Sam has superior table stakes.
Financial backing to make a competitor
Internal knowledge of roadmap
Media focus
Alignment with the 2nd most valuable company on the planet.
I could go on. I strongly dislike the guy but you need to recognize table stakes even in your enemy. Or you’ll be like Ilya. A naive fool who is gonna get wrecked thinking doing the “right” thing in his own mind will automatically means you win.
A true believer is going to act along the axis of their beliefs even if it ultimately results in failure. That doesn't necessarily make them naive or fools - many times they will fully understand that their actions have little or no chance of success. They've just prioritized a different value of you.
Ultimately this is good for competition and the gen-AI ecosystem, even if it's catastrophic for OpenAI.
From my read, Ilya's goal is to not work with Sam anymore, and relatedly, to focus OpenAI on more pure AGI research without needing to answer to commercial pressures. There is every indication that he will succeed in that. It's also entirely possible that that may mean less investment from Microsoft etc, less commercial success, and a narrower reach and impact. But that's the point.
Sam's always been about having a big impact and huge commercial success, so he's probably going to form a new company that poaches some top OpenAI researchers, and aggressively go after things like commercial partnerships and AI stores. But that's also the point.
Both board members are smart enough that they will probably get what they want, they just want different things.
IMO, there are basically two justifiably rational moves here: (1) ignore the noise; accept that Sam and Greg have the soft power, but they don't have the votes so they can fuck off; (2) lean into the noise; accept that you made a mistake in firing Sam and Greg and bring them back in a show of magnanimity.
Anything in between these two options is hedging their bets and will lead to them getting eaten alive.
Any decision that doesn't make the 'line go up' is considered a dumb decision. So to most people on this site, kicking Sam out of the company was a bad idea because it meant the company's future earning potential had cratered.
pushing to call it a coup is an attempt to control the narrative.
Sure, I guess I didn't consider them, but you can lump them into the same "media campaign" (while accepting that they're applying some additional, non-media related leverage) and you'll come to the same conclusion: the board is incompetent. Really the only argument I see against this is that the legal structure of OpenAI is such that it's actually in the board's best interest to sabotage the development of the underlying technology (i.e. the "contain the AGI" hypothesis, which I don't personally subscribe to - IMO the structure makes such decisions more difficult for purely egotistical reasons; a profit motive would be morally clarifying).
Here is what I understand by table stakes: https://brandmarketingblog.com/articles/branding-definitions...
If your objective is to suppress the technology, the emergence of an equally empowered competitor is not a development that helps your cause. In fact there's this weird moral ambiguity where your best move is to pretend to advance the tech while actually sabotaging it. Whereas by attempting to simply excise it from your own organization's roadmap, you push its development outside your control (since Sam's Newco won't be beholden to any of your sanctimonious moral constraints). And the unresolvability of this problem, IMO, is evidence of why the non-profit motive can't work.
As a side-note: it's hilarious that six months ago OpenAI (and thus Sam) was the poster child for the nanny AI that knows what's best for the user, but this controversy has inverted that perception to the point that most people now see Sam as a warrior for user-aligned AGI... the only way he could fuck this up is by framing the creation of Newco as a pursuit of safety.
Please get real.
I'm not sure that's actually true anymore. Look at any story about "growth", and you'll see plenty of skeptical comments. I'd say the audience has skewed pretty far from all the VC stuff.
Or they'll do something hilarious like sell VCs on a world wide cryptocurrency that is uniquely joined to an individual by their biometrics and somehow involves AI. I'm sure they could wrangle a few hundred million out of the VC class with a braindead scheme like that.
"Table stakes" simply means having enough money to sit at the table and play, nothing more. "Having a big pile of GPUs is table stakes to contest in the AI market."