There is a third option where he stayed, they managed to find a compromise, and in so doing kept their influence in the space to a large extent.

I'm pretty sure they tried that before firing him.

[delayed]

The vote for firing him effectively took place on Thursday at the latest, given that Murati was informed about it that evening.

This has been a source of tension at least since the release of ChatGPT, so… yeah it’s not like the problem came out of nowhere. The governance structure itself is indicative of quite elaborate attempts to reconcile it.

I don’t know about that. Yes, there was tension built into the structure, something happened to trigger this. You don’t fire your CEO without a backup plan if this was an on going conflict. And if your backup plan is to keep the current president (who was the chair of the board until you removed him), that’s not a backup plan.

Everything points to this being a haphazard change that’s clumsy at best.

The question was “did they try to find compromise” not “was the firing haphazard.” The answer is definitely yes to the former.

you can interpret it exactly opposite: they tried to negotiate and he lied .

You are assuming there was absolutely no build up to the firing. Just because the disagreements weren’t public doesn’t mean they weren’t happening.

How will they be able to continue doing their things without money?

It seems like people forget that it was the investors’ money that made all this possible in the first place.

Developing new algorithms and methods doesn't necessarily, or even typically, take billions.

Yeah but testing if they work does, that's the problem.

There are probably load so ways you can make language models with 100M parameters more efficient, but most of them won't scale to models with 100B parameters.

IIRC there is a bit of a phase transition that happens around 7B parameters where the distribution of activations changes qualitatively.

Anthropic have interpretability papers where their method does not work for 'small' models (with ~5B parameters) but works great for models with >50B parameters.

Deep NN aren't the only path to AGI... They actually could be one of the worst paths

For Example, check out the proceedings of the AGI Conference that's been going on for 16 years. https://www.agi-conference.org/

I have faith that Ilya. He's not going to allow this blunder to define his reputation.

He's going to go all in on research to find something to replace Transformers, leaving everyone else in the dust.

You mean the person who just signed the letter saying he’d quit if the board doesn’t fire themselves and bring back Sam? That Ilya?

100M users perhaps?

But as I understand it they’re still losing money, as much as $0.30 on every ChatGPT query.

Not true

Sama on X said as of late 2022 they were single digit pennies per query and dropping

The only financial statements I believe are those signed of by external auditors. And even there my trust only goes that far.

Pretty sure that it would be illegal for them to tweet insider information like that if it were false, since it's effectively a statement to shareholders.

I'll take securities fraud for 420, please, but private.

That's exactly the point - by tweeting insider information you are making a public statement. We've learned this very recently...

Parent meant probably meant that there's no securities fraud since no securities are involved as it's not a traded company.

The shareholders are still invested, they still have a 401A Evaluation, and these statements are definitely going to have legal weight.

New models might have different economics…

Still, they must be bleeding money with the humoungous amount of queries they get.

Now that OpenAI is the leader in the field, it has a lot of monetisation avenues above and over the existing income streams of parterships, ChatGPT+ and API access.

Is their deal with Microsoft exclusive tech transfer wise? If not they can always sell/license what they have to Google, Facebook, and Amazon. They should be able to get quite a bit of money to last a while.

> If they didn’t fire him, Altman will just continue to run hog wild over their charter. In that sense they lose either way.

The story would be much more interesting if actually AI had fired him.

Moloch always wins.

[delayed]

LOL