On Friday the board of OpenAI fired its CEO Sam Altman without consulting its biggest investor – Microsoft.
So what? It backfired: by yesterday evening Altman, along with ousted chairman Greg Brockman and most of the company’s employees, had job offers to join a newly created Microsoft subsidiary. Ilya Sutskever, the board director and co-founder who led the charge for Altman’s dismissal, signed a letter calling for his own resignation.
It might look like musical chairs, but last weekend will go down as a pivotal moment for the future trajectory of AI and a case study in corporate mismanagement. We’ll return to this story in the days ahead but some clear winners have already emerged:
The “accelerationists”. Altman has referred to ideological “tribes” within OpenAI. Loosely defined, the main ones are:
- the “accelerationists” – a group of techno-optimists born of Silicon Valley; and
- the “decels” or “doomers” who fear AI presents an existential risk that needs to be tightly controlled.
Since OpenAI was founded the accelerationists have gradually lost ground on its board, but gained it in the sector as a whole. By the end of last Friday four directors were left (previous high-profile departures included Elon Musk and LinkedIn’s Reid Hoffman). Two of them, Tasha McCauley and Helen Toner, are scientists with ties to the effective altruism movement that seeks to minimise harms caused by technology. Sutskever meanwhile used to work for the “godfather of AI”, Geoffrey Hinton, who quit Google earlier this year after warning of the risk posed by “true intelligence”.
Why now? It’s not yet clear why they chose this moment to take down Altman. They said he failed to be “consistently candid” – was that about the safety of OpenAI’s products, or something else? According to tech journalist Kara Swisher, OpenAI’s Dev Day event on 6 November, which involved the CEO pushing consumer-like products, was an “inflection moment of Altman pushing too far, too fast”.
Ousting him was foolhardy, and was supposed to send a signal to take these concerns seriously. They’ve now replaced Altman with the more cautious Emmett Shear, a former Twitch executive, but the risk is that OpenAI loses its main asset – its workforce.
Microsoft. Satya Nadella, chief executive of Microsoft, has played a blinder. Altman’s firing raised the risk that he would lose a key ally at a company into which Microsoft has invested $13 billion. After it became clear the board wouldn’t accept his reinstatement, Nadella offered jobs to Altman, Brockman and other loyalist researchers thinking about leaving.
The upshot: a new AI lab, filled with talent and wholly owned by Microsoft – without the bossy board. An $86 billion subsidiary for a $13 billion investment.
“Power in the industry has shifted from talent to capital,” says Tim Gordon, founder of Best Practice AI. “The new era is not about new techniques or ideas. Fundamentally, it’s about how much data and compute you bring to the table.”
Old-school governance. Attempting to build a company with a “fiduciary duty to humanity” was bound to be tricky. The overarching non-profit model also meant Altman lacked equity incentive – a fact he sought to change by transitioning parts of OpenAI into a “capped” for-profit in 2019.
The weekend exposed a lack of corporate experience on the board, and a lack of teeth. Stopping Altman might have been the responsible move, but the execution was poor. Business 101 says let your chief investor know about your plans to replace the CEO. Microsoft deserved a heads up.
What next? At time of writing 700 of OpenAI’s 770 staff are calling for the resignation of its board and the reinstatement of Altman. That leaves multiple questions unanswered:
- If the board quits, is it status quo ante, with Altman’s power consolidated?
- If it doesn’t, how many staff will join the new Microsoft rival?
- Where does it leave eight investors including Sequoia Capital, Tiger Global Management and Andreessen Horowitz who were set to triple their $300 million investment through a planned sale of shares?
- Was there a specific incident that got Altman fired?
- Did it have anything to do with a lack of clarity on Altman’s pursuits outside OpenAI?
Answers will emerge. In the meantime this much seems clear: Altman and Sutskever had a difference of opinion on the balance between risk and reward in the world’s fastest-growing and probably most consequential business sector. And Altman won.
