It’s been nonstop dramatic action as the OpenAI leadership struggle unfolds—in public. Last week, founder and CEO Sam Altman was dramatically fired from his role at OpenAI, creators of DALLE and ChatGPT, with cofounder and board president Greg Brockman abruptly quitting in outrage at this action. Then, with over 70 percent of the team behind the groundbreaking LLM threatening to quit in protest, it was widely reported that Altman was in negotiations to rejoin the firm—and oust the board that had given him the boot.
We remain committed to our partnership with OpenAI and have confidence in our product roadmap, our ability to continue to innovate with everything we announced at Microsoft Ignite, and in continuing to support our customers and partners. We look forward to getting to know Emmett…
— Satya Nadella (@satyanadella) November 20, 2023
This morning, yet another reverse—Microsoft, a big investor in OpenAI, announced that it has hired Altman and Brockman to spearhead Microsoft’s own in-house AI initiative. And, this afternoon? Ilya Sustskever, the OpenAI chief scientist, said that he deeply regrets having been part of the board’s actions and “will do everything he can to reunite the company”—and Altman himself intimated that, “some way or other,” he and the leadership team are all going to work together, name-checking members of the OpenAI leadership team, including Mira Murati, OpenAI’s CTO, who served briefly as interim CEO until Emmett Shear, formerly Twitch leader, stepped in after being appointed by the board yesterday.
I deeply regret my participation in the board's actions. I never intended to harm OpenAI. I love everything we've built together and I will do everything I can to reunite the company.
— Ilya Sutskever (@ilyasut) November 20, 2023
While Microsoft CEO Satya Nadella expressed excitement that Altman and Brockman were joining the team while also affirming their commitment to working with OpenAI, hundreds of OpenAI employees signed a letter threatening to depart (most likely to join Altman at Microsoft) if he wasn’t reinstated.
It is very dramatic—enough to be captured on film (celebrated AI artist Claire Silver suggested West Wing creator Aaron Sorkin to direct), and even Altman said these few days will go in the history books. For those of us who are not members of that boardroom and are puzzling out what happened, we must make do with the statements—and anonymous leaks—of those who might know more.
Why did the board decide to oust Altman? According to Bloomberg, he’s been going around the world looking for funding for a chipbuilding venture rivaling NVIDIA and a separate venture intended to produce an AI-enabled hardware device.
The OpenAI board, according to Bloomberg’s report, booted him for lack of candor—and given the timing of Altman’s fundraising initiatives, it’s reasonable to assume that these may have been what they feel Altman wasn’t forthcoming about.
To enter into the realm of speculation, this could be because the board wanted OpenAI to have a stake in these ventures—or because the board, which has members linked to the effective altruism movement, wanted to uphold OpenAI’s founding nonprofit ethos and to get safety and alignment issues resolved before going full speed ahead on growth.
The AI industry will be watching closely, with Alphabet’s Bard, Meta’s Llama, Microsoft’s Bing and CoPilot, and X’s Grok all entering the contest to unseat ChatGPT as the most well-known and well-used of the LLM/chatbot applications. Not only underlying science, infrastructure, and alignment will determine the winner—as we are beginning to understand, management and leadership will be absolutely crucial in the race toward mass adoption of AI technology.
From the point of view of the millions of people who have incorporated ChatGPT and other AI tools in their daily lives, it is unlikely that there will be a noticeable difference, according to Przemek Chojecki, AI entrepreneur at ulam.ai. “From the end user, it probably won’t change much. Microsoft already has Bing Chat, which is pretty much ChatGPT. So users would be diverted to entirely Microsoft-owned products (rather than 49% which is how much stake Microsoft has in OpenAI). It seems so far that most of the OpenAI team would go to Microsoft too,” he told nft now.
“From the end user, it probably won’t change much. Microsoft already has Bing Chat, which is pretty much ChatGPT. So users would be diverted to entirely Microsoft-owned products.”– PRZEMEK CHOJEKI
Chojecki agrees with Altman that computing power will be pivotal in AI development. “Computing power is definitely one of the crucial things going forward, as we’re already pushing the existing hardware to the limits. Various teams are working on specialized machine learning chips that make AI training faster & cheaper,” he said.
Despite these corporate travails, as an expert working on the technical side of AI, Chojecki is hopeful for future developments in the field. “I’m looking forward to two things. One is already happening with more multimodality. That means that an AI model can process multiple modalities–for example, text and images. Recent updates to ChatGPT (with Vision) have great applications with people using it to narrate sports live or describe what they want to be painted, creating gaming assets or entire games.
So multimodality is one thing. Another thing, which is a longer shot, is reasoning, and by that, I mean proper reasoning, the one you can see in math proofs. This would largely accelerate our scientific knowledge, allowing AI to help scientists discover new things. Next step after that is only AGI,” he said.