OpenAI boss Altman is back
The dream of AI in the interest of everyone has been shattered
By Hannes Vogel
November 22, 2023, 2:52 p.m
Listen to article
This audio version was artificially generated. More info | Send feedback
Sam Altman once wanted to create artificial intelligence for the benefit of humanity. Then came the billions from Microsoft. In the religious war that is dividing the AI scene, he has prevailed: quick profits now take precedence over safety.
When Sam Altman received a text message from his technology director and co-founder Ilya Sutskever last Friday, he had no idea that his last day at OpenAI had just begun. Sutskever wanted to know whether he could take part in a video call at lunchtime. At the meeting, Sutskever and the rest of his board then told Altman he was fired. Immediately afterwards, his accounts were blocked. This is what the US media reports.
Altman’s expulsion sent shockwaves through the US tech scene, if not worldwide. Because the ChatGPT developer is not just any startup, but the figurehead of the AI revolution. And its founder and boss Sam Altman her face. After the employees openly rebelled and threatened to defect to the largest investor, Microsoft, or to join the competition, the supervisors gave in. After tough negotiations, Altman has now triumphantly returned as CEO – just days after he was fired. Perhaps the most important technology company in the USA is now calm for the time being. But it remains deceptive.
Because tech analysts, journalists and investors continue to puzzle over what exactly happened at OpenAI. Even the AI company doesn’t know and wants to have the events investigated by external investigators. One thing is clear: For some, Sam Altman sacrificed his noble mission for big money. For the others, he has become the target of hyper-cautious and incompetent supervisors. He did win the power struggle with his internal company opponents. But his dream of artificial intelligence, which is developed for the benefit of humanity free of the commercial interests of the tech giants, will probably have to be buried with the drama at OpenAI.
Balancing act between saving the world and commerce
That was exactly the goal with which Altman founded his startup in San Francisco in 2015. The altruistic idea was already evident in the name: OpenAI, open and transparent. Just as in form: not a for-profit company, but a non-profit research organization is still formally behind the startup’s groundbreaking software.
But Altman couldn’t maintain this claim for long: starting in 2019, Microsoft invested a total of up to $13 billion and secured exclusive rights to the AI artists’ programs. OpenAI founded a profit-oriented subsidiary specifically for this purpose, although its profits are capped.
Since then, OpenAI has been stuck in a unique, one might say comical, structure: billion-dollar investors, as is common in Silicon Valley, have no say in OpenAI’s supervisory board. And Altman in a balancing act: developing artificial intelligence for the benefit of humanity and with security in mind. And at the same time expand his non-profit into a billion-dollar business.
Doomsday hysteria or technology hype?
Long before Altman was expelled, there was a deep rift in OpenAI. As early as 2021, some AI researchers led by Dario Amodei left and founded AnthropicAI, Altman’s biggest competitor today – out of fear that their former boss had lost sight of the risks of the groundbreaking technology. Thanks to the deal with Microsoft, Altman was able to put more and more money and computing power into the programs. “We believed that we needed more than just increasingly powerful models: an adaptation to human values - security,” Amodei later explained his departure to Fortune.
His exit was an alarm signal that the religious war that was raging in the AI scene had long since affected OpenAI: On the one hand, there are the doom-mongers who see artificial intelligence as a potentially god-like machine that will one day get out of control humanity and animals could be wiped out. At least three of the supervisors who have now been fired are attributed to this faction. Sutskever is also said to have had doubts about whether Altman still had OpenAI’s noble mission in mind.
The optimists consider these doomsday scenarios to be exaggerated. For them, AI is perhaps humanity’s most important invention, with huge potential benefits for billions. “We’ve gotten to this point because insignificant, tiny risks have been hysterically inflated by the exotic views of sci-fi nerds and click-addicted journalism,” said venture capital veteran Vinod Khosla, one of the earliest open AI investors, commenting on Altman’s expulsion.
Now the profit interests of Internet giants dominate
At OpenAI, these two factions – hyper-cautious supervisors and profit-oriented major investors – have been fighting a covert battle for a long time. Last week it just broke out openly. The exact trigger remains unclear. Altman was “not consistently honest in his communication with the board” and thereby “hindered the board’s ability to carry out his duties,” the press release about his expulsion says cloudily.
This could have something to do with the investor tour in the Middle East that Altman has been on for weeks, according to media reports, to mobilize new money for cheap chips that OpenAI needs for its models. The basis of trust between the charismatic CEO, to whom investors depended, and his supervisors has apparently continued to erode – possibly because his course became too commercial for them.
The OpenAI supervisory board may have only done what is its mandate: pull the ripcord when it sees OpenAI’s charitable mission in danger. The irony is that the “incompetent” overseers, as the OpenAI employees called them in their protest letter, almost ruined the entire company in the process. And thus paved the way to commercialization.
Because three of the four critical, more safety-oriented supervisors who fired Altman are now being kicked off the supervisory board themselves. Ex-US Treasury Secretary Larry Summers and Bret Taylor come on board – ex-head of Salesforce and a proven Altman confidant. Taylor will also become chairman of the supervisory board. Altman has thus installed controllers who are sympathetic to him and who will support his faster, more commercial course and give him “the necessary resources for their success,” as Microsoft boss Satya Nadella wanted. In addition, up to six additional new supervisors are to be installed who may better represent the interests of the billion-dollar donors. They will soon determine how charitable the greatest technology revolution in human history really is.