On November the 17th, 2023, OpenAI’s board of directors announced CEO Sam Altman’s dismissal and President Greg Brockman’s subsequent removal from the board. Just four days later Altman was returned to his post.
When Altman was fired, according to an OpenAI blog post, Altman had not been ‘candid’ with the board, and his communications weren’t sufficiently transparent.
In the chaos that ensued, Altman and Brockman were set to join Microsoft on a new research team. Employees rebelled against OpenAI’s board, with the vast majority stating they’d quit unless the board reversed their decision.
Twitch co-founder Emmet Shear joined as CEO for a matter of days before Altman, through the force of collective action, was reinstated as CEO.
It’s probably the most extraordinary series of events in Silicon Valley company governance history.
Sam Altman’s background at OpenAI
Born in Chicago in 1985, Altman displayed a prodigious interest in technology from an early age. By eight, he was already programming and dismantling computers. In 2015, Altman co-founded OpenAI with other tech bosses like Elon Musk.
Prior to his role at OpenAI, Altman served as the president of Y Combinator, a prominent incubator for tech startups, starting in 2014. Under his leadership, Y Combinator saw a deluge of startup applications totaling 40,000 annually.
OpenAI initially started as a non-profit, but after Musk’s departure in 2018, Altman took the reins as CEO of OpenAI.
Under his guidance, the company transitioned to a for-profit model in 2019, quickly securing a $1 billion investment from Microsoft. This is a key event in OpenAI’s history.
It laid the groundwork for the development and public release of ChatGPT in November 2022. At 38, Sam Altman was leading a company that gained a staggering 100 million users within two months. The age of generative AI had thoroughly begun, and OpenAI sat firmly at the center of this new ecosystem.
ChatGPT’s launch catalyzed a wave of AI funding in Silicon Valley, amounting to $17.9 billion, according to Bloomberg.
Despite ChatGPT’s success, OpenAI, under Altman’s leadership, has faced its share of challenges. Issues of intellectual property, data privacy, and the ethical integration of AI into society have sparked fierce debates.
OpenAI’s unique board
OpenAI operates under a unique governance model. Its board, composed of individuals who do not hold any equity in the company, is driven by a mission to create safe artificial general intelligence (AGI) that benefits humanity broadly rather than focusing on profits. AGI is AI that broadly equals or excels human cognition.
The board’s unusual setup reflects OpenAI’s commitment to its founding principles while reflecting its progression from its initial non-profit status in 2015 to a capped for-profit company.
In addition to Altman and Brockman, OpenAI’s board before the firing of Altman included:
- Adam D’Angelo: The CEO of Quora, Adam D’Angelo, has been a part of OpenAI’s board since April 2018. Known for his thought leadership in AI, D’Angelo’s perspective on the ethical and safe development of AI technologies is seen as a key influence in the board’s decisions. D’Angelo remained on the board after Altman was reinstated.
- Tasha McCauley: A senior management scientist at RAND Corporation, McCauley’s journey from co-founding Fellow Robots to her current role has been marked by a strong focus on innovation and AI governance. Her involvement with the Asilomar AI Principles and guidelines for AI development aligned closely with OpenAI’s founding principles. McCauley was removed from the board after Altman was reinstated.
- Ilya Sutskever: As an OpenAI cofounder alongside Altman and Brockman, Sutskever’s influence was substantial. His computer science background and role as the initial research director and later chief scientist at OpenAI give him a deep understanding of the company’s technological and research objectives. His contributions have been vital in shaping the organization’s direction and research priorities. Sutskever was removed from the board after Altman was reinstated.
- Helen Toner: Toner’s expertise in AI policy and ethics, stemming from her role at Georgetown’s Center for Security and Emerging Technology, has been instrumental in navigating the global implications and ethical considerations of AI development. Toner was removed from the board after Altman was reinstated.
As of 26/11/23, OpenAI’s new board includes D’Angelo and:
- Bret Taylor: Bret Taylor, aged 43, is a prominent figure in the tech industry, known for his role as a computer programmer and entrepreneur. He has served as a board member at Shopify and formerly co-CEO of Salesforce.
- Larry Summers: Larry Summers, 68, is an American economist with a rich background in public service and academia. He served as the US Secretary of the Treasury from 1999 to 2001 and as director of the National Economic Council from 2009 to 2010 under President Obama. Summers is also a professor at Charles W. Eliot University and President Emeritus at Harvard University. Additionally, he sits on the boards of Block Inc. and Skillsoft Corp and is an adviser to Andreessen Horowitz.
OpenAI intends to expand the board to up to nine members.
A timeline of events of Altman’s departure and return
The departure of Altman, who held only indirect shares in OpenAI through an investment by Y Combinator, where he was previously president, was a shock given his central role in the company.
According to sources, he was not involved in the vote to remove him, indicating possible internal disagreements or shifts in vision for the company’s future.
The removal of Brockman from the board and subsequent resignation contributed to the dramatic changes at the top of the organization.
Here’s a timeline of this rollercoaster of events:
Pre-November 17th, 2023
- Leading up to the incident: OpenAI operates normally, with no public indication of the impending leadership changes. He continues his role as CEO, engaging in regular activities and representing OpenAI at various events.
November 17th, 2023
- Early hours: Sam Altman is involved in his usual duties, including speaking at the Asia-Pacific Economic Cooperation Summit in downtown San Francisco.
- Evening: Altman attends a gathering at a warehouse in Oakland, speaking to a group of artists about generative AI. He appears relaxed and engaged, showing no signs of the upcoming announcement.
- Late evening: Altman leaves the Oakland event, mentioning he is late for a meeting.
- OpenAI’s board decision: OpenAI’s board, following a review, concludes that Altman was not consistently candid in his communications with the board. This is perceived as hindering the board’s ability to fulfill its responsibilities.
- Announcement of Altman’s departure: OpenAI publicly announces Sam Altman’s termination as CEO. The board cites a lack of confidence in his ability to continue leading the company as the reason for his dismissal.
November 17th, 2023 (post-announcement)
- Greg Brockman’s reaction: Shortly after the announcement, Greg Brockman, co-founder and former board chair of OpenAI, resigned. Despite initial plans for him to remain at the company in a different capacity, he decides to leave entirely.
- Mira Murati’s appointment: Mira Murati, OpenAI’s chief technology officer, is named interim CEO following Altman’s departure.
- Sam Altman’s response: Altman made his first public statement after his dismissal on social media platform X. He did not provide specific details about the incident but indicated he would discuss his future plans later.
- Other colleagues leave: Several other colleagues leave the company in solidarity.
Sam and I are shocked and saddened by what the board did today.
Let us first say thank you to all the incredible people who we have worked with at OpenAI, our customers, our investors, and all of those who have been reaching out.
We too are still trying to figure out exactly…
— Greg Brockman (@gdb) November 18, 2023
Post-November 17th, 2023
- Industry reaction: The tech and AI community reacts with surprise. Partners, including Microsoft, and insiders express shock, having been unaware of the board’s deliberations and decision.
- Uncertainty: The tech industry begins to speculate about the future direction of OpenAI, particularly in the context of its rapid growth and influence following the launch of ChatGPT.
first and last time i ever wear one of these pic.twitter.com/u3iKwyWj0a
— Sam Altman (@sama) November 19, 2023
November 20th, 2023
- New OpenAI CEO: New OpenAI CEO: Emmet Shear, co-founder of Twitch, is appointed as OpenAI’s new CEO.
- Microsoft hire Altman and Brockman: Microsoft hired Sam Altman and Greg Brockman. They’re set to head a new AI research lab. However, reports had previously indicated rising tensions between the two companies.
- Employees rebel: Some 550 of 700 OpenAI employees, including most top researchers, sign a letter accusing the board of acting recklessly, leading to mass fallout and subsequent drama on X. The signees threaten to leave to join Microsoft, which will allegedly give any OpenAI leavers jobs alongside Altman and Brockman.
- Pressure grows: The letter was eventually signed by virtually all OpenAI employees.
- Microsoft swoops in: Microsoft says there are spaces for all OpenAI employees at their company, threatening a major power grab and the demise of OpenAI. Observers grow wary that Microsoft will own all the power and talent in AI, including Elon Musk, who said this would be the worst outcome.
- Speculation: Many speculated that OpenAI will fall into obscurity if the mass exodus precipitates. Microsoft might become the new leader in generative AI, absorbing OpenAI’s expertise into its mature for-profit structure.
Today I got a call inviting me to consider a once-in-a-lifetime opportunity: to become the interim CEO of @OpenAI. After consulting with my family and reflecting on it for just a few hours, I accepted. I had recently resigned from my role as CEO of Twitch due to the birth of my…
— Emmett Shear (@eshear) November 20, 2023
November 21st, 2023:
- Sam Altman reinstated: After immense pressure from employees and investors, the board returned Altman to his post as CEO, and OpenAI essentially reformed. Brockman also returned to his post.
- X explodes: Altman, Brockman, and others post numerous times on X to emphasize their triumph in returning OpenAI to its former glory.
Post-November 21st, 2023:
- Diagnosis: Supporters and critics of OpenAI begin trying to figure out what happened. Ulterior motives are unraveled, but none are confirmed.
- Q*: Some speculate that OpenAI is sitting on a dangerous model, which caused the board to fire Altman in fear that he doesn’t take AI safety seriously. More on that here.
- Business as usual: As the dust settles, Altman, Brockman, etc., continue their mission.
OpenAI’s board U-turned, but why? And what does it mean?
The removal and reinstatement of Altman from OpenAI was shocking, not merely because it involves a high-profile figure in the tech world, but because it symbolizes the deep-seated challenges of AI governance.
AI is fast-moving and subject to intense public debate, speculation, and regulatory impact.
The decision by OpenAI’s board to remove Altman was reportedly driven by disagreements over the speed of AI commercialization and differing views on AI safety. It showcases the conflict between the rapid advancement of AI and the need for responsible, ethical oversight.
It seems like Altman, Brockman, and most OpenAI employees, wanted to continue at their current trajectory. They aim to create artificial general intelligence (AGI) which exceeds human cognition.
In the meantime, OpenAI’s value has rocketed towards the $100 billion mark, hovering at around $86 billion currently.
OpenAI was a non-profit but transitioned to a novel for-profit structure in 2019, so it’s a remarkable rise.
However, rather than representing financial backers, OpenAI’s board was – and still is – tasked with acting in the interests of humanity. This was reflected in the board’s composition, which, until recently, included three of OpenAI’s co-founders and three independent members.
Critics argued that such a structure, with its self-governing approach and lack of external accountability, may not be sustainable in the long term. That view has been proved true for now, and the new board will likely not take the risks of their predecessors.
However, this has shown that if the board is concerned about the company’s direction, affirmative action is far from guaranteed.
In many tech companies, founders often influence the company’s direction. Their vision and charisma can overshadow traditional corporate governance structures, but this took that to a new level.
“King Sam,” as some major news outlets have called him, cannot be ousted by the board.
Safety issues motivated Altman’s firing, maybe
Most agree that the board fired Altman over safety concerns, but despite this becoming the de facto account of events, there is little hard evidence.
One theory supporting this suggests an underlying debate about safety from Dustin Moskovitz, an early Facebook employee and founder of Asana, who founded Open Philanthropy. Moskovitz is an influential figure in AI but has been ramping up talks over AI safety and the need for restraint.
Moskovitz’s connections with board members like Adam D’Angelo, Helen Toner, and Tasha McCauley, who have ties to organizations he supports, suggest his potential influence over OpenAI’s governance.
Toner also criticized the organization for releasing GPT models too early, allegedly prompting a “race to the bottom” in AI development. This implies that OpenAI’s approach to releasing AI models may be reckless or unethical, contrasting with more cautious approaches like that of Anthropic, which Toner referred to in a paper as a better model of AI governance.
Some viewed that as a valid critique of OpenAI’s strategies, while others questioned its quality and the appropriateness of publicly criticizing the company one is part of.
Finally, there was speculation that OpenAI was sitting on a dangerously powerful model called Q*.
Reuters reported that insiders contacted them to say that employees were worried about Q* and warned the board – but that doesn’t seem to be in keeping with the solidarity that brought Altman back to the company.
However, some have mentioned that employee shares go up for sale soon. Altman’s departure would have delayed this, at least.
So, was safety ever an issue?
We don’t yet know. It certainly seems OpenAI’s unique structure, blending non-profit objectives with for-profit operations, has lost its shine this year, however.
Sarah Kreps, director of the Tech Policy Institute at Cornell University, said, “Sam Altman and Greg Brockman seem to be of the view that accelerating AI can achieve the most good for humanity. The plurality of the [old] board, however, appears to be of a different view that the pace of advancement is too fast and could compromise safety and trust.”
As OpenAI has launched commercially successful products and attracted significant investments, it’s feasible to think that Altman, Brockman, and most employees are simply not prepared to put the brakes on – it goes against human nature, not to mention Silicon Valley culture.
OpenAI has rapidly become one of the world’s most influential countries. They won’t want to let their rivals sail past them by pausing progress.
Who benefited from all of this?
OpenAI is forming a new board, currently including Bret Taylor (ex-Salesforce), Larry Summers (former United States Secretary of the Treasury), and the retention of Adam D’Angelo.
This will expand to nine directors and will undoubtedly shape the organization’s future trajectory.
Regaining his position, Altman is now the uncontested leader of OpenAI. With new board allies and less internal resistance, Altman’s vision for AI will likely proceed with less opposition. It seems that relying solely on internal governance is insufficient to manage AI’s complex impacts.
we have more unity and commitment and focus than ever before.
we are all going to work together some way or other, and i’m so excited.
one team, one mission.
— Sam Altman (@sama) November 20, 2023
OpenAI will continue at its same relentless pace, hurtling towards AGI. We’ll know what the impacts are when we get there, as it doesn’t seem like anything will stop it.
At least, not for longer than about four days.
