OpenAI approached Anthropic about merger

I think it's pretty clear by this point that the board consists of several people who are woefully ignorant about the way the world actually works.

This just reinforces something I've been thinking for a while. When you look people at the top in medium to large companies, you'd think that they are extremely qualified and got there through merit alone. Although this is certainly true in many cases, it's probably not the case most of the time - a combination of luck and contacts often boost people to places they have no business being.

In general, people in high places can just surf the waves and go with the flow in collecting cash and not really make that much difference, relying on the other people closer to the work (engineers and frontline managers). But when making the right decisions can decide the fate of a company, you then see who's supposed to be there and who isn't.

Apparently they not only asked Dario to become CEO, they also asked Nat Friedman (Former GitHub CEO) and Alex Wang (Scale AI CEO). Flailing indeed.

OpenAI was planning to add more members to the board. Meanwhile Sam and GDB are fully productizing the company, even training GPT5. The safety group has been increasingly concerned and feels the non-profit charter has been slowly losing over it its $ minded startup side. They see the writing on the wall, as most employees care about those sweet $ ppu’s, and decide they have a last ditch effort, maybe 25% chance, of averting this successfully. They feel they have to do it otherwise the company will just be Microsoft accelerating GPT development into our oblivion. They aren’t quite successful.

Remember, this company was founded, 8 years ago, with Sama writing “Why You Should Fear Machine Intelligence.”

Although vaguely possible, there’s something else to the story as Emmett Shear said “The board did not remove Sam over any specific disagreement on safety, their reasoning was completely different from that.”

https://blog.samaltman.com/machine-intelligence-part-1

Only explanation acceptable for me now is Hanlon's razor mixed with a Despair Inc poster.

"Never attribute to malice that which is adequately explained by stupidity."

And

"Meetings : None of us is as dumb as all of us."

Two camps: one camp was into moving fast, cranking out products and profit. The other group watching from the side, claiming to be worried about the future of humanity. The group watching though had the power to fire the leaders of the other group, and they did.

It didn’t have to be any deeper philosophical disagreements or anything, just plain old rivalry and power grab. Of course it wouldn’t be written down as it’s just childish, but that’s how people behave, especially when newly found fame and power gets to their heads.

The most plausible sounding reason I’ve heard is that Open AI was originally a charity. People donated money to a cause. Most who donated got nothing in return. Those who donated a lot got board seats.

Now, the the employees of the for-profit arm of the organization are all trying to get rich while those who donated the original money get nothing. But… they still have their board seats, and they’re weaponizing them.

>I’ve interacted with board members with various companies over my career, and everyone on the board is extremely junior.

Yes, you've hit on the thing that bothered me when I read the members' bios after Altman's firing but didn't figure out until reading your comment. It's not so much what they say, as much as what they don't say. There is no greybeard, some seasoned current/former CEO of another company (no, the CEO of Quora does not quality). There is no VC representative.

I realize that OpenAI has this weird nonprofit/profit ownership arrangement which I still don't get, and that the VC board members/observers are presumably on the profit side, but that still doesn't answer the implied question!

Also if they are exploring mergers, they might consider merging with a university like stanford or cern or a bigger non profit.

Never been happier to be wrong. Altman basically unionized the openai labor force and the investors. He contained the board. Instead of scattering the board to 3 corners of the planet, it will be further contained by adding 6 board members including Sam and microsoft and hopefully brockman. Order will have been restored. And they lived happily ever after !

It has never been about AI safety. They just threw their little mission statement on the end like an email signature.

Nothing could possibly weaken their position more than the last 72 hours

You have a point, but one could easily frame it differently.

- the openai board is the non-profit, meant to make "good" decisions, and not necessarily business sense decisions.

- altman was fired because that board doesn't feel it can rely on him (both in action and in communication), he'll do what he wants and he won't be obvious about the fact he's not listening to them.

- now altman and co. are essentially going to kill openai as a company, leaving it only with the existing IP, so the board figures that they should bring in people who are interested in working / researching on top of openai's previous work.

- anthropic is composed of people who left openai due to safety concerns, so they're good people to partner with.

The fact that it would be better safety-wise if there was competition (maybe?) is immaterial when facing the fact that 1) openai appears to be collapsing, and 2) microsoft anyway has rights to openai's work, and Microsoft appears to be preparing to cannibalize openai.

In which sense less choice for developers harms AI safety?

(I'm saying this as someone that doesn't believe in AI safety as a worthwhile goal)

Maybe they’re aiming to start mining cryptocurrency.

Zero chance OpenAI ends up seeing any more compute if they merge.

They’re fools, and thought they could oust Altman over petty grievances and that no one would care, and when it blew up, they struggled to find a solution that saves face and their careers and avoids adverse legal judgments against them individually, and they’re still frantically searching? I mean, they didn’t even have replacements in mind. They are not planners.

Since the board hasn’t given any remotely convincing explanation for the firing [0], their miscalculation seems to have rather been that nobody would be asking any questions. Which is really bewildering.

[0] see https://news.ycombinator.com/item?id=38356534

wasnt over saftey though

For their sake it had better be the AI safety thing because if it wasn't they are in a lot of trouble even if that hasn't registered yet. Board member is not a job without risks.

First time having the curtain pulled back?