Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Greg Brockman quits OpenAI

twitter.com

Post reply on HN
Print view View original

631–640 of 701 posts

Re: Greg Brockman quits OpenAI

#631

Post by vikramkr »

Paul-Craft wrote: post #596
vikramkr wrote: post #590

Earlier quoted context omitted.

They also said Greg was going to stay at the company and then he immediately quit. I find it very hard to believe that smart, savvy, and self interested are adjectives that apply to a board who doesn't know what their own chairman thinks.

Even smart, savvy, and self interested people can't always predict what individual humans are going to do. It's certainly an interesting wrinkle, but I don't think it's relevant to the limited scope of the analysis I've presented here.

He was the chair of the board. And they were wrong very quickly. It very much sounds like they spoke for him. Or he pretended that he was going to stay and then backstabbed them. Which, given how strongly aligned with altman he seems to be, not really a surprise. I have yet to see a single action from them that means towards saavy rather than incompetent.

Re: Greg Brockman quits OpenAI

#632

Post by Rastonbury »

bananapub wrote: post #629
danbmil99 wrote: post #615

Earlier quoted context omitted.

This. There were no investors on the board -- as Jason @ all-in said "that's just crazy".

> as Jason @ all-in said lol > "that's just crazy". why is it crazy? the purpose of OpenAI is not to make investors rich - having investors on the board trying to make money for themselves would be crazy.

Exactly, if we assume Altman wanting to pursue commercialization at the cost of safety was the issue, the board did its job by advancing its mandate of "AI for the benefit of humanity" although not sure why they went with the nuclear option.

Re: Greg Brockman quits OpenAI

#634

Post by CapitalistCartr »

ascorbic wrote: post #611
leptons wrote: post #542

Earlier quoted context omitted.

Steve Jobs would be nothing without Wozniak to design something people wanted.

But the point is that Woz was replacable, because he was replaced. Jobs on the other hand was replaced and the company nearly died, he came back and turned it around. Of course it only because a trillion dollar company after Tim Apple took over, which I guess just shows that nobody is irreplacable.

Tim Cook?

Re: Greg Brockman quits OpenAI

#635

Post by voitvod »

Paul-Craft wrote: post #575
eslaught wrote: post #372

Earlier quoted context omitted.

This doesn't make any sense. If it was a disagreement, they could have gone the "quiet" route and just made no substantive comment in the press release. But they made accusations that are specific enough to be legally enforceable if they're wrong, and in an official statement no less. If their case isn't 100% rock solid, they just handed Sam a lawsuit that he's virtually guaranteed to win.

I agree. None of this adds up. The only thing that makes any sense, given OpenAI has any sense and self interest at all, is that the reason they let Altman go may have even been bigger than even what they were saying, and that there was some lack of candor in his communications with the board. Otherwise, you don't make an announcement like that 30 minutes before markets close on a Friday.

[flagged]

Re: Greg Brockman quits OpenAI

#636

Post by spangry »

visarga wrote: post #616
mlyle wrote: post #456

Earlier quoted context omitted.

Clearly there's a difference, because the architectures we have don't know how to persist information or further train. Without persistence outside of the context window, they can't even maintain a dynamic, stable higher level goal. Whether you can bolt something small to these architectures for persistence and do some small things and get AGI is an open question, but what we have is clearly insufficient by design. I…

But context windows got to 100K now, RAG systems are everywhere, and we can cheaply fine-tune LoRAs for a price similar with inferencing, maybe 3x more expensive per token. A memory hierarchy made of LoRA -> Context -> RAG could be "all you need". My beef with RAG is that it doesn't match on information that is not explicit in the text, so "the fourth word of this phrase" won't embed like the word "of", or "Bruce Wil…

I've been wondering along similar lines, although I am for all intents and purposes here a layman so apologies if the following is nonsensical.

I feel there are potential parallels between RAG and how human memory works. When we humans are prompted, I suspect we engage in some sort of relevant memory retrieval process and the retrieved memories are packaged up and factored in to our mental processing triggered by the prompt. This seems similar to RAG, where my understanding is that some sort of semantic search is conducted over a database of embeddings (essentially, "relevant memories") and then shoved into the prompt as additional context. Bigger context window allows for more "memories" to contextualise/inform the model's answer.

I've been wondering three things: (1) are previous user prompts and model answers also converted to embeddings and stored in the embedding database, as new "memories", essentially making the model "smarter" as it accumulates more "experiences" (2) could these "memories" be stored alongside a salience score of some kind that increases the chance of retrieval (with the salience score probably some composite of recency and perhaps degree of positive feedback from the original user?) (3) could you take these new "memories" and use them to incrementally retrain the model for, say, 8 hours every night? :)

Edit: And if you did (3), would that mean even with a temperature set at 0 the model might output one response to a prompt today, and a different response to an identical prompt tomorrow, due to the additional "experience" it has accumulated?

Re: Greg Brockman quits OpenAI

#637

Post by tsimionescu »

rich_sasha wrote: post #630
tsimionescu wrote: post #579

Earlier quoted context omitted.

No, the onus would be on Sam Altman to prove that the statement was materially false, AND intended to slander him, AND actually succeeded in affecting his reputation. When you're a public person, the bar for winning a defamation case is very high.

I don't know. The board statement, peeling away the pleasantries, says he lied to the board repeatedly. That's a very serious accusation. I don't know how US law works here, but in the UK you can sue and win over defamation for far milder infractions.

Even in the UK, if you sue, it is on you to prove that you didn't lie, not on the person you're sueing to prove that you did.

Also, as long as you are a public person, defamation has a very high bar in the USA. It is not enough to for the statement to be false, you have to actually prove that the person you're accusing of defamation knew it was false and intended it to hurt you.

Note that this is different from an accusation of perjury. They did not accuse Sam Altman of performing illegal acts. If they had, things would have been very different. As it stands, they simply said that he hasn't been truthful to them, which it would be very hard to prove is false.

Re: Greg Brockman quits OpenAI

#638

Post by tsimionescu »

Rastonbury wrote: post #632
bananapub wrote: post #629

Earlier quoted context omitted.

> as Jason @ all-in said lol > "that's just crazy". why is it crazy? the purpose of OpenAI is not to make investors rich - having investors on the board trying to make money for themselves would be crazy.

Exactly, if we assume Altman wanting to pursue commercialization at the cost of safety was the issue, the board did its job by advancing its mandate of "AI for the benefit of humanity" although not sure why they went with the nuclear option.

Very true.

Though I would go further than that: if that is indeed the reason, the board has proven themselves very much incompetent. It would be quite incompetent to invite this type of shadow of scandal for something that was a fundamentally reasonable disagreement.

Re: Greg Brockman quits OpenAI

#639

Post by justcool393 »

jumploops wrote: post #10

Greg is the one who announced GPT-4. Sam enabled Greg and vice-versa. The next AI winter may have just begun...

we can only hope

i'm sick and tired of everyone sticking a chatbot on random crap that doesn't need it and has no reason to ever need it. it also made HN a lot less interesting to read

Re: Greg Brockman quits OpenAI

#640

Post by rich_sasha »

tsimionescu wrote: post #637
rich_sasha wrote: post #630

Earlier quoted context omitted.

I don't know. The board statement, peeling away the pleasantries, says he lied to the board repeatedly. That's a very serious accusation. I don't know how US law works here, but in the UK you can sue and win over defamation for far milder infractions.

Even in the UK, if you sue, it is on you to prove that you didn't lie, not on the person you're sueing to prove that you did. Also, as long as you are a public person, defamation has a very high bar in the USA. It is not enough to for the statement to be false, you have to actually prove that the person you're accusing of defamation knew it was false and intended it to hurt you. Note that this is different from an ac…

In a specific case, perhaps. But surely, I can't go out, make a broad statement like, "XYZ is a liar and fornicator" and leave it there. And how would XYZ go around proving they are not a liar and fornicator? Talk to everyone in the world and get them to confirm they were not lied to or sexually involved?

Surely, at some level, you can be sued for making unfounded remarks. But then IANAL so, meh.

Post reply on HN