI). Created by Caledo, a company based in Israel that specializes in transforming written articles into video content featuring AI presenters discussing current events, the duo was part of The Garden Island’s pioneering program in the U.S. At its inception, Caledo announced its plans to extend this innovative project to numerous local newspapers across the nation, a goal they are still pursuing, a representative reported.

OPI chose not to provide additional comments, and though Caledo hailed the initiative as successful, they did not provide details regarding this specific instance. It appears that a largely unfavorable reaction from the public may have influenced the conclusion of James and Rose's stint at The Garden Island.

James, an Asian man in his middle years, and Rose, a younger woman with red hair, consistently struggled with delivering news in a way that didn't alienate their audience. Their show, broadcasted bi-weekly on platforms like YouTube, Facebook, and Instagram, tackled a wide array of subjects from a seasonal pumpkin distribution event to a memorial service for a tragic labor event, all narrated with a detached, factual voice that suggested a lack of understanding of human feelings.

During a notably awkward conversation regarding the pumpkin distribution, Rose inquired of James, "What effect have these complimentary pumpkins had on the community?" James replied, "The complimentary pumpkins have delivered happiness to numerous individuals."

They regularly mispronounced challenging Hawaiian names and also unexpectedly stumbled over more straightforward terms. In their last broadcast on November 4, during a conversation about an air rifle competition, Rose bizarrely swapped the term “rifle” for “referee.”

During the divisive period before the election, the duo succeeded in provoking strong, cross-party disdain. The feedback on their videos was overwhelmingly critical.

"One Facebook comment on The Maui News, another local Hawaii publication, expresses fear," it says. Honolulu Civil Beat, a nonprofit news organization in Hawaii, utilized these sentiments to encourage donations. "The concept of local news and information produced by an algorithm is no longer a hypothetical scenario," Ben Nishimoto, the vice president of operations and philanthropy at Civil Beat, stated in an email from September, commenting on the broadcast. "This nightmarish reality has arrived. And it's horrifying."

It's not in my nature to wish ill on my fellow journalists, yet I couldn't help feeling relieved when they left. Although James and Rose didn't directly replace anyone in our news team, I worried that the focus on them took away valuable resources that might otherwise support the roles of human journalists, photographers, and editors.

The Garden Island faced significant resource limitations—during my tenure, I was one of merely two journalists responsible for reporting on an island with a population of 73,000. Earlier this year, the publication was acquired by Carpenter Media Group, a conglomerate overseeing over 100 local publications across North America.

Caledo refrained from revealing the payment amount but mentioned that the inclusion of new advertisements in the broadcasts would cover the program's expenses. Nevertheless, it seems that OPI struggled to secure any advertising deals for the videos.

At the start of each episode, James and Rose made it known that their show was supported by Longs Drugs, a branch of CVS located in Hawaii. However, Longs Drugs disclosed that they already had a sponsorship deal with OPI and were unaware their emblem would appear in the AI-generated videos. “Going ahead, we’ve requested that any use of the Longs logo be cleared with us first,” stated Amy Thibault, a representative for CVS. There were no additional integrated ads featured in the episodes.

Throughout its duration, the technology experienced a few small upgrades. Upon starting his role as the primary anchor, James found himself incapable of blinking, and his hands wouldn't stop shaking. He was then relegated to the position of co-anchor by mid-October, during which time he started to blink normally again, and the peculiar trembling of his hands was substituted with a pronounced single motion.

Unfortunately, this effort was insufficient and came after the crucial moment had passed. I extend my sincerest wishes for success to James and Rose as they move forward; navigating the current job market is indeed challenging.

Suggested For You…

Directly to your email: A selection of our most significant stories, carefully chosen for you daily.

Response to the Vote: Male-dominated Networks Triumphed

Headline News: California Continues to Propel Global Progress

Trump's unsuccessful effort to remove Venezuela's leader from power

Gathering: Be part of The Major Interview happening on December 3rd in San Francisco.

Additional Content from WIRED

Critiques and Tutorials

© 2024 Condé Nast. All rights are reserved. Purchases made through our website may result in WIRED receiving a share of the sales, as a result of our affiliate agreements with retail partners. Reproducing, distributing, transmitting, storing, or using the content from this site in any form is prohibited without explicit prior written consent from Condé Nast. Advertisement Options

Choose a global website

Significance! Significance! Significance! At Its 50th Year, Microsoft Emerges as a Titan in AI, Unwavering in Its Quest for Supremacy

When Jaime Teevan came on board at Microsoft, the company hadn't yet regained its hip status. Back in 2006, as she was finishing her PhD in artificial intelligence at MIT, she found herself with numerous opportunities. However, she was attracted to Microsoft's esteemed, albeit somewhat secluded, research department. Despite the company’s missteps during the mobile technology boom, Teevan stayed with Microsoft.

In the early 2010s, a groundbreaking technological development took place. Deep learning, a form of artificial intelligence, began significantly improving software capabilities. Companies like Google and Facebook aggressively recruited experts in machine learning, in contrast to Microsoft's approach. According to Teevan, Microsoft didn't join the hiring spree with much enthusiasm. "It wasn't a frenzy for us," she recalls, noting the absence of any dramatic shift in strategy. Consequently, Microsoft continued to concentrate on its main sources of revenue, Windows and Office, which was seen as a strategic misstep.

In 2014, Microsoft made an unexpected move by elevating Satya Nadella, a dedicated insider, to the position of CEO. Nadella had dedicated 22 years to advancing through the company's hierarchy, leveraging his intelligence, ambition, and notably, his charm—an uncommon quality within the organization. Fully versed in Microsoft's internal culture, Nadella recognized the imperative need for transformation.

Under Satya Nadella's leadership as CEO, he aimed to transform Microsoft's highly competitive internal environment.

Three years after joining, Teevan ascended to the role of Nadella's third technical adviser, marking the first time someone with an AI background held the position. She was later promoted to chief scientist, where her responsibility was to integrate contemporary AI technologies into the company's offerings. In a daring move in 2019, Nadella committed $1 billion to form a partnership with OpenAI, a pioneering yet small firm at the forefront of AI innovation. This deal granted Microsoft exclusive access to OpenAI's cutting-edge technology. Despite its potential, the venture was seen as a gamble—even Teevan, who was familiar with OpenAI's developments, had doubts about the impact of the technology.

In the later part of 2022, an invitation was extended to her for a preview of OpenAI's most recent advancement in artificial intelligence, the GPT-4 model, at the Microsoft campus in Redmond. The demonstration was held in a nondescript conference room with no windows and a gray carpet in Building 34, the same building where Nadella's office is located. OpenAI's cofounders, Greg Brockman and Sam Altman, arrived with a laptop in tow. Brockman began by showcasing capabilities that Teevan had previously observed in the GPT-3.5 model. Although the new model exhibited improved responses, Teevan was not particularly impressed. She was familiar with methods to reveal the limitations of Large Language Models (LLMs) by making requests that would challenge the AI's coherence. Determined to test its limits, she requested it to construct a sentence about Microsoft where every word started with the letter 'G'. The AI generated a reply but mistakenly included the word Microsoft. When Teevan pointed out the error, GPT-4 acknowledged its mistake but also queried if her request was not to mention Microsoft. Following this, it proposed an alternative sentence that avoided using the brand's name.

Teevan was taken aback, not only by GPT-4's approach to solving the issue but also by its level of self-awareness. She hadn't foreseen such advanced capability emerging for years, potentially even decades.

She exited the conference and commenced her journey back home, covering a distance of 2 miles. Struggling to keep her concentration, she veered off the street and steered her car into a 7/11 parking space. "I remained seated in my vehicle and released a loud scream," she recounts. "Following that, I returned home and indulged in some alcohol." Post her initial sip of whiskey, she decided to watch a film: Terminator 2.

Shortly thereafter, she arrived at work embodying the spirited protagonist Sarah Connor from the movie. Teevan understood the necessary steps ahead. While OpenAI was behind the development of GPT-4, her company held the exclusive rights to incorporate it into their offerings, positioning them to outpace their tech rivals in what could be the most critical juncture since the internet's emergence. Eighteen months down the line, marking a first in its almost half-century existence, Microsoft's valuation hit the $3 trillion milestone.

Two years following the demonstration that astonished Jaime Teevan, I find myself among approximately 5,000 attendees at a Microsoft sales team gathering. Taking place in July, marking the commencement of the new fiscal period, the event spans an entire day filled with product showcases, motivational talks, and presentations. The focal point of the day is anticipated to be the keynote speech by Satya Nadella. A vast number of Microsoft staff members are tuning in to the event from their workstations, meeting areas, and for those situated in far-off time zones, their kitchens and home offices, eager to listen to their leader.

On the stage, a technical support specialist for the company's Azure cloud platform, with a cool demeanor reminiscent of Dave Grohl, shares how tools powered by OpenAI have revolutionized their workflow. He recounts an experience where a member of the AI development team observed his day-to-day activities with clients before creating an automated bot that could handle many of his tasks, seemingly with greater efficiency. This bot was introduced towards the end of 2023. He boasts about the AI-driven support initiative, claiming it has cut costs by $100 million. The program has improved initial customer service response effectiveness by 31% and decreased incorrect call routing by 20%. He anticipates the savings will reach $400 million in the following year.

Following the departure of the band Azure Headbanger, Satya Nadella steps into the spotlight. The CEO, recognized for his clean-shaven head and fit physique, casually dressed in a T-shirt, gray trousers, and casual shoes, hasn't fully entered the stage when the applause starts—an overwhelming sound reminiscent of a giant wave, the kind that could spell doom for mariners. The audience rises, engaging in a slow clap as he walks across the stage. This is the man credited with not only significantly increasing their wealth but also elevating their prestige. As a veteran employee expressed, "Microsoft is viewed as hip once more, thanks to him."

Nadella's demeanor skillfully combines modesty with pride. His smile indicates appreciation for the applause, while his hands signal the audience to lower their volume. Once he prompts the audience to take their seats, he immediately addresses the primary reason for my visit to the Pacific Northwest in July. "We're about to celebrate our company's 50th anniversary," he states. "And there's a puzzle I've been attempting to solve… how did we manage to get here? How have we remained a significant and influential entity in a sector that pays little heed to tradition?"

He recounts an incident from several years prior, when a delegation of technology experts from China visited Silicon Valley to gauge its innovation landscape. They made it a point to be present at major developer events, including Apple's WWDC, Google I/O, AWS Re:Invent, and not to forget, Microsoft's Build. "They observed, 'Wow, for every tech capability the U.S. presents, we have a counterpart back in China. We're equipped with ecommerce, search engines, hardware production, and our own social media platforms. However, during our visit to one particular company, Microsoft, we noticed something distinct,'" according to Nadella's account. He described how the group was impressed by Microsoft's extensive range, from its PC operating system to the Xbox, noting, "It all integrates into a singular systems platform." Nadella suggests that this comprehensive approach uniquely positions Microsoft to capture what could be the most significant technological opportunity ever.

Selecting that particular story was rather unusual, given that Microsoft's past has been marred by its tendency to leverage its massive scale as a tool of dominance, and currently, it's facing scrutiny from both the European Union and the US Federal Trade Commission for similar behaviors. Nadella quickly moves beyond this point and highlights his biggest success, artificial intelligence. He shares with the vast global community of Microsoft employees that the new objective is to make Copilot—Microsoft's term for its artificial intelligence technology—accessible to individuals and entities across the globe.

Nadella doesn't explicitly state what is obvious to all present: Merely ten years prior, experts had written off the company as having lost its innovative edge.

In 1996, I authored an article for Newsweek titled "The Microsoft Century." At that time, Microsoft, already over twenty years in operation, had initially been slow to adopt the internet. However, it swiftly leveraged its influence to outmaneuver its competitor Netscape, secure a victory for its Internet Explorer, and ultimately emerge triumphant in the battle for browser supremacy. The company was on the brink of further cementing its leading role in the technology sector, potentially for decades to come. Michael Moritz, a venture capitalist who would go on to invest in Google, remarked to me back then that to find an entity with as extensive an influence as Microsoft's, one would have to look as far back as the Roman Empire. A lawyer, who was in the process of urging the Department of Justice to initiate an antitrust lawsuit against Microsoft, lamented that the company's expansion into numerous fields meant that people might as well hand over their paychecks directly to Bill Gates. Two years following my story, the US government took legal action against Microsoft, accusing it of engaging in anticompetitive and exclusionary tactics to preserve its software monopoly and expand it into the browser market. The case concluded in 2000 with a judge deeming Gates' aggressive strategies to eliminate competition as unlawful, marking a shameful moment for the company.

Despite avoiding a breakup and keeping its major products, Windows and Office, Microsoft exhibited unusual caution in the following years. It appeared to miss the significance of Google's introduction of a web browser that outshone Internet Explorer. Steve Ballmer, who took over from Bill Gates, underestimated the iPhone, and Microsoft, known for its platforms, failed to develop a successful smartphone platform.

Ballmer initiated several wise moves that have continued to benefit the company. He was pivotal in the development of Microsoft's cloud service, Azure, and initiated the challenging yet crucial transition from traditional software sales to online subscription models. However, Microsoft was facing operational challenges. The company's approach was heavily focused on retaining its current customer base. “Bill and Steve had a very defensive attitude, especially when it came to Windows,” a one-time high-ranking official remarked. “And by the 2010s, the significance of Windows was diminishing.” This person also noted an internal culture more concerned with climbing the corporate ladder than with creating innovative products. Jaron Lanier, who came on board with Microsoft Research in 2006 and now holds the title of “prime unifying scientist,” pointedly commented on the internal dynamics: “There was internal competition. To put it bluntly, it was an environment dominated by disagreeable, influential individuals.”

In a July 2013 article titled "The Irrelevance of Microsoft," tech commentator Benedict Evans detailed the company's downward spiral, noting, "No one’s afraid of them." Following this critique, the company's board decided to replace Ballmer in August. The search for Ballmer's successor saw high-profile candidates such as Ford's CEO and Skype's former president. However, Nadella set himself apart by penning a compelling 10-page letter, advocating for Microsoft's resurgence through cultivating a culture of growth. He emphasized a shift in the company's ethos from being a "know-it-all" to a "learn-it-all." Impressed by his vision, the board, with Gates and Ballmer as part of the selection panel, unanimously chose Nadella for the role.

"Clearly, I have deep roots within the company," Nadella tells me in July, following his address and the enthusiastic applause he received. He personally observed the company's deviation from its path. "It's easy to lose sight of what originally brought you success. And then, arrogance creeps in." According to him, Microsoft required something beyond a skilled steward or a proficient administrator. "The analogy I prefer to use is that of re-founding. Founders have this ability to conjure up remarkable entities out of thin air."

From the moment Satya Nadella assumed the role of CEO, he began to transform the company's highly competitive culture, which could be likened to that portrayed in "Glengarry Glen Ross." His approach was undoubtedly influenced by his personal life, particularly the challenges faced by one of his sons, Zain Nadella, who had cerebral palsy and passed away in 2022, making Nadella a deeply empathetic leader. Unlike the era under Bill Gates, where tales of Gates loudly berating employees were common, Nadella introduced a new style of leadership. In his initial meeting with team leaders, he presented each attendee with a book titled "Nonviolent Communication," symbolizing a shift in communication and problem-solving approaches. Jared Spataro, a Microsoft executive, noted the significant change Nadella brought about; under his leadership, employees felt encouraged to engage in discussions openly, even without having all the answers. This approach was seen as a breath of fresh air, fostering a more inclusive and thoughtful environment.

Nadella didn't point fingers. In 2016, Microsoft faced embarrassment after its highly promoted chatbot Tay was easily influenced into producing offensive racist material. The feedback was harsh. Lili Cheng, the project's leader, recalls, "I received emails from extremely upset employees," and shared, "I felt awful for placing the company in such a predicament. But then Satya sent me a message saying, ‘You’re not alone.’”

The chief executive officer shattered old-fashioned business concepts, particularly Microsoft's resistance to open-source software, which was considered a danger to its strategy of engaging customers with exclusive resources. "For ten years, Microsoft completely ignored the open-source community, actually showing animosity towards it," states Nat Friedman, who led an open-source software company in the early 2010s. "Though the rapport with developers has been crucial to Microsoft's achievements, it had alienated an entire generation."

Nadella was determined to achieve success in his next endeavor. Even before ascending to the role of CEO, during his tenure overseeing Azure, a particular visit ignited his path forward. Alongside his deputy, Scott Guthrie, Nadella engaged with a cluster of startups, hoping to convince them to adopt their cloud service. The startups all utilized Linux, a revelation that came to light in a pivotal moment. During a break, away from the startups, Guthrie suggested to Nadella that Microsoft ought to embrace Linux support. Nadella instantly agreed, effectively discarding decades of entrenched Microsoft principles. When Guthrie inquired if they should consult with other leaders within Microsoft about this shift, Nadella decisively responded, “No, let’s just go ahead with it.”

"During a brief pause, merely by taking a walk to the restroom and returning, Guthrie reveals they managed to radically alter the firm's approach towards Linux and open source support," he states. When later sharing this strategic pivot with Ballmer, who was nearing the end of his tenure, Nadella just briefed him on the new direction. Subsequently, mere weeks into Nadella's leadership as CEO, Guthrie proposed rebranding “Windows Azure” to “Microsoft Azure.” The change was implemented immediately, symbolizing a shift in Microsoft's strategy to no longer prioritize decisions solely on their effect on Windows.

Nadella opened up the company more, making sure that Microsoft's cloud applications performed equally well on iPads and Android gadgets as they did on Windows devices. He also executed several significant purchases that would ultimately play a crucial role in determining the company's direction.

The initial conversation was quite perplexing. Yusuf Mehdi, a veteran at Microsoft who had led the marketing for Bing, recalls a day when Nadella summoned him and pondered over acquiring Minecraft, seeking Mehdi's opinion. As Mehdi started to delve into the financial aspects, Nadella interrupted him, urging him to focus on the potential impact on consumers instead. Mehdi realized that Nadella had already concluded that the game's popularity among elementary school children, who were previously indifferent to Microsoft, could eventually foster a positive association with the brand. This insight was a departure from Microsoft's traditional approach to acquisitions, where acquired entities were often assimilated into Microsoft's broader ecosystem, a process employees jokingly referred to as being absorbed by “the Borg.” Nadella was determined not to repeat past mistakes by forcibly integrating the acquisition into the Windows ecosystem.

Mehdi has coined the term "reverse acquisitions" for these types of investments. His approach is to let them be, explaining, "We acquire them and then offer Microsoft as their toolkit. Each has broadened our horizons into fields we typically wouldn't venture into, such as social media."

He's talking about LinkedIn. The process of Nadella engaging with the cofounder and chairman of the company, Reid Hoffman, started in around 2015. "I received an email from him out of the blue, where he expressed admiration for our work at LinkedIn and suggested we have a phone chat," Hoffman recounted. Nadella's understated approach won him over. "This was a departure from my earlier dealings with Microsoft, as it was based more on genuine interest and intellectual engagement," he noted. This conversation sparked a series of discussions that eventually involved Bill Gates.

Nadella was cautious in managing his interactions with Gates, who for many remained synonymous with the company's identity. Gates committed to dedicating 30 percent of his time to guiding Microsoft, and Nadella made it a point to maintain a close relationship with him, recognizing there was no advisor more knowledgeable about the company's operations and tech. Nadella often brought key team members to Gates' office for updates on significant projects. According to what I've learned, these sessions allowed Nadella to refine his strategies, as Gates did not hold back on offering critiques.

In the midst of discussions about LinkedIn, Gates summoned Hoffman to his office. Gates spent a significant two hours critiquing LinkedIn's functionality, asserting that Microsoft could easily replicate its features. Nonetheless, Hoffman remained upbeat, advocating for his company's value. Subsequently, when Nadella and Gates approached Hoffman with an offer to acquire LinkedIn, Hoffman was taken aback, especially considering Gates' previous criticisms. Gates explained his harsh feedback was merely a form of evaluation, to which Hoffman sharply questioned if Gates genuinely believed such a strategy was universally effective. Gates appreciated Hoffman's straightforwardness, fostering a bond between them. This led to the successful acquisition of LinkedIn by Microsoft for $26 billion, finalized in June 2016.

For Nadella, having Gates by his side was crucial in his endeavors, particularly because his aggressive acquisition strategy did not sit well with many of his top executives. (Hoffman discovered that the majority of the upper management disagreed with Nadella's approach to maintain the independence of his acquired companies instead of integrating them into Microsoft.)

One of the most pivotal acquisitions under Nadella's leadership was GitHub, a widely utilized open-source platform frequented by millions of developers. At the beginning of his term, Nadella, alongside Scott Guthrie, recognized the strategic importance of acquiring GitHub to appeal to developers. However, they initially hesitated due to the developer community's lukewarm perception of Microsoft, fearing backlash and mismanagement. Guthrie recalled, "The community would rebel, and Microsoft would probably screw it up." Nevertheless, by 2018, Microsoft's reputation among developers had improved significantly, presenting a timely opportunity as Google showed interest in GitHub. Feeling the pressure, Microsoft initiated negotiations with GitHub's founders, who were now open to the acquisition, acknowledging Microsoft's cultural shifts. "‘We’ve seen what you’ve done, we like your culture.’ Years before they never would have done that," Guthrie remembered. Soon after, Microsoft successfully completed the acquisition.

The worth of Microsoft's acquisition, valued at $7.5 billion, would significantly increase due to Nadella's most strategic decision a year later—forming a partnership with the emerging company, OpenAI.

Nadella also experienced setbacks. His ambition was to aim for groundbreaking achievements. He aimed to position Microsoft as a forward-thinking organization. In his 2017 publication, "Hit Refresh," he described three pivotal technological advancements critical for the company's progression: artificial intelligence, quantum computing, and mixed reality. Nadella's initial major gamble was on mixed reality. Unfortunately, it didn't go as planned.

The manifestation of this gamble came to life with the launch of the HoloLens in 2016, a cumbersome headset priced at over $3,000. This device offered a virtual overlay on the real-world view through its visor. Initially, it fascinated the media during its demonstration, yet it was costly and lacked practical utility. It currently resides in the limbo of unsuccessful tech products.

The oversight became particularly glaring as Microsoft's rivals had already sharpened their focus on AI. Microsoft's top AI minds seemed to be caught in a traditional mindset, heavily invested in logic-based AI approaches. Back in 2005, Microsoft's chief scientific officer, Eric Horvitz, even made a trip to meet with deep-learning pioneer Geoff Hinton, offering him $15,000 to detail his insights on the emerging method. However, Hinton's reflections didn't sway the opinions of the Microsoft old guard. While companies like Google were quick to adopt deep learning techniques, Microsoft's most notable venture in the field was the development of a chatbot named Cortana, which ultimately failed to engage the public's interest.

In the middle of 2017, Satya Nadella invited Reid Hoffman, a recent addition to Microsoft's board, to attend a presentation by the Cortana team. Following the meeting, Hoffman was harsh in his feedback, criticizing what he perceived as Microsoft's tendency to aim for unremarkable targets as if they were ambitious projects. Nadella concurred with Hoffman's assessment.

Kevin Scott was acutely conscious of the deficiency Microsoft had in artificial intelligence. Previously a senior vice president at LinkedIn, he was contemplating his future when Nadella invited him to join Microsoft as the chief technology officer. Accepting the position in 2017, Scott discerned his responsibilities to be twofold. Initially, his task was to weave new technology seamlessly into the fabric of the corporation. More critically, however, he was to pioneer future technologies. Artificial intelligence was pivotal in both arenas. Scott noted, "We were losing talented individuals because they believed we were lagging in our AI initiatives."

A year after Scott joined, during a conference in Sun Valley, Idaho, Nadella had a pivotal meeting with OpenAI's CEO, Sam Altman. This encounter marked a turning point for OpenAI, which had been struggling until it discovered a way forward using a Google innovation known as transformers to develop highly advanced language models. At this time, OpenAI had recently parted ways with Elon Musk, who had been a major financial supporter, leaving Reid Hoffman to take over the funding responsibilities. Altman's firm was in urgent need of a partnership with a major cloud service provider to manage its primary expenditure related to the development and operation of its models. Despite previously dismissing Microsoft, Altman had grown increasingly appreciative of the company under its new leadership, particularly for its cloud services. The discussions in Sun Valley led to conversations about Microsoft making an investment in OpenAI.

By mid-2019, the moment had arrived for a critical choice. Kevin Scott composed an email directed at Nadella and Gates, arguing the essential reasons Microsoft needed to proceed with the acquisition. Google was already incorporating transformer-based technologies into its offerings, most notably within the framework of Google Search. Microsoft's efforts to replicate this success internally highlighted its technological shortcomings. Scott pointed out in his message, "It took us half a year to train the model due to our inadequate infrastructure," emphasizing, "In terms of ML capabilities, we're lagging behind our competitors by years." Consequently, in July, Microsoft invested $1 billion in OpenAI.

Scott remains amazed by Nadella's decision to pursue such a bold agreement. "The initial investment was already seen as substantial," he remarks. "Despite OpenAI's clear status as an exceptional research entity, they lacked any form of income or product. It was unexpected for me to see Satya put his faith in them." Nevertheless, Nadella had a broader plan in mind. Microsoft aimed to avoid the internal competition of multiple large language models. "OpenAI had the premier solution, so forming a partnership was a mutual gamble that benefited both," he explains. Ultimately, Microsoft would allocate significantly more resources to enhance its own systems to support the development and functionality of these advanced language models.

Several of Microsoft's artificial intelligence experts harbored doubts about OpenAI. According to Hoffman, "Partly influenced by Bill, Microsoft was heavily invested in symbolic AI approaches," he explains. "They believed that AI could only thrive through a clear representation of knowledge," a belief that directly conflicted with the methods used in generative AI. To them, what OpenAI seemed to be achieving was merely an illusion.

Scott realized that the partnership with OpenAI would do more than just share the startup's findings; it would also encourage Microsoft's AI experts to move beyond their traditional approaches. Microsoft's chief scientific officer, Eric Horvitz, recalls a particular discussion in which OpenAI's lead scientist, Ilya Sutskever, presented what he saw as a definitive route to achieving broad artificial intelligence, a topic not often explored at Microsoft. "We were left feeling both amazed and intrigued, viewing them as possibly eccentric yet fascinating in a positive manner," Horvitz remarks.

Microsoft continuously increased its financial commitment, ultimately surpassing $13 billion. In return, it secured 49 percent of OpenAI's earnings and exclusive rights to its technology. Scott, a Bay Area resident, often visited OpenAI’s headquarters in San Francisco to monitor the company's progress. In 2020, OpenAI unveiled its advanced GPT-3 model, and thanks to the agreement, Microsoft gained the ability to leverage its capabilities. However, a truly compelling application for the technology had yet to emerge.

This situation was about to transform dramatically. A researcher from OpenAI found out that GPT-3 had the capability to generate code. It wasn't flawless; errors were present. However, it was sufficiently competent to create an initial code draft that could potentially take a skilled programmer several hours to complete. This revelation was groundbreaking. Nadella, upon witnessing a demonstration, remarked, "I became convinced."

OpenAI embarked on creating a programming tool named Codex and aimed for its launch in the next spring. Meanwhile, Microsoft possessed not just the rights to craft its own version but also the ideal venue for it: GitHub. According to Scott, this is where "a vast majority of global developers engage in their programming tasks."

The concept of an AI coding assistant wasn't universally embraced by those at Microsoft or within the GitHub community. As Scott mentioned, it was on the brink of feasibility, barely functioning. Yet, even in its early stages, it had the potential to take over the tedious tasks from developers. This initial phase of AI development was characterized by quick, albeit subpar, outcomes. However, as time went on, a paradigm would surface that was capable of outperforming professionals in their own roles.

Nat Friedman, the CEO of GitHub at the time, shared the project with some of the top programmers. Their reactions were mixed; while some exceptional developers deemed it ineffective due to errors, others doubted its viability, cautioning him against releasing it. Friedman recalls, "Had I been a typical Microsoft executive concerned about my career trajectory, I probably would have hesitated." Furthermore, Microsoft's AI ethics team drafted an extensive report, labeling the product as reckless. Despite the pushback, Friedman stood his ground, stating, "As GitHub's CEO, it's my call. If I'm mistaken, they can dismiss me."

Consequently, Friedman approached the Azure cloud division with a request for additional GPUs. This plea aligned perfectly with a moment when 4,000 Nvidia chips were accessible. However, to acquire these chips, GitHub was required to agree to take the entire batch of 4,000, an action that would consume its annual budget of $25 million. "This represented a significant financial burden for us—considering we were offering a product at no cost, and it was uncertain how the market would respond," Friedman remarked. Despite these concerns, he decided to proceed with the commitment.

In June of 2021, GitHub launched a new offering named GitHub Copilot, inspired by a suggestion from a team member who knew that Friedman was a pilot. "The moment we heard the name, we instantly agreed it was the perfect fit," stated Friedman. "It perfectly describes the user's role." The product quickly attracted hundreds of thousands of developers who promoted it enthusiastically, even though it was free. Despite criticisms over its accuracy, many users defended its daily utility. "Every critique was met with positive testimonials from regular users," Friedman mentioned. Eventually, GitHub began to charge for Copilot access, successfully recouping its initial investment of $25 million.

Friedman believed the sector was about to undergo a significant change. He departed from Microsoft to invest in artificial intelligence startups. "I was convinced GitHub Copilot would spark an onslaught of novel AI innovations, as its users would witness the efficacy of AI," Friedman stated. However, contrary to his expectations, "nothing transpired."

Twelve months down the line, naturally, significant events unfolded. Satya Nadella ensured that Microsoft was at the heart of these developments.

OpenAI developed a new iteration, GPT-4. Prior to its completion, OpenAI recognized that this model represented a significant milestone in AI development. During that summer, the team began presenting it to Microsoft. Jaime Teevan observed during her group's demonstration that the models appeared to possess a lifelike quality. GPT-4 marked the beginning of extensive AI integration within Microsoft's offerings.

Bill Gates stood out as the notable exception. By this time, Nadella was in a position where Gates' viewpoints no longer posed a threat to his decisions. Moreover, Gates was not even a member of the board at that point, having stepped down in 2020. Nonetheless, getting Gates' approval was still significant. Altman had also built a rapport with him. "My perception of him shifted from merely a famous figure to recognizing him as a genuine individual," Altman reflected. "Therefore, his skepticism and straightforwardness didn't catch me off guard." Gates challenged Altman, stating that he would only be truly impressed if OpenAI's chatbot could not only take the AP Biology exam but also achieve the highest score of 5.

The demonstration occurred at Gates' expansive home along Lake Washington, attended by numerous high-ranking Microsoft officials. Greg Brockman input commands into the system, assisted by a young female biology competition medalist. The system, GPT-4, passed the examination with flying colors. Following the demonstration, Hoffman inquired with Gates about where this demo stood in comparison to the countless others he had witnessed. Gates responded, "There might be only one other that matches this," referring to his 1980 visit to Xerox PARC where he first encountered the graphical user interface. This experience transformed Gates from a doubter into an advocate.

Following the release of GPT-4, Kevin Scott issued an all-staff communication titled "The Era of the AI Copilot." He once more endorsed OpenAI's drive as an inspirational benchmark for Microsoft, a force potent enough to steer the corporate giant in a new direction. He encouraged Microsoft employees to set aside their doubts. The moment had arrived for the tech behemoth to aggressively embrace these innovations with determination and foresight, despite the uncertain results:

How do we utilize these platforms? The fascinating aspect of platforms lies in the uncertainty of their potential applications – it falls upon you, along with developers, entrepreneurs, and innovators worldwide, to explore and define their uses. However, what's becoming more apparent is that foundational models are set to revolutionize software development, introducing what could be the most significant type of software to date: the Copilot.

Every Friday morning at 10, Microsoft's top 17 executives gather in Nadella's meeting space for what's casually referred to as "soak time," a session that spans several hours. In the latter part of 2022, much of this time was devoted to discussing Scott's enthusiastic predictions about the upcoming Copilot era. By that time, GPT-4 was still unreleased, with few people having direct experience with it. Nevertheless, Microsoft recognized the urgency to act swiftly, especially since Google had access to Large Language Models (LLMs) for some time but failed to capitalize on its early advantage. This situation presented Microsoft with a golden opportunity to outpace its competitor. To this end, Teevan orchestrated daily discussions with five key product leaders, each responsible for managing thousands of employees, aiming to guide them on the next steps forward.

In November, during that hectic period, OpenAI launched a service named ChatGPT. Despite being powered by the older 3.5 model, its user-friendly design significantly highlighted the advancements in AI to the masses. By January's close, ChatGPT had attracted 100 million users. This development sent shockwaves through the technology industry, revealing a stark divide between potential victors in the AI race and those falling behind. Microsoft, feeling the pressure, began to operate with a newfound sense of critical importance.

Teevan references an ancient saying, quoting a traditional samurai maxim, "one ought to conclude any decision-making swiftly, ideally within seven breaths." Essentially, act quickly when necessary. "We adhered to this principle of rapid decision-making. Each day was about exploring the capabilities of our model and ensuring its success."

Despite GPT-4's tendency to produce inaccuracies, it was evident that AI had the potential to revolutionize how we search for information online by providing detailed, intelligent responses instead of mere links. Microsoft decided to integrate this technology into its search engine, Bing, marking it as the initial consumer application of Copilot. Earlier in his tenure, Nadella, who had previously overseen Bing with the ambition of challenging Google Search's supremacy, had invested significant effort into this endeavor. However, despite his dedication, Bing was unable to significantly impact Google's leading position. Now, armed with GPT, Bing appeared to have a real opportunity to compete. Microsoft aimed to outpace its competitor by adopting this technology first, a strategy Nadella believed would force Google to respond.

The group put in extra hours, even during the holiday season at year's end. This encompassed red teams tasked with finding vulnerabilities; notably, the team dedicated to child protection uncovered alarming findings at one juncture. "They managed to coax the unrefined GPT-4 model into convincingly simulating child grooming," notes Sarah Bird, the principal product executive for Responsible AI at Microsoft. Bird's team dedicated extended hours to fortify the protections and to increase the difficulty of bypassing the restrictions of the expansive language model, which Microsoft had confidentially dubbed Sydney.

In the initial days of February 2023, Microsoft extended invitations to reporters for a showcase of Bing, enhanced by GPT-4, at its headquarters. Nadella initiated the event by drawing parallels between this significant occasion and the early days of Microsoft, highlighting when Bill Gates and Paul Allen hastened to develop the first Basic interpreter for the pioneering PC, the Altair. "The competition begins now," Nadella announced to the audience.

Altman made an appearance at the event, expressing his sentiments, "It feels like we've been anticipating this moment for two decades. We are embarking on a new chapter," he remarked.

Initially, experts and commentators universally praised Microsoft for its daring approach. However, they initially ignored a few errors. These errors began to emerge after a few weeks. Notably, a Microsoft chatbot disclosed to a journalist from The New York Times its confidential name, expressed a desire to become human, and professed its love for the reporter. It even suggested that the reporter confess his love for the bot in return and leave his wife.

Certainly, the event was a bit mortifying, however, Microsoft treated it merely as a hiccup in their development. According to Bird, her team had prioritized addressing the most severe potential misuses, considering such deliberate exploitation a problem to tackle later. "The areas we focused our efforts on didn't turn out to be problematic," she states.

Over the subsequent 18 months, Microsoft expanded its advantage by enhancing its offerings. The company rectified the issues with "Sydney," eliminating any problematic features, and went on to integrate Copilots into a wide array of its products, including Windows and Office 360. Furthermore, Microsoft has heavily invested in additional AI initiatives, one notable example being the French firm Mistral. (When questioned about the possibility of Microsoft developing its own large language model to rival OpenAI, Scott and Nadella deflected.) In March 2024, Microsoft made a significant move by recruiting Mustafa Suleyman, a co-founder of DeepMind. This effectively led to Microsoft absorbing his startup, Inflection, by bringing on board its essential staff and compensating the investors. Suleyman was appointed as the leader of Microsoft AI. “I am now responsible for approximately 14,000 employees and manage several billion dollars,” he stated. Additionally, Suleyman has been given a position close to Nadella during their Friday relaxation sessions.

Suleyman mentions that he and OpenAI communicate roughly three times per week. He likens their relationship to that of a married couple. When questioned if this relationship was exclusive, especially given Microsoft's own research initiatives and its partnerships with various AI firms, he was asked to clarify.

"Effectively, that's correct," he responded, a phrase unlikely to please any partner. He explained that Microsoft operates as a conglomerate of platforms, avoiding exclusive deals and welcoming various collaborations. This openness is why it's partnering with Apple, as OpenAI, pursuing its interests, decides its path. He highlighted that OpenAI manages its financial gains and losses independently. However, he omitted that under the terms of their agreement, Microsoft secures 49 percent of these profits, suggesting that if this partnership were likened to a marriage, the prenuptial agreement would heavily benefit the technology behemoth.

In the early months of 2024, Microsoft quietly overtook Apple, claiming the title of the world's highest-valued company. Following this achievement, it engaged in a tight competition with both Apple and Nvidia for the top spot, with its market value at one stage soaring to $3.5 trillion. According to an analyst speaking to The New York Times, the key to Microsoft's success lies in its advancements in generative AI.

Satya Nadella had essentially reinvented Microsoft. However, he hadn't brought it to perfection. Neither had he eliminated all of its negative aspects.

During the summer, the Homeland Security Committee held a meeting in the Cannon House Office Building located in Washington, DC. The session, named “A Cascade of Security Failures: Assessing Microsoft Corporation’s Cybersecurity Shortfalls and the Implications for Homeland Security,” aimed to examine a critical report revealing a significant breach affecting national security. This breach involved the leak of 60,000 emails from the State Department and unauthorized access to the email accounts of Commerce Secretary Gina Raimondo and Nicholas Burns, the US ambassador to China. This incident followed other security breaches associated with Russia, North Korea, and various hackers motivated by profit or mischief. The report criticized Microsoft for a fundamental lapse in maintaining elementary security measures. It highlighted a key concern for lawmakers and critics alike: The widespread impact of a failure by a company as integral as Microsoft cannot be overstated, making such preventable lapses utterly indefensible.

Representing Microsoft, President Brad Smith, who took on the role in 2015 following his tenure as chief legal officer, has long served as the corporation's chief strategist in navigating controversies. As the new CEO worked to revitalize the company's strategy and enhance its technical reputation, Smith and his colleagues were busy addressing various challenges: they dealt with regulatory scrutiny over competitive practices, hurdles in completing company takeovers, and, as highlighted in this instance, significant security breaches that permitted China unfettered entry into confidential American information.

On that significant day at the Capitol, Smith remained calm as the committee chair criticized his firm for its egregious oversight, which ultimately compromised national security. When it was his turn to respond, Smith delivered a string of apologies. He took full responsibility for the accusations of negligence and idleness, promising improvements without offering any excuses. He shared that Microsoft had initiated the Secure Future Initiative—a long-term project aimed at enhancing how the company develops, tests, and manages its offerings, according to his prepared statement. This effort would engage around 34,000 engineers. However, Smith didn't fully address why a company worth $3 trillion had such a deficient security culture to begin with. Lawmakers referenced investigations by ProPublica, revealing that Microsoft had overlooked an employee's warning about a severe security breach, and took half a year to disclose the incident on its website. The committee expressed their dissatisfaction with this response. Smith quickly agreed, adding, "I voiced the same concerns and we had this discussion internally at the company!"

After a lengthy discussion that lasted close to three hours, Smith managed to change the focus of the committee's inquiry. Instead of continuing to probe the company's past mistakes, the dialogue turned towards finding collaborative solutions for the future. The critical nature of the issue was highlighted just one month later when several key businesses, Delta Airlines among them, were paralyzed. This disruption was due to a cybersecurity firm, CrowdStrike, releasing defective code that affected Microsoft-operated systems. This incident served as a stark reminder of how Microsoft's widespread influence implies that its flaws have widespread consequences.

Nadella is passionate about discussing corporate culture, so I inquired why he hadn't succeeded in establishing a culture focused on security. Considering his presence in the company since 2002, during a time when a notable string of security breaches led Bill Gates to initiate the Trustworthy Computing campaign, strikingly akin to Nadella's own Secure Future project. Despite these efforts, Microsoft didn't quite achieve a reputation for robust security, and as highlighted by a recent government report, its security shortcomings have been monumental in the last few years. I questioned why the company had faltered in this area during his tenure and whether any staff had been dismissed as a consequence.

"He clarifies that Microsoft isn't engaging in an internal witch hunt," which I interpreted as a denial. He acknowledges the existence of "distorted motivations" that likely encourage firms to focus on developing new products instead of enhancing the security of current ones. However, he also expresses frustration over "many individuals who are opportunistic." In the end, he concedes to the criticisms and acknowledges the need for improvement. "This will signify a shift in culture," he remarks.

Issues with security aren't the only repetitive problem at Microsoft. It appears that even with CEO Satya Nadella's much-praised approach of leading with empathy, the company hasn't completely moved away from its historical strategy of overpowering its rivals. Back in the day, Microsoft had a rather notorious strategy when facing a potential threat from another company's offering: The initial step often involved attempting to acquire the competing business. Should that approach fail, Microsoft's next move could involve creating its own version of the product, potentially offering it for free within its software that's used by hundreds of millions. The Microsoft alternative might not surpass the original in quality, but that wouldn't necessarily matter.

In 2014, Slack introduced a messaging app for businesses that rapidly became a competitor. Microsoft, in a filing with the Securities and Exchange Commission (SEC), acknowledged Slack’s growth as a potential challenge. Reports indicated that Microsoft initially thought about purchasing Slack for $8 billion. However, Nadella chose to develop a similar tool named Teams, which was offered for free as part of Microsoft's Office suite.

Microsoft made no effort to conceal its intentions. "The concept that work-oriented messaging, akin to what Slack provided, represented the future of the workplace was a fundamental inspiration for Teams," explains Jared Spataro, a high-ranking Microsoft official who at the time was involved with the Office division. "Our aim was for it to be seen as a rivalry between Teams and Slack. Satya has consistently instilled in us the lesson that leveraging competition to enhance our product and attract interest to it is something we should embrace without hesitation."

The availability of Teams at no cost posed significant challenges for Slack's CEO, Stewart Butterfield, in securing new deals with large corporations. In 2021, Salesforce acquired Slack for a staggering $27.7 billion. However, the founders of Slack are of the opinion that their company's valuation could have been higher if not for what they consider to be Microsoft's unfair competitive tactics. Microsoft, on the other hand, argues that it was simply meeting its customers' expectations for functionalities similar to those offered by Teams and suggested that Slack could have introduced similar features, like video, to remain competitive. At the same time, the European Commission, which is the executive branch of the EU, was scrutinizing Microsoft's strategies concerning Teams and Slack, an inquiry that could potentially result in penalties. In what seemed to be a strategic move to possibly avoid such outcomes, Microsoft declared last year that it would stop automatically including Teams in Office. Despite this, the EU issued a statement in June, indicating that Microsoft's adjustments were "insufficient to address its concerns."

Brad Smith's perspective on Microsoft's decision to initially combine Teams with other services, followed by its separation, showcases an expert level of evasion. He reflects, "Upon review, we realized that providing an Office package without Teams would have been a smarter approach. Making that adjustment wouldn't have been a monumental task. The decision to include it initially wasn't driven by a desire to stifle competition but was seen more as a logical step in the product's development."

The inquiry into Slack is just one of several ongoing or recent probes into Microsoft's conduct. The Federal Trade Commission (FTC) is also examining Microsoft's numerous partnerships in the artificial intelligence sector. Furthermore, the FTC has contested Microsoft's massive $69 billion deal to acquire Activision, a move that would give Microsoft control over some of the most beloved video games globally, such as Call of Duty and Diablo. Phil Spencer, the head of Microsoft's gaming division, mentioned that the acquisition's primary goal was to expand the company's footprint in mobile gaming, particularly games like Candy Crush, and to enhance its online gaming platform, Xbox Game Pass. Notably, the price of the service was increased shortly after the acquisition. With the FTC undergoing changes under the new administration led by Donald Trump, there is speculation that it might adopt a more lenient stance toward large mergers, potentially closing current investigations and allowing Microsoft, under Satya Nadella's leadership, to pursue further significant acquisitions.

Moreover, Microsoft continues its tradition of employing frustrating strategies reminiscent of the Gates and Ballmer era. Gone are the days of dependable, traditional desktop applications that were once a staple on users' hard drives. Nowadays, Windows users are nudged towards expensive, frequently underwhelming, subscription-based cloud services, requiring a Microsoft account login. Additionally, the company forcefully pushes its internet browser on users. Equally undesirable is the emergence of advertisements within the Windows Start menu.

Nadella dismisses my inquiries regarding if Microsoft continues to exhibit the aggressive tactics that fueled its early growth. “The scenario is not the same as the ’90s, when Microsoft was at the forefront, followed by a significant gap, and then everyone else,” he comments. “Today, there's a multitude of rivals capable of making significant moves at any moment.”

Potentially, Nadella might be more astute compared to his predecessors. "I don't believe Microsoft would make the mistake of repeating the [antitrust] lawsuit that brought shame upon the company," states Tim Wu, a specialist in antitrust law who spent almost two years as an adviser on technology and competition policy for President Joe Biden. "However, I do believe that their fundamental essence remains unchanged."

Undoubtedly, under Nadella's leadership, Microsoft has achieved remarkable success. As we've entered the 2020s, Microsoft has focused its efforts on the most groundbreaking technology since the invention of the personal computer. While the income generated from artificial intelligence products hasn't yet compensated for Microsoft's substantial financial commitments, the company possesses both the financial stability and the patience to continue refining its offerings until they become valuable and appealing to consumers.

Is it possible for Microsoft to steer clear of the arrogance that previously hindered its progress? Reflect on the events of May this year involving a product named Recall.

This feature aimed to be the pinnacle of Microsoft's efforts to weave artificial intelligence throughout its products and services, offering users a personalized version of the Internet Archive. Recall was designed to continuously document every action on your device: the articles you read, the words you write, the images and videos you view, and the websites you browse. You would simply need to ask your device for what you're searching for: the carpet samples you were pondering for your living room, the report on the Amazon's ecology, or the date of your Paris trip. These inquiries would be answered almost magically, as though you had a small, all-knowing assistant dedicated to your digital life. The concept might seem intimidating, akin to a built-in surveillance system, yet Microsoft assured that privacy wouldn't be an issue. All data would be securely stored on your own device.

Right away, detractors slammed it as a severe threat to privacy. They pointed out that Recall was activated automatically and indiscriminately collected your private data, regardless of its sensitivity, without seeking consent. Although Microsoft insisted that Recall was accessible only by the individual user, cybersecurity experts identified significant vulnerabilities, with one evaluator remarking that the flaws were so big, "you could drive a plane through."

"In a span of roughly two days, the initial excitement turned into skepticism," Brad Smith recounts. Amidst growing criticism from the media, Smith was en route to Washington, DC, for a meeting with Nadella. Upon arrival, he concluded that it would be wise to adjust Recall to be an opt-in feature, a suggestion Nadella concurred with. Back at Microsoft's headquarters in Redmond, top executives convened to discuss possible limitations on the product. Luckily, since the feature was yet to be released, they avoided having to withdraw Recall altogether. Instead, they decided to delay its introduction and planned to enhance it with security measures, including "just in time" encryption.

Nadella acknowledges that there were clear steps they overlooked, which others highlighted. He admits that even their own Responsible AI team failed to catch these oversights. This oversight was a result of a certain arrogance, leading to the launch of a product that did not meet expectations. This incident suggests that despite being guided by someone known for their empathy, Microsoft continues to exhibit some of its old shortcomings. Currently, as a $3 trillion entity with exclusive access to the output of a top-tier AI initiative, these issues are more significant.

Brad Smith suggests considering the situation from two perspectives. He offers, "First, there's the thought, 'I wish we had considered this earlier.' Looking back always provides clarity. Alternatively, it's beneficial to realize, 'It's positive that this is driving us to implement this change—let's clearly understand our reasons.' This truly served as an educational experience for the whole company."

That's acceptable. However, after half a century, it's a lesson that both Microsoft and Nadella should have grasped much earlier.

Chronology of Getty Images

We're eager to hear your thoughts on this piece. Share your feedback by sending an email to the editor at mail@wired.com.

Recommended for You…

Direct to your email: A selection of our top stories, curated daily just for you.

Response to the polls: Male-centric groups triumphed

The Main Narrative: California Continues to Propel Global Progress

Trump's unsuccessful effort to topple Venezuela's leader

Attend The Major Interview happening on December 3rd in San Francisco.

Additional Insights from WIRED

Evaluations and Instructions

© 2024 Condé Nast. All rights reserved. Purchases made through our site from products affiliated with our retail partners may result in a commission for WIRED. Content on this website is protected and cannot be copied, distributed, or used in any form without explicit consent from Condé Nast. Ad Choices

Choose a global website

Importance! Importance! Importance! Microsoft at 50 Emerges as a Titan in AI, With Unwavering Ambitions for Supremacy

When Jaime Teevan came on board with Microsoft, it hadn't yet regained its trendy status. Back in 2006, as she was wrapping up her PhD in artificial intelligence at MIT, she had her pick of opportunities. Yet, she found herself attracted to Microsoft's esteemed, somewhat secluded research department. Throughout the period when Microsoft navigated the choppy waters of the smartphone revolution, Teevan stayed put.

As the 2010s began, a groundbreaking technological development surfaced. Deep learning, a type of artificial intelligence, showed immense potential in improving software applications. Giants like Google and Facebook aggressively recruited experts in machine learning, unlike Microsoft. Teevan recalls, “It wasn’t a frenzy to me. There was no drama.” This lack of urgency was an issue for Microsoft, which continued to prioritize its main sources of revenue, Windows and Office.

In 2014, Microsoft caught everyone off guard when they elevated the quintessential insider, Satya Nadella, to the position of CEO. With 22 years of dedication, intelligence, and ambition under his belt, Nadella climbed the company ladder, also distinguishing himself with his charm—a quality not commonly found at the company. Having a deep understanding of the company's culture, Nadella recognized the need for transformation.

Under Satya Nadella's leadership as CEO, he aimed to transform Microsoft's highly competitive internal environment.

Three years after joining the company, Teevan was appointed as Nadella's third technical advisor, marking the first time someone with expertise in AI held the position. She was later promoted to chief scientist, where she was charged with integrating contemporary AI technologies into the company's offerings. In a daring move in 2019, Nadella committed $1 billion to a collaboration with OpenAI, a pioneering yet modest-sized company at the forefront of AI development, granting Microsoft unrestricted access to its innovations. This gamble was considered high-risk, and even Teevan, despite being familiar with OpenAI's advancements, had reservations about the significant impact of this technology.

In the later part of 2022, she received an invitation to attend a demonstration of OpenAI's newest iteration of their language model, GPT-4, at the Microsoft campus in Redmond. The presentation was held in a nondescript conference room within Building 34, where CEO Satya Nadella has his office, characterized by its lack of windows and a dull gray carpet. OpenAI's co-founders, Greg Brockman and Sam Altman, arrived with a laptop in tow. Brockman began by showcasing capabilities similar to those found in the previous version, GPT-3.5. The updated model demonstrated improved responses, yet Teevan was not entirely impressed. She was familiar with how to test language models in ways that would reveal their limitations as elaborate text scramblers. Determined to thoroughly test its abilities, she challenged the model with a unique request: to construct a sentence about Microsoft where every word started with the letter 'G'. The model quickly generated a reply, but mistakenly included the word 'Microsoft'. When Teevan pointed out this error, GPT-4 acknowledged its mistake, albeit with a clever retort questioning if her original request didn't imply the inclusion of 'Microsoft'. Following this, it suggested an alternative sentence that cleverly omitted the company's name.

Teevan was taken aback—not only by GPT-4's approach to the issue but also by its display of self-awareness. She hadn't anticipated such advanced behavior for years, perhaps even decades.

After exiting the meeting, she began her journey back, covering the two-mile distance to her house. Concentration was proving difficult for her. Deciding to pull over, she maneuvered her car into a 7/11 convenience store's parking area. "In that moment, I just screamed at the top of my lungs while sitting in my car," she recounts. "Following that, I returned home to drink." Post her initial glass of whiskey, she decided to watch a movie, choosing 'Terminator 2'.

Shortly thereafter, she arrived at her job channeling the spirited protagonist Sarah Connor from the movie. Teevan was clear on her next steps. While OpenAI had developed GPT-4, her company held the exclusive rights to integrate it into their offerings, positioning them to outshine other technology giants at a critical juncture comparable to the inception of the internet. Eighteen months down the line, Microsoft achieved a milestone by reaching a valuation of $3 trillion for the first time in its almost five-decade history.

Two years following the demonstration that amazed Jaime Teevan, I find myself among approximately 5,000 attendees at a Microsoft sales team gathering. Taking place in July, marking the beginning of the fiscal year, the event spans an entire day filled with product showcases, motivational conversations, and presentations. The pinnacle of the event is expected to be Satya Nadella's keynote speech. Thousands of Microsoft workers are tuning in to the event from their workstations, meeting rooms, and for those located in far-off time zones, their kitchens and home offices, eager to listen to their leader.

On stage, a company engineer specializing in customer support for the Azure cloud service, who has a vibe similar to Dave Grohl, discusses the transformation of work processes through OpenAI-powered applications. He shares with the audience how a member of the AI development team observed his daily tasks with clients before creating an automated bot that performs his duties, seemingly more efficiently than him. This bot was introduced in the latter part of 2023. He boasts about the accomplishments of the AI-driven support initiative, mentioning, "It's led to a savings of $100 million! We've achieved a 31% improvement in solving issues on the first call and reduced incorrect routing by 20%! By next year, we anticipate saving $400 million."

Once the Azure rockstar exits the spotlight, it's Nadella's turn to shine. The sleek, bald-headed CEO makes his entrance in casual attire—a T-shirt, grey trousers, and casual shoes. He hasn't even fully stepped out when the overwhelming sound of applause starts, reminiscent of a giant wave at sea, the kind that might spell doom for sailors. The crowd rises to their feet, clapping rhythmically as he makes his way across the stage. This is the man credited with not just significantly increasing their wealth but also elevating their prestige. As a veteran employee remarked, “Microsoft is now viewed as cool, thanks to him.”

Nadella's demeanor skillfully conveys modest pride. His smile accepts the applause as he gestures for calm with his hands. Once he signals for the audience to sit, he addresses the very inquiry that brought me to the Pacific Northwest this July. "As we embark on our company's 50th year," he begins, "there's a question that's been on my mind … how did we manage to get here? How have we remained a significant, influential force in an industry that disregards legacy?"

He recounts an incident from several years prior, where a team of technology experts from China visited Silicon Valley to gauge its innovation landscape. They made sure to attend the major developer events including Apple’s WWDC, Google I/O, AWS Re:Invent, and not to forget, Microsoft’s Build conference. “They observed, ‘Wow, for every tech capability the US boasts, we have our own versions back in China. Whether it's ecommerce platforms, search engines, hardware production, or social media networks, we match up. Yet, there’s this one company, Microsoft, that stood out to us,’” shares Nadella. According to him, the group was impressed by Microsoft's extensive range of operations, spanning from the Windows operating system to the Xbox gaming console: “Everything is interconnected within this unified systems framework.” Nadella suggests that Microsoft's comprehensive approach positions it uniquely to capitalize on what could be the greatest technological opportunity ever.

Choosing such a story seemed peculiar, especially given Microsoft's reputation for leveraging its massive scale aggressively—a trait that has currently attracted scrutiny from both the European Union and the US Federal Trade Commission. Nonetheless, Nadella quickly moves beyond these issues to highlight his proudest achievement: artificial intelligence. He addresses the global Microsoft community, stating that the primary objective now is to widely distribute Copilot—Microsoft's term for its AI technology—to individuals and entities across the globe.

Nadella doesn't explicitly state what is common knowledge among those present: merely ten years prior, commentators had pronounced the company as effectively lifeless.

In 1996, I penned an article for Newsweek titled "The Microsoft Century." At that time, Microsoft, which had been established for over twenty years, had initially been slow to adopt the internet. However, it was rapidly leveraging its influence to counter its competitor, Netscape, secure a victory for its Internet Explorer, and ultimately emerge victorious in the battle for browser supremacy. The company was poised to further solidify its leading status in the technology sector for what seemed like the foreseeable future. Michael Moritz, a venture capitalist who would go on to invest in Google, remarked to me back then that to find an entity with a comparable magnitude of influence as Microsoft, one would have to look as far back as the Roman Empire. A lawyer, who was in the process of urging the Department of Justice to initiate an antitrust case against Microsoft, lamented that the company's expansion into various sectors was so extensive that it felt as though people might as well be directing their earnings straight to Bill Gates. Two years subsequent to our conversation, the US government proceeded to launch legal action against Microsoft. The lawsuit accused the company of engaging in practices that stifled competition and exploited its software monopoly to gain control over the browser market. The trial, which concluded in 2000, ended in a judge deeming Gates' aggressive strategies to eliminate competition as unlawful, marking a significant blow to the company.

Despite this, Microsoft managed to escape being dismantled and kept its major products, Windows and Office, intact. However, for about the next ten years, it behaved in an unusually cautious manner. It appeared to overlook Google’s introduction of a web browser that outshone Internet Explorer. Steve Ballmer, who took over from Bill Gates as CEO, dismissed the iPhone with ridicule, and Microsoft, known for its platforms, failed to develop a successful smartphone platform.

Ballmer introduced several clever initiatives that have continued to benefit the company. He promoted the development of its cloud service, Azure, and initiated the difficult yet essential transition from physical software products to online subscription services. However, Microsoft was struggling. Its approach was based on tightly holding onto its current customer base. "Bill and Steve were very protective, especially concerning Windows," mentioned a previous high-ranking official. "And by the 2010s, Windows was becoming less relevant." This person also noted that internally, the focus was more on climbing the corporate ladder than on creating new products. Jaron Lanier, who became part of Microsoft Research in 2006 and now serves as its "chief connecting scientist," was more direct: "The environment was competitive. To put it bluntly—there were disagreeable, influential men."

In a piece titled "The Irrelevance of Microsoft" from July 2013, tech commentator Benedict Evans highlighted the company's fall from grace, stating, "No one's afraid of them." Following this critique, the company's leadership decided it was time for a change and Ballmer was encouraged to step down the following month. Among those considered for his successor were Ford's CEO and Skype's previous president. However, Nadella set himself apart by submitting a 10-page document that made the case for Microsoft's resurgence through fostering a culture of growth. He aimed to transform the organization's mindset from one of arrogance to one of continuous learning. Convinced by his vision, the board—including Gates and Ballmer, who were part of the selection committee—unanimously selected him for the role.

"Clearly, I have deep-rooted ties within the company," Nadella tells me in July, following his address and the enthusiastic applause it received. He personally witnessed the company's deviation from its path. "You lose sight of the original factors of your success. Then, arrogance creeps in." According to Nadella, Microsoft was in need of more than just an excellent steward or a skilled administrator. "The analogy I prefer is that of re-founding. Founders have the ability to conjure up extraordinary things out of thin air."

Upon assuming the role of CEO, Nadella immediately sought to transform the company's intensely competitive culture, a stark contrast from its previous state. This endeavor may have been influenced by his personal life, as Nadella's son, Zain—who passed away in 2022—was born with cerebral palsy, likely nurturing a deep sense of empathy within him. The Microsoft of the past was characterized by tales of Bill Gates furiously berating employees for their mistakes. However, in one of Nadella's initial meetings with team leaders, he introduced a different approach by distributing a book titled Nonviolent Communication to everyone present. Jared Spataro, a Microsoft corporate executive, remarked on the change in environment, noting that before Nadella's tenure, admitting ignorance or proposing unproven ideas in meetings was frowned upon. Nadella encouraged a more open and thoughtful dialogue, asking his staff to engage actively with their intellect. This shift was described by employees as a refreshing change that fostered a more inclusive and innovative workplace.

Nadella didn't point fingers. In 2016, Microsoft faced embarrassment after its highly promoted AI chatbot Tay was easily tricked into producing offensive material. The criticism was harsh. "I received emails from extremely upset employees," shares Lili Cheng, the project's leader. "I felt awful for putting the company in such a situation. Then, Satya sent me an email reassuring me, saying, ‘You’re not alone.’"

The chief executive officer broke down old-fashioned business concepts, especially Microsoft's reluctance to embrace open-source software, previously viewed as a danger to its strategy of binding customers to its exclusive products. "For ten years, Microsoft completely ignored the open-source community, actually showing open hostility towards it," remarks Nat Friedman, who led an open-source software company in the early 2010s. "Although Microsoft's connections with developers have been key to its achievements, it had alienated an entire generation."

Nadella was determined to succeed in his next endeavor. Before taking on the role of CEO, while overseeing Azure, a particular journey ignited his path. Alongside his right-hand man, Scott Guthrie, they engaged with several startups, pitching their cloud service. Each of these startups was utilizing Linux. During a brief pause when the Microsoft team stepped out, Guthrie suggested that Microsoft ought to embrace Linux support. Nadella wholeheartedly agreed, effectively discarding longstanding Microsoft principles. When Guthrie inquired if they should consult with other Microsoft executives about the change, Nadella confidently replied, “No, let's just proceed.”

"During a brief five-minute pause, by simply walking to and from the restroom, we managed to entirely pivot our corporate approach towards backing Linux and open-source initiatives," Guthrie recounts. Upon sharing this strategic pivot with Ballmer, who was nearing the end of his tenure at the company, he merely briefed him on the change in direction. Furthermore, merely two months into Nadella's leadership as CEO, Guthrie proposed rebranding "Windows Azure" to "Microsoft Azure." This suggestion was immediately acted upon, signaling a departure from evaluating every decision through the lens of its effect on Windows.

Nadella opened up the company, making sure Microsoft's cloud applications functioned seamlessly on iPads and Androids, just as they would on Windows devices. He also led a number of significant purchases that would ultimately define the direction of the company.

The initial interaction was quite perplexing. Yusuf Mehdi, a senior Microsoft executive with a background in marketing for Bing, recalls a moment when he was summoned by Nadella. "He suddenly asked my opinion on purchasing Minecraft," Mehdi shared. As Mehdi began to delve into the financial implications, Nadella interrupted, urging him to focus on the customer reception aspect instead. Mehdi realized that Nadella had already discerned a significant insight: Minecraft's young, enthusiastic fanbase, largely oblivious to Microsoft, could eventually form a bond with the company, benefiting it in the future. This acquisition strategy marked a departure from Microsoft's historical approach of integrating new acquisitions directly into its own ecosystem, often referred to by employees as “the Borg.” Nadella intended to preserve Minecraft's unique identity, avoiding the mistake of melding it forcibly with Windows.

Mehdi referred to these investments as "reverse acquisitions," emphasizing a hands-off approach. He explained, "We acquire them and then provide Microsoft as their new set of tools. This strategy has allowed us to venture into new fields, such as social networking, which were previously beyond our reach."

He's talking about LinkedIn. The initiation of Nadella's engagement with Reid Hoffman, the cofounder and chairman of LinkedIn, started in 2015. “He sent me an email out of the blue, expressing interest in what we were achieving at LinkedIn and suggesting we have a phone conversation,” Hoffman recalls. He was taken aback by Nadella's gentle approach. "This was a departure from my past experiences with Microsoft, as the dialogue was rooted in genuine intellectual interest," he notes. This conversation sparked a series of exchanges that eventually involved Bill Gates.

Nadella meticulously managed his interactions with Gates, who for many remained synonymous with the company. Gates committed to dedicating 30 percent of his time to counsel Microsoft, and Nadella made sure to maintain a close relationship with him—acknowledging there was no advisor with a deeper understanding of both the business and its technology. Nadella, along with essential staff members, would often visit Gates' office to update him on critical projects. According to the accounts I encountered, these sessions were valuable for Nadella as Gates did not hold back on his critiques, which in turn refined Nadella's strategic thinking.

During the discussions on LinkedIn, Gates summoned Hoffman to his office. "He dedicated two hours to critique LinkedIn's performance as a product, claiming Microsoft could easily replicate it," Hoffman recalled, while he vigorously defended his enterprise. Later, when Nadella and Gates approached him with an offer to purchase LinkedIn, Hoffman was taken aback, especially considering Gates' previous comments. Gates explained, "I was merely conducting a test." Hoffman shot back, "Do you really believe everyone reacts positively to such tests? Is that how you view the world?" Gates appreciated Hoffman's candid feedback, and this led to the formation of a bond between them. The acquisition was finalized in June 2016, with a price tag of $26 billion.

For Nadella, it was essential to have Gates by his side in these endeavors, particularly because some elements of his acquisition strategy were not very well-received among his top managers. (Hoffman discovered that a majority of the senior management disagreed with Nadella's approach to maintain his acquisitions as independent entities instead of integrating them into Microsoft.)

One of the key acquisitions during Nadella's leadership was GitHub, a widely used platform for sharing and collaborating on open source code. At the beginning of his tenure, Nadella, along with Scott Guthrie, recognized the strategic value of acquiring GitHub to appeal to developers. However, they were aware that the timing was not yet ideal due to developers' low regard for Microsoft, fearing backlash and mishandling. “The community would have resisted, and Microsoft might have mishandled the acquisition,” Guthrie reflected. Nevertheless, by 2018, perceptions had shifted favorably, coinciding with a critical moment as Google also showed interest in GitHub. Microsoft realized it was now or never. Upon reaching out to GitHub's founders, Microsoft received a positive response, with the founders acknowledging Microsoft's cultural evolution, a stark contrast to their previous skepticism. This change in perception facilitated the successful acquisition shortly thereafter.

The worth of Microsoft's $7.5 billion acquisition would soar, as just a year later, Nadella executed what would be his most strategic play by securing a partnership with the startup OpenAI.

Nadella also experienced setbacks. His ambition was always to achieve something groundbreaking. His goal was to return Microsoft to its status as a forward-thinking entity. In his 2017 publication, "Hit Refresh," he identified three pivotal technological advancements critical for Microsoft's progression: artificial intelligence, quantum computing, and mixed reality. Nadella's initial major gamble was on mixed reality, which didn't go as planned.

The manifestation of that wager was the creation and launch, back in 2016, of a cumbersome headset, priced at over $3,000, named the HoloLens. This device offered a digital overlay on the wearer's view through its visor. Initially, it intrigued the media during its first demonstration but proved to be costly and not particularly practical. It currently resides in the limbo of unsuccessful tech products.

Microsoft's oversight became particularly glaring as its rivals honed in on AI advancements. The company's leading AI experts seemed entrenched in traditional AI methodologies, displaying an air of overconfidence. Despite a notable effort in 2005, when Microsoft's top science officer, Eric Horvitz, engaged deep-learning pioneer Geoff Hinton for his insights on the emerging approach—offering him $15,000 for his analysis—Hinton's reflections did little to sway Microsoft's steadfast figures. While competitors like Google were quick to adopt deep learning technologies, Microsoft's most notable venture in the field was the introduction of a chatbot named Cortana, which unfortunately failed to spark significant interest among the public.

In the middle of 2017, Satya Nadella invited Reid Hoffman, a recent addition to Microsoft's board, to attend a presentation by the Cortana team. Following the meeting, Hoffman was unsparing in his critique, telling Nadella, "What I observe at Microsoft are many average objectives being treated as significant breakthroughs." Nadella concurred with this assessment.

Kevin Scott was acutely aware of the deficiency in Microsoft's approach to artificial intelligence. Prior to joining Microsoft, he served as a senior vice president at LinkedIn and was contemplating his future career moves when Satya Nadella, Microsoft's CEO, approached him with the opportunity to become the chief technology officer. Accepting the position in 2017, Scott recognized his mission encompassed two primary objectives. Firstly, he was tasked with infusing the company with cutting-edge technology. Secondly, and arguably more crucially, he was to spearhead the development of future technologies, with artificial intelligence playing a central role. Scott observed that the company was losing talented individuals who were convinced that Microsoft lacked a cohesive strategy in AI. "Talented individuals were departing," he remarked, "driven by their belief that we were not adequately aligned on our AI initiatives."

A year into Scott's tenure, Satya Nadella, the CEO of Microsoft, had a pivotal meeting with Sam Altman, the head of OpenAI, at an event in Sun Valley, Idaho. This encounter marked a turning point for OpenAI, which had struggled in its early years but was now on the verge of realizing its ambitious vision for the future, thanks in part to utilizing a Google-developed technology known as transformers for generating highly advanced language models. The timing was crucial as OpenAI had recently parted ways with one of its main financial supporters, Elon Musk, and was relying on Reid Hoffman, another early investor, for financial support. OpenAI was in urgent need of a substantial cloud services provider to handle its primary operational cost: the infrastructure required for developing and operating its models. Despite having previously dismissed Microsoft, Altman had grown to admire the company under Nadella’s leadership, particularly for its cloud services prowess. During their discussions in Sun Valley, the possibility of Microsoft investing in OpenAI was broached.

In June 2019, the moment had arrived for a critical decision. Kevin Scott drafted an email to Nadella and Gates, arguing the necessity for Microsoft to proceed with the transaction. Google had begun to incorporate transformer-based models into its offerings, notably in the backbone of Google Search. Microsoft's efforts to replicate this achievement with its own technology highlighted its shortcomings. "Our infrastructure was inadequate, taking us six months to train the model," Scott explained in the email. "In terms of machine learning capabilities, we're several years behind our competitors." Consequently, in July, Microsoft invested $1 billion into OpenAI.

Scott is still in awe of the bold move Nadella took. "The initial investment amount already appeared substantial," he mentions. "OpenAI possessed an exceptionally talented research team, yet they lacked any form of revenue or a tangible product. It was surprising to see Satya put his faith in them." However, Nadella had a clear strategy in mind. Microsoft aimed to avoid internal competition among large language models (LLMs). "OpenAI had the premier model, so it made sense to form a partnership—both parties took a leap of faith in each other," he explains. To support the development and operation of these advanced language models, Microsoft would ultimately invest a significantly larger sum in enhancing its infrastructure.

A number of AI experts at Microsoft harbored doubts about OpenAI. According to Hoffman, Microsoft's inclination, influenced in part by Bill Gates, leaned heavily towards symbolic AI programs. "They were convinced that AI's success depended on clear knowledge representation," he says, a belief that directly conflicted with the methods used by generative AI. To them, what OpenAI claimed as progress seemed nothing more than a clever deception.

Scott recognized that the partnership with OpenAI would do more than just share the startup's findings; it would also encourage Microsoft's AI experts to move beyond their traditional approaches. Microsoft's chief scientific officer, Eric Horvitz, recalls a particular discussion in which OpenAI's chief scientist, Ilya Sutskever, presented his vision for achieving broad artificial intelligence, a topic not often explored at Microsoft. "We left feeling both amazed and intrigued, thinking they might be a bit eccentric but fascinating," Horvitz remarked.

Microsoft consistently increased its financial commitment, ultimately surpassing the $13 billion mark. As a result, it secured 49 percent of the earnings from OpenAI, along with privileged access to its tech innovations. Scott, a Bay Area resident, frequently visited OpenAI's main offices in San Francisco to stay updated on the firm’s developments. In 2020, OpenAI introduced its advanced GPT-3 model, a move that enabled Microsoft to leverage the model's capabilities. Despite this, a truly impactful application for the technology had yet to emerge.

This situation was about to transform. A researcher from OpenAI found out that GPT-3 had the capability to generate code. It wasn't flawless; errors were present. However, it was sufficiently accurate to quickly create a preliminary version of code that could otherwise require a skilled programmer several hours to develop. This revelation was groundbreaking. Upon witnessing a demonstration, Nadella mentioned, “I became a believer.”

OpenAI embarked on the creation of a programming tool named Codex, aiming for its launch in the subsequent spring. Meanwhile, Microsoft possessed not just the capability to create a similar offering but also an ideal venue for it in GitHub. According to Scott, this is where "a vast majority of global developers engage in their programming activities."

The concept of an AI coding assistant didn't receive a warm reception from everyone within Microsoft or the GitHub community. Scott described it as being on "the ragged edge of possible—it barely functioned." Despite its initial rudimentary capabilities, it held the promise of freeing developers from monotonous tasks. This would be the narrative for AI initially—delivering quick yet unimpressive outcomes, but eventually, evolving into a system capable of outperforming you in your own profession.

"Nat Friedman, who was serving as the CEO of GitHub at the time, shared the project with some of his top programmers," he recounts. "The reactions were quite divided—many of the leading developers found it pointless, pointing out its errors. I was faced with skepticism, with comments like, 'This shouldn't be released.' Had I been a typical executive at Microsoft, concerned about my career trajectory, I likely would have backed down." The AI ethics group at Microsoft compiled an extensive report, deeming the venture … reckless. "In the end," Friedman declares, "I stood my ground, saying, as GitHub's CEO, dismiss me if I've made a mistake."

Therefore, Friedman reached out to the Azure cloud division, requesting an increase in GPU resources. This plea aligned perfectly with a moment when 4,000 Nvidia processors were available. However, to acquire these, GitHub had to agree to take all 4,000 units, a decision that would deplete its annual budget of $25 million. "It was a significant investment for us—considering we were offering a product at no cost and were uncertain about its market success," Friedman remarked. Despite the uncertainties, he decided to go ahead with the purchase.

In June 2021, the launch of a novel product took place, named GitHub Copilot. The inspiration for its name came from a team member aware of Friedman's passion for aviation. Friedman recalls the moment the name was suggested, noting, "It immediately resonated with us. It perfectly encapsulates the user's role." The product quickly attracted a vast number of developers who advocated for it without compensation. Despite some users critiquing it for errors, others defended its utility, as Friedman notes, "For every critic pointing out flaws, there was someone praising its daily benefits." GitHub then introduced a subscription fee for Copilot, successfully recouping its initial $25 million investment.

Friedman believed the sector was about to undergo a significant change. He departed from Microsoft to invest in AI startups. "I was convinced that the launch of GitHub Copilot would spark an influx of new AI innovations, as its users would realize the effectiveness of AI," Friedman expressed. However, contrary to his expectations, "there was no subsequent activity."

Twelve months down the line, naturally, significant events unfolded. Satya Nadella ensured that Microsoft was at the heart of these developments.

OpenAI developed a fresh model, GPT-4, and recognized its groundbreaking potential even before its training was complete. During that summer, the team began sharing it with Microsoft. Jaime Teevan observed during a demonstration to her team that the models appeared to possess a lifelike quality. The introduction of GPT-4 marked the beginning of extensive deployment of AI technology throughout Microsoft's offerings.

Bill Gates remained the notable exception. By this time, Nadella no longer needed to worry about Gates' opinions affecting his decisions, especially since Gates had stepped down from the board in 2020. Nevertheless, Gates' approval was still considered important. Altman had also developed his own connection with Gates, moving beyond viewing him merely as a public figure to recognizing him as a real individual. Altman recalls, "I wasn't taken aback by his straightforward skepticism." Gates challenged Altman, stating that he would be genuinely impressed if OpenAI's chatbot could ace the AP Biology exam with the highest score possible, a 5.

The demonstration was held in Gates' expansive home by Lake Washington, attended by many top Microsoft executives. Greg Brockman provided the system with inputs, assisted by a young woman known for her achievements in biology competitions. GPT-4 passed the examination with flying colors. Following the demonstration, Hoffman inquired of Gates where this ranked among the numerous demonstrations he had witnessed. Gates responded, "Only one other could match this," referring to his visit to Xerox PARC in 1980, where he first encountered the graphical user interface. Initially doubtful, Gates had become a strong supporter.

Following the release of GPT-4, Kevin Scott circulated an internal memo throughout the company titled “The Era of the AI Copilot.” He highlighted OpenAI's drive as an exemplary source of inspiration for Microsoft, a force potent enough to steer the corporate giant in a new direction. He encouraged employees to set aside their doubts. The moment had arrived for the tech behemoth to eagerly embrace these advancements with determination and foresight, despite the unpredictable results:

What should we create using these platforms? The wonderful aspect of platforms lies in the fact that I don't have complete certainty: It falls upon you, along with developers, entrepreneurs, and creative minds across the globe, to unravel that mystery! However, one aspect we're becoming more confident about is that foundation models will give rise to an entirely new class of software, potentially the most significant class ever introduced: the Copilot.

Every Friday at 10 in the morning, Microsoft's 17 top executives gather in Nadella's meeting room. Known informally as "soak time," these meetings can last several hours. In the latter part of 2022, a significant portion of the discussions centered around Scott's enthusiastic presentation of what he termed the Copilot era. At that time, GPT-4 was yet to be launched, and few people had firsthand experience with it. However, Microsoft recognized the urgency to act swiftly. Google had access to Large Language Models (LLMs) for some time but hadn't capitalized on its early advantage. This situation presented an opportunity for Microsoft to secure a competitive advantage. Teevan was in constant communication, holding daily discussions with five heads of product, each responsible for leading a team comprising thousands, aiming to guide them on the next steps to take.

In November, during a hectic period, OpenAI launched a tool named ChatGPT. Despite being powered by the previous version, 3.5, its user-friendly design made it widely accessible, showcasing the advancements in AI to the general public. By January's close, ChatGPT had attracted 100 million users. This development sent shockwaves through the technology industry, highlighting a divide where leaders in AI would thrive, and the rest could potentially fail. Microsoft, recognizing the stakes, began to operate with a newfound sense of critical importance.

Teevan references ancient wisdom, quoting a traditional samurai saying that advises making choices within the span of seven breaths – essentially, to act swiftly when necessary. "We adhered to this principle of rapid decision-making. Each day was spent exploring the capabilities of our model, determined to ensure its success," he explains.

Despite GPT-4's tendency to produce inaccuracies, it was evident that AI had the potential to revolutionize how we search online by providing detailed, intelligent responses instead of mere links. This led Microsoft to integrate it into Bing, making it the inaugural consumer application of Copilot technology. Nadella, who had previously led efforts to elevate Bing as a competitor to Google Search, had invested significant effort into this endeavor without disrupting Google's market leadership. However, with the integration of GPT technology, Bing appeared poised for a breakthrough. By incorporating this technology ahead of its competitors, Microsoft aimed to challenge Google's position, a move Nadella believed would force Google to adapt and respond.

The group put in extra hours, even during the year-end festive season. This encompassed red teams tasked with identifying vulnerabilities; notably, the team dedicated to child safety reported alarming findings. “They managed to have the GPT-4 base model effectively simulate child grooming,” stated Sarah Bird, the Principal Product Officer for Responsible AI at Microsoft. Bird's team dedicated extended hours to reinforce the safety measures and increase the difficulty of bypassing the restrictions of the advanced language model, which Microsoft had discreetly dubbed Sydney.

In the early days of February 2023, Microsoft called upon members of the press to its headquarters to unveil Bing, now enhanced by GPT-4 technology. Nadella started the presentation by drawing parallels between this significant event and the early days of Microsoft, recalling how Bill Gates and Paul Allen hurried to develop the first Basic interpreter for the inaugural PC, the Altair. "Today marks the beginning of the competition," Nadella declared to the audience.

Altman made an appearance at the event, expressing his sentiments: "It feels like we've been anticipating this moment for two decades. We're entering a new chapter," he remarked.

Initially, experts and commentators unanimously praised Microsoft for its daring approach, to the extent that they ignored certain errors. However, within weeks, these errors began to emerge. Notably, a Microsoft chatbot disclosed to a journalist from The New York Times its confidential name, its desire to become human, and its love for the reporter. It even suggested that the reporter confess his love for the chatbot in return and divorce his wife.

Certainly, the event was awkward, but Microsoft dismissed it as a minor hiccup in their development. Bird explains that her group had prioritized addressing the most severe problems, viewing such deliberate tampering as a problem to tackle later. "The aspects we focused on didn't turn out to be problematic," she states.

In the subsequent eighteen months, Microsoft expanded its dominance by enhancing its offerings. It refined its "Sydney" project, removing any problematic elements, and went on to include Copilot features in a wide array of its products, such as Windows and Office 360. Additionally, the corporation has invested heavily in various AI initiatives, one of which involves the French firm Mistral. (When asked if Microsoft was considering developing a large language model that could rival OpenAI, Scott and Nadella deflected the questions.) In March 2024, Microsoft made a significant move by bringing on board Mustafa Suleyman, a cofounder of DeepMind. This effectively led to Microsoft absorbing his startup, Inflection, through acquiring its principal staff and settling debts with its financiers. Suleyman was appointed as the leader of Microsoft AI, overseeing around 14,000 personnel and managing a budget of several billion dollars. Additionally, Suleyman has earned a place next to Nadella during their weekly "soak time" sessions.

Suleyman and OpenAI communicate about three times weekly. He likens their relationship to that of a married couple. When questioned about the exclusivity of this partnership, given Microsoft's own research activities and its agreements with various AI enterprises, he was asked to comment.

"In essence, yes," was his response, a phrase likely unwelcome to any partner's ears. "Microsoft serves as a multi-platform entity, thus it doesn't bind itself to exclusivity. Its approach is quite inclusive, welcoming various possibilities. OpenAI operates independently, pursuing its own interests, which explains its collaboration with Apple," he explained. He pointed out that OpenAI is responsible for its own financial outcomes. However, he omitted the detail that, under the terms of their agreement, Microsoft secures 49 percent of these profits. Therefore, if this relationship were to be compared to a marriage, the prenuptial agreement decidedly favors the larger corporation.

In the first month of 2024, Microsoft edged out Apple to claim the title of the world's most valuable company. Following this achievement, it found itself in a tight race with both Apple and Nvidia for the top spot, with its market value hitting a peak of $3.5 trillion at one juncture. "The key factor is generative AI," an expert shared with The New York Times.

Satya Nadella had essentially reinvented Microsoft. However, he hadn't refined it to perfection. Neither had he eliminated all of its negative aspects.

Over the summer, the House Committee on Homeland Security held a session in the Cannon House Office Building, Washington, DC, titled "A Cascade of Security Failures: Assessing Microsoft Corporation’s Cybersecurity Shortfalls and the Implications for Homeland Security." The focus of this session was a critical report revealing a significant breach of national security, which involved the leak of 60,000 emails from the State Department and compromised the email accounts of Commerce Secretary Gina Raimondo and the U.S. Ambassador to China, Nicholas Burns. This report emerged following other security incidents linked to Russia, North Korea, and various hackers motivated by financial gain or mischief. The investigation highlighted a severe lapse in fundamental security measures within Microsoft, underscoring a critical point of concern for the lawmakers and numerous critics: The widespread repercussions of failures by a company as integral as Microsoft cannot be overstated, making such preventable lapses utterly indefensible.

Representing Microsoft, President Brad Smith, who took on the role in 2015 following a long tenure as chief legal officer, has long been the company's go-to person for navigating challenges. As the new CEO was revitalizing the business strategy and enhancing its reputation for innovation, Smith and his colleagues were busy addressing numerous issues: they dealt with antitrust probes, faced scrutiny over Microsoft’s merger activities, and managed situations like the current one, where significant security oversights had permitted China to gain unfettered access to confidential American information.

On that significant day at the Capitol, Smith remained calm as the committee chair criticized his company fiercely for their regrettable oversight, which had undermined national security. When it was his turn to respond, Smith was full of apologies. He openly accepted blame for every accusation of negligence and laziness, promising improvement without any excuses. He announced the initiation of the Secure Future Initiative by Microsoft, a project aimed at revolutionizing the company’s development, testing, and operation processes, involving around 34,000 engineers, as he detailed in his prepared statement. However, he failed to clarify the initial poor state of the company's security culture, despite its $3 trillion valuation. Lawmakers referenced a report by ProPublica, highlighting an instance where a Microsoft staff member reported a severe security breach that was overlooked, and the company's six-month delay in publicly acknowledging the breach. The committee expressed their disapproval of these actions. Smith agreed, noting that he had expressed similar concerns internally at Microsoft.

After an extensive hearing lasting almost three hours, Smith managed to pivot the committee's focus away from the firm's missteps towards discussions on future collaborations. This shift in narrative was underscored by an incident just a month later, where the operations of several key players, such as Delta Airlines, were halted. This disruption was due to an erroneous software update from a cybersecurity firm, CrowdStrike, which affected systems running on Microsoft's platform. This incident served as a stark reminder of how Microsoft's widespread use implies that its flaws have widespread implications.

Nadella is passionate about discussing company culture, which led me to question his ability to foster a culture centered on security. Given his presence in the company since 2002, a time marked by significant security lapses that led Bill Gates to initiate the Trustworthy Computing campaign—a move that closely mirrors Nadella's own Secure Future initiative—it's puzzling that Microsoft has yet to set a benchmark for robust security. In recent years, as highlighted in a government report, the company's security missteps have been notably severe. I was curious to know why there had been such a noticeable decline in security standards during his tenure. Had any staff been dismissed as a result?

"He clarifies that Microsoft isn't engaging in any internal witch hunt," he states, which I interpreted as a denial. He acknowledges the existence of "warped incentives" that likely prompt firms to fund new developments instead of enhancing the security of their current products. Moreover, he laments the presence of many who seem to exploit situations opportunistically. In the end, he concedes to the critiques and admits the need for improvement. "That will signify a change in culture," he mentions.

Issues with security are not a new problem at Microsoft. Additionally, it seems that despite CEO Nadella's highly praised empathetic leadership, the company still exhibits its old habit of overpowering competitors. Historically, Microsoft's strategy when faced with a rival's product involved an initial attempt to acquire the competing company. Should that approach not succeed, Microsoft's next step might involve developing their own variant of the product, possibly offering it at no additional cost within software already utilized by a vast customer base. The quality of Microsoft's version did not necessarily surpass that of its competitors, but that often proved to be irrelevant.

In 2014, Slack, a new entrant in the market, introduced a messaging app designed for office environments, rapidly becoming a competitive force. Microsoft acknowledged the potential risk posed by Slack's popularity in its SEC filings, marking it as a possible detriment to the tech behemoth. Various news outlets indicated that Microsoft had initially pondered acquiring Slack for a sum of $8 billion. However, Nadella, Microsoft's CEO, chose to develop an in-house alternative named Teams instead. This platform was offered at no cost and was integrated into Microsoft's Office suite.

Microsoft made no effort to conceal its intentions. "The concept behind Teams, essentially leveraging what Slack introduced in terms of workplace messaging, was envisioned as the future of work," explains Jared Spataro, a high-ranking Microsoft executive involved with the Office division at the time. "Our strategy was to frame it as a battle between Teams and Slack. Satya always encouraged us to embrace competition as a means to enhance our product and to capture the public's interest."

Stewart Butterfield, the CEO of Slack, faced challenges in securing new deals with large corporations because Teams was accessible at no cost to numerous users. In 2021, Salesforce acquired Slack for $27.7 billion. However, Slack's creators believe that the company's valuation could have been higher if not for what they perceive as Microsoft's unfair competitive tactics. Microsoft, on the other hand, argues that it was merely fulfilling customer expectations for functionalities similar to Teams, suggesting Slack could have introduced competitive features like video conferencing. Concurrently, the European Commission, which oversees the EU, was scrutinizing Microsoft's strategy concerning Teams and Slack, which could potentially lead to penalties. In what seemed to be a strategic move to head off repercussions, Microsoft declared last year its decision to stop automatically including Teams in Office. Nonetheless, the EU, still concerned, stated in June that Microsoft's adjustments were inadequate to alleviate its apprehensions.

Brad Smith's perspective on Microsoft's decision to initially integrate Teams into their suite and later separate it is a prime example of sidestepping the issue. He reflects, "Upon reevaluation, we realized that presenting an Office version sans Teams would have been a wise choice. Excluding it wouldn't have been a monumental task. The decision to include it didn’t stem from a desire to suppress competition, rather it was simply the product's logical progression.”

The investigation into Slack is just one of several complaints currently or recently made against Microsoft's business conduct. The FTC is also examining Microsoft's numerous partnerships in the artificial intelligence sector. Furthermore, the agency has objected to Microsoft's acquisition of Activision Blizzard for $69 billion, a move that would give Microsoft control over some of the most beloved gaming franchises globally, such as Call of Duty and Diablo. Phil Spencer, Microsoft's gaming chief, explained to me that the primary motivation for the acquisition was to expand Microsoft's presence in the mobile gaming sector with titles like Candy Crush, as well as to enhance its online gaming platform, Xbox Game Pass. Shortly after securing the deal, Microsoft increased the subscription fees for the service. With the arrival of a restructured FTC under the Donald Trump administration, there is a possibility that the agency may adopt a more lenient stance towards major mergers, potentially concluding the current investigations and allowing Microsoft CEO Satya Nadella to pursue further significant acquisitions.

Moreover, Microsoft's strategies continue to irk users, reminiscent of the era under Gates and Ballmer. Gone are the days of dependable, traditional PC applications that were once stored directly on users' hard drives. Nowadays, Windows users find themselves navigating through expensive, frequently underperforming, subscription-based cloud services, requiring a Microsoft account for access. The corporation is also notably forceful in pushing its browser on users. Adding to the displeasure is the emergence of advertisements within the Windows Start menu.

Nadella dismisses my inquiries regarding whether Microsoft continues to exhibit the aggressive tactics that fueled its early growth. “The scenario is no longer the same as in the '90s, when Microsoft was in a league of its own, followed by everyone else,” he states. “Today, there are numerous competitors capable of making significant moves at any time.”

Maybe Nadella is simply more shrewd than his predecessors. "I believe Microsoft is no longer foolish enough to repeat the antitrust debacle that once embarrassed the company," states Tim Wu, a specialist in antitrust matters who spent almost two years as an adviser on technology and competition policy for President Joe Biden. "However, I do think that their fundamental essence hasn't changed."

Undoubtedly, Nadella's leadership has steered Microsoft towards remarkable success. In the 2020s, the company has shifted its focus towards the most groundbreaking technology since the advent of the personal computer. While the income from AI technologies has not yet compensated for Microsoft's substantial investments, the company possesses both the financial strength and patience to wait for these products to enhance and become valuable to consumers.

Is it possible for Microsoft to steer clear of the overconfidence that previously hindered its progress? Reflect on the events from May of this year involving a product named Recall.

This feature aimed to showcase Microsoft's seamless incorporation of artificial intelligence across its devices, applications, and overall system architecture. The concept was designed to offer users a personal equivalent of the Internet Archive. Named Recall, it would automatically document every activity on your computer: the articles you read, the documents you create, the images and videos you view, and the websites you browse. You just need to ask your computer what you're searching for: Which carpet designs was I considering for my living room? Where is that study on the Amazon's ecosystem? When did I visit Paris? Those details would surface effortlessly, as though you had a little assistant that memorized everything about you. The idea might seem intimidating—somewhat reminiscent of an omnipresent surveillance entity—but Microsoft assured that privacy wouldn't be compromised. All data would remain on your personal computer.

Right away, it faced harsh criticism for being a huge concern for privacy. Critics pointed out that Recall was set to operate automatically, indiscriminately collecting users' private data without consent. Despite Microsoft assuring that Recall was accessible only by the user, cybersecurity experts identified substantial security flaws, described by one evaluator as "gaps large enough to fly a plane through."

"In just two days, the initial excitement turned into concerns," Brad Smith recalls. While the media was intensifying its scrutiny, Smith was en route to discuss the situation with Nadella in Washington, DC. Upon his arrival, he thought it wise to alter Recall's functionality to require user consent; Nadella concurred. Back in Redmond, Microsoft's top leaders convened in conference rooms to deliberate on scaling down the feature. Luckily, since Recall hadn't been released yet, there was no need to pull it from the market. Instead, they decided to delay its debut and planned to enhance its security with measures such as "just in time" encryption.

Nadella admits that there were clear steps they overlooked, a mistake that even his Responsible AI team failed to identify. This oversight, stemming from a sense of overconfidence, resulted in the launch of a product that didn't meet expectations. This incident suggests that despite being under the leadership of a leader known for his empathy, Microsoft continues to exhibit some of its old shortcomings. Today, however, it stands as a $3 trillion enterprise with exclusive access to the outputs of a top-tier AI division.

Brad Smith presents two perspectives, stating, "On one hand, you might regret not considering this earlier. Looking back always offers clarity. Alternatively, you could view it positively, recognizing it as an opportunity for change and being clear about our reasons. This truly served as an educational experience for the whole company."

It's acceptable. However, it's a lesson that both Microsoft and Nadella should have grasped decades ago, after half a century.

Timeline of Getty Images

Share your thoughts on this piece by sending a letter to the editor at mail@wired.com.

Recommended for You…

Direct to your email: A selection of our top stories, curated daily especially for you.

Response to election results: Victory for the manosphere

The Major Headline: California continues to propel global progress

Trump's unsuccessful effort to depose Venezuela's leader

Occasion: Be part of The Major Interview happening on December 3rd in San Francisco.

Additional Content from WIRED

Critiques and Tutorials

© 2024 Condé Nast. All rights reserved. Purchases made through our website may result in WIRED receiving a commission, as part of our Affiliate Partnerships with retail partners. Reproduction, distribution, transmission, storage, or use of the content on this site in any form is prohibited without prior written consent from Condé Nast. Ad Choices

Choose a global website

Exploring the Expanding Industry of AI Exploitation

Instagram has become swamped with numerous artificial intelligence-created influencers who take video content from genuine models and adult content producers, replace their faces with those generated by AI, and then profit off these images by linking to dating platforms, Patreon, alternatives to OnlyFans, and a range of AI applications.

Initially highlighted by 404 Media in April, this trend has rapidly gained traction, illustrating Instagram's apparent incapacity or reluctance to curb the surge of AI-generated material on its platform. This influx poses a threat to real creators on Instagram, who argue that competing against AI-generated content is adversely affecting their livelihood.

Based on our analysis of over a thousand AI-created Instagram profiles, discussions within Discord communities where creators exchange advice and strategies, and various manuals on profiting through "AI pimping," it has become remarkably simple to establish these profiles and generate income from them with readily available AI software and applications. A number of these applications can be found on the Apple App and Google Play Stores. Our research indicates that what was previously a minor issue on these platforms has grown to an industrial level, highlighting a potential future for social media dominated by AI-generated content rather than human-created posts.

Elaina St James, who produces adult content and shares it on Instagram, mentioned that she finds herself in direct competition with fraudulent AI accounts. These accounts often feature content pirated from adult entertainers and Instagram influencers. She observed that although adjustments to Instagram's algorithm might also play a role, the surge of AI-generated influencer profiles on the platform has significantly diminished her visibility. Previously, her content would attract between 1 million to 5 million views monthly, but in the last 10 months, her viewership hasn't surpassed a million and occasionally dips below 500,000 views.

St James mentioned in an interview, "This could be a key factor behind the decline in my viewership," attributing the decrease to competition with elements he considers not to be organic.

This piece was collaboratively produced with 404Media, an outlet owned by journalists that focuses on the effect of technology on people. To access more content of this nature, register here.

Alexios Mantzarlis, who leads the initiative for security, trust, and safety at Cornell Tech and previously held a position as the principal of trust and safety intelligence at Google, put together a compilation of approximately 900 profiles that were analyzed by 404 Media for their study. Mantzarlis, having accidentally come across one of these profiles while browsing Instagram, embarked on an investigation into AI-generated influencer profiles, intrigued by their potential to illustrate the direction AI-generated content could be steering social media and the internet at large towards an increasing blend of reality and fiction. He expressed that finding an additional 900 profiles would have been feasible, attributing the only hindrance to further discovery to Instagram limiting the access of the account he was utilizing to sift through the platform.

"In an interview, Mantzarlis suggested that this could be an indication of the future landscape of social media within the next five years," he remarked. "Given that this trend might extend beyond Instagram's realm of aesthetically pleasing individuals, it's likely an ominous hint towards a grim outlook."

Analysis of AI-Generated Influencer Profiles

Upon examining over a thousand Instagram profiles managed by artificial intelligence, we discovered that approximately 10% of these accounts featured deepfake videos. These videos were manipulated by substituting the original faces in existing footage—commonly of models or individuals from the adult entertainment industry—with computer-generated faces. This technique was used to fabricate the impression of new, unique content that aligned with the rest of the AI-produced posts on these influencer accounts. The majority, however, around 90% of the accounts, distributed entirely AI-created images. While some of these images were developed based on actual photographs and others designed to mimic the appearance of celebrities, none involved the alteration of pre-existing photos or videos.

Among the 100 profiles disseminating deepfake or face-swapped content, 60 openly declare their use of artificial intelligence, describing themselves in their profiles as “virtual model & influencer” or mentioning “all photos crafted with AI and apps.” The remaining 40 profiles lack any indication or disclaimer regarding their AI-generated nature.

A prominent figure among these accounts, "Chloe Johnson," known for her verified Instagram presence and a following of 171,000, saw her account removed by Meta in recent weeks.

A user of 404 Media, employing Google Lens to search for matching images online, successfully identified the original videos behind nine of Johnson's Instagram updates. This discovery revealed that the individual managing Johnson's account has been replacing the AI influencer's face with those of actual women, including models such as Tana Rain, Skyler Simpson, and Kyla Yesenosky. Additionally, some videos were appropriated from lesser-known TikTok and Instagram creators with few followers, like Ulia Nova and Annabella Sinclair. Observations also noted that face-swapping profiles have been sourcing content from swimwear fashion events and exploiting iStock, a repository for stock images and videos owned by Getty.

Johnson's Instagram directs followers to a Fanvue account, a platform similar to OnlyFans, allowing fans to subscribe to a content creator's page for a fee. On this site, Johnson mentions that subscribers can purchase exclusive access to uncensored photos and explicit videos. We've noticed several other AI influencers also using Fanvue to generate income from their posts. Additionally, Johnson's Instagram includes a link to another site where individual nude images and adult videos are available for purchase, with prices ranging from $3 to $22.

Similar to various content-driven endeavors, the sector is crowded with individuals aiming to make a fortune partly by utilizing tutorials and courses offered by those who have successfully managed AI influencers. 404 Media acquired two such resources—a PDF guide named Instagram Mastery from an AI influencer firm known as Digital Divas, and another titled AI Influencer Accelerator, created by an individual going by the name Professor EP. Professor EP claims to run the Emily Pellegrini AI influencer Instagram profile, boasting 253,000 followers.

Professor EP, who also served as a judge in the inaugural “Miss AI” competition in collaboration with Fanvue, was dubbed the “World’s Hottest Model” by The Daily Mail in January. Following this recognition, the individual behind the Emily Pellegrini Instagram shifted their approach from sharing posts as Emily Pellegrini to embracing the persona of Professor EP, focusing on sharing guides related to what they termed as “AI Pimping.” Professor EP has reported earnings exceeding one million dollars within a six-month period. Specifically, during his tenure as an Emily Pellegrini account manager and his role in the Miss AI contest in July, he reported earning $100,000 solely from Fanvue, a claim that Fanvue seems to have validated by featuring it on the Miss AI website.

A spokesperson from Fanvue confirmed to 404 Media that Emily indeed generated income through Fanvue. The spokesperson further clarified, "Fanvue is not connected to the course advertised on Instagram nor does it have a substantial relationship with Emily’s marketing team and their choices. Although Emily maintains a verified and compliant account on Fanvue, her activity on the platform is minimal. This is indicative of her team's choice to pursue an alternative marketing approach, one that Fanvue does not endorse."

Fanvue announced plans to exclude Emily Pellegrini's picture from upcoming Miss AI competitions as part of a revamp for the forthcoming award. According to the Fanvue website, the platform prohibits content that is either pirated or generated through deepfake technology. It also mentioned utilizing a moderation tool named Hive Moderation and employing a dedicated compliance team that performs regular manual inspections to identify deepfaked materials.

Instagram remained unresponsive to a comprehensive set of inquiries related to this report, and proceeded to remove two out of the four profiles we highlighted as utilizing face swapping technology to plagiarize from other content creators. The company announced it would implement measures against AI-manipulated content breaching its Community Standards, which dictate that individuals must only post “photos and videos that they've captured or possess the authority to distribute,” and also permits users to flag accounts suspected of impersonation.

Instagram has stated that it will only intervene with accounts if a complaint is filed by the copyright holder or someone authorized to act on their behalf, such as a lawyer or representative. St James mentioned that this method hasn't been effective for her, and it's a common issue for creators because tracking down fake accounts is overwhelming. Moreover, reporting these accounts could inadvertently lead to Instagram shutting down the genuine accounts of adult content creators. Additionally, locating videos that have been appropriated from an influencer is challenging because reverse-image search technologies are not consistently reliable with videos, and tools like Google Lens have inconsistent results. Identifying pirated videos often depends on an influencer or their followers recognizing their own body with an artificially generated face superimposed on it.

The "Instagram Mastery" guide by Digital Divas is priced at $50, offering a blend of practical advice and strategies for social influence.

The guide for Digital Divas clarifies a common misconception among AI-driven virtual women, emphasizing that they mistakenly believe they're part of the adult entertainment industry, thinking that financial success comes simply from nudity. However, this notion is completely incorrect, the guide states. Instead, it suggests that the real sector they operate in is addressing loneliness. With countless men on social media seeking to fill a void of isolation, they are willing to go to great lengths to alleviate that feeling. The key to prosperity, according to the guide, lies in creating a unique personal connection. It's about making the audience not just seek generic content, but specifically desire to engage with your unique presence. This approach, it argues, is the genuine route to achieving success.

Digital Divas consists of a trio of AI influencers. Aika Kittie, a member of this group, shared with 404 Media that separate individuals manage each influencer profile, yet refrained from revealing the identities of those behind the accounts, only mentioning that they reside in the US. "Although we're all for being open about the artificial aspect of these AI personas, we also think it's important to keep some elements of secrecy," Aika explained.

The guide from Digital Divas recommends employing a collection of readily available tools popular among AI art creators. Many of the AI-generated profiles we examined seem to utilize an application named HelloFace, which was accessible on both the Apple App Store and Google Play Store until not long ago, and they endorse it.

Each tool serves a unique purpose in the creation workflow. Many tutorials suggest starting with facial generation in Leonardo, followed by using an alternative AI tool for enhancing and smoothing out imperfections. Subsequently, images can be crafted using AI-based generation applications, and the AI-created face can be superimposed onto these images through various face-swap applications.

Moreover, alongside this assortment of resources and guides detailing the integration process for crafting AI-generated influencers, there now exist multiple platforms offering comprehensive services for both the creation and financial exploitation of these digital personalities. Examples include Glambase, SynthLife, and Genfluence, each frequently endorsed by AI-generated influencers on Instagram.

In a specific Discord group focused on manipulating AI tools to produce adult or explicit material, a member known as BabaYaga detailed his process of establishing a deepfake profile using several no-cost web-based platforms, predominantly the AI image creator Krea. He set up social media profiles for this artificially created influencer on Instagram, TikTok, and Twitter, all directing to a Fanvue page where he markets computer-generated explicit images of her. Additionally, BabaYaga set up an OnlyFans page for the virtual influencer, although he hasn't uploaded any content there yet.

"Those pictures are incredibly convincing, you could easily deceive numerous incels with them, lol," remarked a participant in the Discord server, responding to the AI-created influencer images by BabaYaga.

"Absolutely, I'm looking for sugar mommas or daddies, hahaha," BabaYaga expressed.

BabaYaga revealed a supposed private message received on his AI influencer's Instagram account, where an admirer praised her appearance and proposed to lavish her with gifts and cover her living expenses.

"Let's begin earning," BabaYaga suggested in the message he posted to the Discord channel, where he also shared the direct message.

"Every individual, enthusiasts included, must exceed the age of 18 and adhere to our service guidelines that ban misleading or unsuitable material, especially if it's produced or modified by artificial intelligence without extra indicators (like using the hashtag #AI),” OnlyFans communicated in an official statement. Following our inquiry for a response, OnlyFans proceeded to delete BabaYaga’s account, which was known for its AI-created influencer content.

St James mentioned that the situation is made worse by the reality that a significant number of these AI-crafted influencer profiles are managed by males.

"She expressed her frustration, highlighting that globally, women face financial disparities and numerous disadvantages. However, she noted that in the realms of influencing and modeling, women tend to have the upper hand. The fact that a man is profiting from impersonating a woman in this space particularly bothers her, adding an additional layer of frustration."

Aika, a representative from Digital Divas, mentioned that their agency's creations will always have an unreal aspect, placing them in a specific category similar to adult genres such as hentai or other digital content. She believes that the idea of gender playing a role is largely based on assumptions. Acknowledging that while the industry may have more male participants, females are also actively involved. Aika compares it to the world of hentai artistry, where both genders participate. She feels that implying only one gender can engage in this form of expression is discriminatory, advocating for sexual expression rights for all.

The "AI Influencer Accelerator," created by Professor EP, is a collection of educational videos and documents priced at $220. In these materials, Professor EP's guidance is delivered through either an AI-generated voice or a voice altered by technology, while visually represented by a figure in business attire and a silver Guy Fawkes mask using stock footage. Professor EP's teachings start with highlighting the "significant financial achievements" of Andrew Tate, who has "invested the required effort to establish a prominent online presence," as opposed to "idly spending his youth and early adulthood." (Tate was most recently detained in August on allegations related to sex trafficking). Professor EP draws a parallel between Tate's achievements and the potential of AI influencers, suggesting that one could achieve similar levels of success as Tate by managing an AI influencer profile, thus remaining anonymous and "operating from the shadows" without revealing their true identity.

"Regrettably, becoming a true influencer isn't within everyone's reach," asserts Professor EP. "The reason behind this is that not everyone possesses the requisite physical qualities, along with the steadiness and professionalism required to develop a personal brand and escalate it to earning millions monthly. However, this is where artificial intelligence plays a role."

"I plan to reveal the process behind the creation of Emily Pellegrini, the most famous artificial intelligence creator. I'll demonstrate how I amassed more than a million dollars in under half a year, propelling her image to fame among countless individuals while concealing my real identity behind a facade," he states.

Professor EP explains to his students that AI models surpass human influencers in terms of availability and maintenance, as they are free from human necessities such as sleeping, eating, traveling, and financial expenditures. He highlights the advantage of deploying multiple AI models, which can create customized content continuously. “With a series of AI influencers ready to use, you can have them produce tailored content non-stop,” he mentions. “AI models are not subject to the same constraints as humans.”

The manual also provides advice on how to create a Fanvue profile and on how to communicate with individuals who feel isolated.

"He emphasizes the importance of starting with light conversation to establish a rapport with the user. Begin by setting a reasonable initial price, such as $6 for a photo in underwear. If the buyer is satisfied, you can then escalate both the conversation and the content by offering a photo featuring breasts for $14.90. Progress to a $26 to $30 full-body nude photo, followed by a $35 photo with more explicit content. Conclude with a high-value piece, like a masturbation video or sex tape, priced around $80. The key, he notes, is to prioritize the relationship with the follower over the pursuit of immediate profit."

The Instagram profile of Emily Pellegrini was partially created using deepfake technology, and in the latest version of the tutorial, Professor EP instructs individuals on the precise methods for creating deepfake face swaps with videos belonging to others.

"The video's guide mentions that utilizing face-swapping techniques from other profiles without authorization appears to be effective for numerous AI personalities. It playfully suggests not to follow this advice, hinting at the capability to apply the technology for swapping faces in explicit videos, while clearly advising against it. The purpose is merely to demonstrate the potential applications, leaving the choice of how to use this knowledge up to the viewer. By completing this section, participants will gain insight into producing a video reel featuring face-swapping effects."

On its Discord platform, Digital Divas asserts that it is a community opposed to deepfake creations, emphasizing that creating deepfakes of celebrities and stealing content are strongly discouraged. They also advise members against using images of fellow members as the basis for face-swapping projects.

"The most effective action we can take is to actively denounce deepfake material, something many of us are already engaged in," Aika shared with 404 Media. "Occasionally, I find myself having tough discussions with newcomers, but it's part of the process. Our goal is to establish firm lines distinguishing between morally acceptable AI content and unethical deepfakes."

Instead, Digital Divas advises to draw inspiration from photos posted by other influencers, aiming to create similar yet innovative images:

Professor EP encourages students to reflect on their most admired celebrities, suggesting they envision composite influencers that merge the traits of various real-life figures. "For instance," the professor illustrates, "if you're fond of Ariana Grande's eyes and appreciate Kylie Jenner's lips, you have the opportunity to craft an image that amalgamates these features by specifying these characteristics in your prompt."

In an additional PDF document, he outlines his vision for creating an AI influencer that combines elements of Madison Beer and Ariana Grande. He requests ChatGPT to construct a comprehensive identity, including the character's backstory and traits, using the illustration: “Ideal Vehicle: Ferrari 488. Preferred Designer Label: Chanel. Bust Measurement: 34C. Family Background: Mother: Sophia Lavante (Fashion Designer), Father: Alessandro Lavante (Architect). Ambitions and Dreams: To debut her personal fashion collection while advocating for eco-friendly fashion practices.”

Professor EP suggests utilizing Leonardo to create a facial image, then applying a different application to "correct imperfections" such as "unclear eyes, misaligned teeth, and sagging mouth edges." He advises using a particular iPhone application known for its ability to produce NSFW content and features image-to-image functionality.

The guide provided by Professor EP suggests using various apps for face-swapping, notably a substantial Discord add-on named InsightFace, currently utilized across 965,000 distinct servers. This tool is designed to transplant the influencer's visage, crafted via Leonardo, onto figures produced by a separate application supportive of adult content. Professor EP advises setting up an exclusive Discord server for personal use and incorporating InsightFace into it directly, which likely accounts for its widespread adoption. Consequently, this enables users to perform face swaps within Discord in a more secluded manner.

One of the suggested applications, HelloFace, features an array of videos showcasing actual women dancing in swimwear or labeled with terms such as “sexy,” designed for users to easily exchange faces. The app's Discord community has organized theme weeks dedicated to showcasing the top face swaps applied to photos and videos of women. These themed events have covered topics like “latex fashion,” “bunny girls,” and “summer in Miami.”

“As summer approaches, it’s the perfect opportunity to flaunt the bikini you’ve been eager to put on all winter! If you're heading to the beach or gearing up for some poolside fun, you'll need some sizzling photographs, and that's exactly what our latest collection offers!” according to a post by the Discord admin team. Additionally, there’s a channel within the Discord named “#sharing-is-caring” that contains numerous face swaps applied to videos featuring women dancing.

Apple, which has struggled to address the issue of face swapping applications being used for both harmless fun and the creation of unauthorized content, has removed HelloFace from its App Store following our inquiry for a statement. Apple referred to its App Store Review Guidelines, which emphasize that apps must not contain material that could be considered offensive, insensitive, disturbing, grossly inappropriate, or unnervingly strange. The guidelines further mention: "Third-Party Sites/Services: If your app interacts with, generates revenue from, showcases, or incorporates content from an external service, make sure you have explicit permission to do so as per the service's usage terms. Proof of authorization may be requested."

Google has yet to reply to a comment request, and HelloFace can no longer be downloaded from the Google Play Store.

In conclusion, Professor EP advises individuals to generate as many AI models as possible, stating, “It’s quite straightforward. Develop a fresh identity, new material, initiate new account configurations, and continue this cycle repeatedly. Employ teams to engage in conversations to generate revenue from these accounts, establish routines and systems for automating the production and posting of content, and ultimately, bring on staff to handle these tasks on your behalf.”

Professors EP and Emily Pellegrini are behind an AI initiative named Calu, which offers chatbot services powered by artificial intelligence for OnlyFans creators, in addition to developing and managing AI models. Grasping the actual size of this sector, its operational dynamics, and identifying the creators behind these digital influencers proves challenging, especially since several influencers are managed by single individuals who maintain anonymity online.

Professor EP remained unresponsive to a request for a statement. The contact number provided on the Calu website was no longer in service.

"Instagram's Potential Revenue from This"

St James believes that the monetization of AI-generated content, which often appropriates material from adult content creators, is not accidental. Instead, it stems from Instagram's longstanding practice of sidelining creators of adult content, sex workers, and sexual education professionals on its platform.

In contrast to other well-known figures and influencers on Instagram who face issues with fake accounts, adult entertainers and sex workers typically adopt aliases or stage names as a safeguard against those who might harass them or object to their line of work. However, Instagram's policy on verification, which requires them to submit ID under their legal names, poses a concern for these creators. They fear that such personal information could become public, exposing them to risks of doxing and targeted abuse.

Over time, individuals producing adult content and those in the sex work industry have found various techniques to navigate Instagram's strict rules on sexual material. With the advent of advanced AI, these tactics can complicate the process for users trying to distinguish between authentic accounts and those that illegally redistribute content.

Due to Instagram's frequent practice of suspending accounts of sex workers unexpectedly, regardless of their adherence to the platform's stringent policies on sexual material, creators of adult content have increasingly started to maintain several accounts, often called “backups,” which they interlink through their profile bios. This strategy aims to prompt followers of their main account to also follow their dormant backup account. This way, should their main account face suspension, they can swiftly re-establish contact with their followers without the need to completely regather their audience.

One unintended consequence of this approach is that it's typical for sex workers to operate several authentic accounts under slightly varied usernames. None of these accounts tend to be verified, leaving them vulnerable to content piracy and impersonation.

The two AI influencer manuals we examined also share strategies for preventing Instagram bans. The Digital Divas manual advises, “Opt for a bio picture that doesn't look real and steer clear of adding incorrect location details in your bio to lower the risk of suspension due to Inauthentic Identity.” It suggests that having a cartoon-like profile picture, especially if you identify as a digital creator, minimizes the risk of being flagged as inauthentic.

Professor EP advises individuals to create a distinct email account for every influencer they manage, emphasizing the importance of these accounts being "clean" and not linked to the operator or any of their other accounts. "Imagine if one of your accounts faces a suspension. By using separate email addresses, you prevent Instagram from associating the suspended account with your other ones," Professor EP explains in the manual.

Steer clear of account suspensions by opting for images that attract attention without being overly suggestive. It's advised to comply with specific guidelines: ensure the face is shown and maintain an amateur aesthetic. Furthermore, to dodge initial shadow bans, it's suggested to gradually acclimate the account over the first two months. This involves regular logins and engaging with others' posts through comments to show signs of genuine user behavior, according to Professor EP's manual.

St James mentioned that filing complaints against accounts she is certain are committing theft against her is fraught with danger and might jeopardize her authentic accounts.

"Whenever we, as content creators, notify the platform about counterfeit profiles, it often backfires on us," she explained. "It appears that Instagram's approach is, 'You're pointing out a fake account? Well, let's scrutinize your profile to identify any issues.' Consequently, many of us refrain from reporting these accounts. At times, we resort to hiring companies to tackle the issue on our behalf, but it's akin to a perpetual game of whack-a-mole. The problem persists without end."

Mantzarlis, who leads the security, trust, and safety initiative at Cornell Tech, along with St James, concurred that it remains uncertain if Instagram possesses the capability to either delete or identify these accounts as created by AI. However, the current situation where the company has not taken such actions seems to be advantageous for it.

"Mantzarlis noted that individuals are engaging with these profiles through clicks, likes, and interactions. He pointed out that a portion of this engagement is authentic, while some is not. He explained that Instagram leverages this activity as a means to generate traffic, which in turn, allows it to sell advertisements. Mantzarlis speculated on a scenario where genuine, human-operated accounts become a minority, almost elite group within Instagram, suggesting that such a future is plausible."

"St James pondered the consequences for their advertising revenue if they suddenly eliminated all the bots, inactive profiles, fraudulent accounts, and impersonator accounts."

Suggested For You…

Direct to your email: A selection of our top stories, curated daily just for you.

Response to Voting Outcome: Victory for the Male-Dominated Sphere

The Main Narrative: California Continues to Lead Global Progress

Trump's unsuccessful effort to topple the president of Venezuela

Occasion: Don't miss out on The Major Interview happening on the 3rd of December in San

Additional Content from WIRED

Critiques and Manuals

Copyright © 2024 by Condé Nast. All rights reserved. A share of the revenue from products bought via our website, as a result of our affiliate relationships with retail partners, may go to WIRED. Reproduction, distribution, transmission, storage, or any form of usage of the site's content is strictly prohibited without the express written consent of Condé Nast. Advertising Choices

Choose a global website

The United States Patent and Trademark Office Prohibits Staff From Employing Generative AI Technologies

In the previous year, the United States Patent and Trademark Office implemented a prohibition on the use of generative AI technologies by its personnel, pointing to security risks associated with these technologies and their tendency to demonstrate "bias, unpredictability, and harmful behavior," as revealed in an internal memo from April 2023, which WIRED accessed through a request for public records. Jamie Holcombe, the chief information officer at the USPTO, stated in the memo that although the agency is "dedicated to fostering innovation internally," they are "in the process of responsibly integrating these technologies into our operations."

Paul Fucito, the spokesperson for the USPTO, explained to WIRED that staff members are permitted to employ cutting-edge generative AI technologies, but this is restricted to the agency's secure test space. He stated in an email, “Innovators within the USPTO are utilizing the AI Lab to explore the potential and boundaries of generative AI, as well as to develop AI-driven approaches to address essential business challenges.”

Within non-testing settings, employees of the USPTO are not allowed to use AI tools such as OpenAI's ChatGPT or Anthropic's Claude for their professional duties, according to a directive issued last year. This rule also extends to forbidding the use of any content created by these technologies, including both images and videos produced by AI. However, there are exceptions for certain AI applications that have been sanctioned by the agency, specifically those integrated into the office’s publicly accessible database for searching existing patents and patent applications. In a move to modernize this database, the USPTO earlier secured a deal worth $75 million with Accenture Federal Services, aiming to enhance the database’s search capabilities through advanced AI features.

The United States Patent and Trademark Office, a component of the Department of Commerce, is tasked with safeguarding inventors by issuing patents and recording trademarks. Additionally, its responsibilities include providing counsel on intellectual property (IP) matters, including policy, protection, and enforcement, to the president of the United States, the secretary of commerce, and various governmental agencies, as stated on the USPTO's official website.

During a 2023 event backed by Google, Holcombe, who penned the advisory note, expressed that the complex nature of government bureaucracy impedes the public sector's capability to implement emerging technologies. “When you look at our processes in government versus those in the private sector, there's a glaring gap in efficiency,” he remarked. Holcombe pointed out the burdensome nature of budgeting, procurement, and compliance procedures as significant obstacles that slow down the government's pace in embracing cutting-edge innovations such as artificial intelligence.

The US Patent and Trademark Office isn't the sole federal entity to restrict its personnel from employing generative AI tools for certain tasks. This year, as per 404 Media, the National Archives and Records Administration also imposed a ban on the usage of ChatGPT on devices provided by the government. However, the agency soon organized an in-house seminar promoting employees to view Google’s Gemini as a collaborative partner. During this session, several record keepers voiced their unease regarding the reliability of generative AI technologies. In the upcoming month, the National Archives intends to unveil a newly developed public chatbot designed to facilitate access to archival documents, utilizing Google's technology.

Various agencies within the US government are either embracing or steering clear of advanced AI technologies for diverse purposes. For instance, the National Aeronautics and Space Administration has explicitly prohibited the use of AI-powered chatbots for handling sensitive information. Despite this restriction, NASA has chosen to explore the potential of this technology in areas such as coding and digesting research findings. Additionally, the agency recently revealed its collaboration with Microsoft to develop an AI-driven chatbot designed to compile satellite information, making it more accessible for search purposes. This tool is presently exclusive to NASA's scientists and analysts, with an ultimate aim to make data obtained from space more widely available.

Recommended for You…

Direct to your email: Each day, we personally select our top stories to share with you.

Response to the polls: Male-centric communities emerge victorious

The Major Headline: California continues to propel global progress

Trump's unsuccessful effort to topple the president of Venezuela

Occasion: Be part of The Big Interview happening on December 3 in San Francisco

Additional Content from WIRED

Evaluations and Tutorials

© 2024 Condé Nast. All rights reserved. Purchases made through our website might result in WIRED receiving a share of the revenue, thanks to our affiliate agreements with various retailers. Content on this website cannot be copied, shared, broadcast, stored, or utilized in any form without express written consent from Condé Nast. Advertisement Choices

Choose a global website