2) Where is the board? At a bare minimum, issue a public statement that you have full faith in the new CEO and the leadership team, are taking decisive action to stabilize the situation, and have a plan to move the company forward once stabilized.
The only thing I've read about Shear is he is pro-slowing AI development and pro-Yudkowsky's doomer worldview on AI. That might not be a pill the company is ready to swallow.
https://x.com/drtechlash/status/1726507930026139651
> I specifically say I’m in favor of slowing down, which is sort of like pausing except it’s slowing down.
> If we’re at a speed of 10 right now, a pause is reducing to 0. I think we should aim for a 1-2 instead.
> - Emmett Shear Sept 16, 2023
No explanation beyond "he tried to give two people the same project
the "Killing the company would be consistent with the companies mission" line in the boards statement
Adam having a huge conflict of interest
Emmet wanting to go from a "10" to a "1-2"
I'm either way off, or I've had too much internet for the weekend.
It would also comport with Ilya's regret about the situation: perhaps he wanted to slow things down, board members convinced him Sam's ouster was the way to do it, but then it has actually unfolded such that development of dangerous AI/AGI/ASI might accelerate at Microsoft while weakening OpenAI's own ability to modulate the pace of development.
[1]: Given all the very public media brinkmanship, I'm not so quick to assume reports like these two explanations are true. E.g. the "Sama is returning with demands!" stories were obviously "planted" by people who were trying to exert pressure on the negotiations; would be interested to have more evidence that Ilya's explanations were actually this sloppy.
Even worse, if we don't have near constant updates, we might realize this is not all that important in the end and move on to other news items!
I know, I know, I shouldn't jest when this could have grave consequences like changing which uri your api endpoint is pointing to.
A statement from the CEO/the board is a standard descalation.
Haven't we gotten statements from them? The complaint seems to be that we want statements from them every day (or more) now.
The board has not given a statement besides the original firing of Sam Altman that kicked the whole thing off.
"All PR is good PR" is a meme for a reason. Many cultures thrive on dysfunction, particularly the kind that calls attention to themselves.
You're saying we're in a less attention-seeking culture today than in pre-social media times?
Maybe "attention seeking" isn't the right way to look at this. Getting bad press always does reputational damage while giving you notoriety, and I think GP's suggestion that the balance between them has changed is compelling.
In an environment with limited connectivity, it's much more difficult for people to learn you even exist to do business with. So that notoriety component has much more value, and it often nets out in your favor.
In a highly connected environment, it's easier to reach potential customers, so the notoriety component has less value. Additionally, people have access to search engines, so the reputational damage becomes more lasting; potential customers who didn't even hear about the bad press at the time might search your name and find it. They may not have even been looking for it, they might've searched your name to find you website (whereas before they would have needed to intentionally visit a library and look through the catalog to come across an old story). So it becomes much less likely to net out in your favor.
PSA: If you or your culture is dysfunctional and thriving - think about how much more you'll thrive without the dysfunction! (Brought to you by the Ad Council.)
Unless you're TNT, cause they "know drama"
I’m done pretending they’re important. It’s a lie they and the boards have sold us and investors. The real meat of a company is who their smartest people are, and how much the company enables those people.
Pretty easy to see the difference if you consider between a company full of smart people who actually make things vs a company full of CEOs, which one will do better.
For the record, I don't think it's true. I think it was a power play, and a failed coup at that. But it's about as substantiated as the "serious" hypotheses being mooted in the media. And it's more fun.
As they learn to construct better and more coherent conceptual chains, something interesting must be happening internally.
There is nothing special about human intelligence threshold.
It can be surpassed by many different models.
Could you at least elaborate what they are “not”? Surelly you are not having a problem with “LLMs predict language”?
I smell a fallacy. Parent has moved from something you can parse as "LLMs predict a representation of concepts" to "LLMs construct concepts". Yuh, if LLMs "construct concepts", then we have conceptual thought in a machine, which certainly looks interesting. But it doesn't follow from the initial statement.
I think trying to model the world based on a single projection won't get you very far.
I personally think a far more fundamental change is necessary to reach AGI.
Your comment sounds like a rhetoric way to say that GPT is in the same class as autocomplete and that what autocomplete does sets some kind of ceiling to what IO functions that work a couple of bytes at a time can do.
It is not evident to me that that is true.
Are there any papers testing how good humans are at predicting the next word?
I presume us humans fail badly:
1. as the variance in input gets higher?
2. Poor at regurgitating common texts (e.g. I couldn't complete a known poem).
3. When context starts to get more specific (majority of people couldn't complete JSON)?
https://nonint.com/2023/06/10/the-it-in-ai-models-is-the-dat... The ultimate model, in his (author's) sense, would suss out all patterns and then patterns among those patterns and so on, so that it delivers on compute and compression efficiency.
To achieve compute and compression efficiency, it means LLM models have to cluster all similar patterns together and deduplicate them. This also means successively levels of pattern recognition to be done i.e. patterns among patterns among patterns and so on , so as to do the deduplication across all hierarchy it is constructed. Full trees or hierarchies won't get deduplicated but relevant regions / portions of those trees will, which implies fusing together in ideas space. This means root levels will be the most abstract patterns. This representation also means appropriate cross-pollination among different fields of studies further increasing effectiveness.
This reminds me of a point which my electronics professor made on why making transistors smaller has all the benefits and only few disadvantages. Think of these patterns as transistors. The more deduplicated and closely packed they are, the more beneficial they will be. Of course, this "packing together" is happening in mathematical space.
Another thing which patterns among patterns among patterns reminds me of homotopies. This brilliant video by PBS Infinite Series is amazing. As I can see, compressing homotopies is what LLMs do, replace homotopies with patterns. https://www.youtube.com/watch?v=N7wNWQ4aTLQ
From this, we get comedy. A funny statement is one that ends in an unpredictable manner and surprises the listener brain because it doesn't have the meaning of that one already calculated, and hence why it can take a while to "get the joke"
https://www.nytimes.com/2023/03/08/opinion/noam-chomsky-chat...
I absolutely disagree in the strongest terms possible.
I get that this is a "just for fun" hypothesis, which is why I have just for fun questions like what incentive does anyone have to keep clearly observed ai risk a secret during such a public situation?
But if there's one thing that seems very easy to discern about Ilya, it's that he fully believes that when it comes to AI safety and alignment, the buck must stop with him. Giving that control over to government bureaucracy/gerontocracy would be unacceptable. And who knows, maybe he's right.
Yes, actually. This is overwhelmingly true for most people. At the end of the day, we all fear being alone. I imagine that fear is, at least in part, what drives these kinds of long-term "existential worries," the fear of a universe without other people in it, but now Ilya is facing the much more immediate threat of social ostracism with significantly higher certainty and decidedly within his own lifetime. Emotionally, that must take precedence.
His existential worries are less important than OpenAI existing, and him having something to work on and worry about.
In fact, Ilya may have worried more about the continued existence of OpenAI than Sam after he was fired, which looked instantly like a: "I am taking my ball and going home to Microsoft.". If Sam cared so much about OpenAI, he could have quietly accepted his resignation and help find a replacement.
Also, Anna Brockman had a meeting with Ilya where she cried and pleaded. Even though he stands by his decision, he may ultimately still regret it, and the hurt and damage it caused.
* Current-gen AI is really good at tricking laypeople into believing it could be sentient
* "Next-gen" AI (which, theoretically, Ilya et al may have previewed if they've begun training GPT-5, etc) will be really good at tricking experts into believing it could be sentient
* Next-next-gen AI may as well be sentient for all intents and purposes (if it quacks like a duck)
(NB, to "trick" here ascribes a mechanical result from people using technology, not an intent from said technology)
Uh, or investors and customers will? Yes, people are going to speculate, as you point out, which is not good.
> we might realize this is not all that important in the end and move on to other news items!
It's important to some of us.
News
Company which does research and doesn't care about money makes a decision to do something which aligns with research and not caring about money.
From the OpenAI website...
"it may be difficult to know what role money will play in a post-AGI world"
Big tech co makes a move which sends its stock to an all time high. Creates research team.
Seems like there could be a "The Martian" meme here... we're going to Twitter the sh* out of this.
If they had openly given literally any imaginable reason to fire Sam Altman, the ratio of employees threatening to quit wouldn't be as high as 95% right now.
I think he is correct, being the CEO twitch is a position known by no one in many places, e.g. how many developers/users in China even heard of Twitch? Being the CEO of OpenAI is a completely different story, it is a whole new level he can leverage in the years to come.
Like taking a sword to the gut.
If everything goes well, he can claim that he is the man behind all these to reunite the OpenAI team. If something goes wrong, well, no one is going to blame him, the board screwed the entire business. He is more like a emergency room doctor who failed to save a poor dude who just intentionally shot himself in the head with a shotgun.
It's now one day later and Altman is back as CEO - what can Emmett Shear claim exactly?
- helped to stabilized the situation
- the first to propose the idea of independent investigation into the matter
- sided with impacted employees all the way through such difficult moments
- supported a smooth transition period
depending on how he reacted when 90% employees signed that letter asking for altman's return, don't be too surprised if he claims to be guy how helped to push for that as well.
However, the OpenAI board has no such obligation. Their duty is to ensure that the human race stays safe from AI. They've done their best to do that ;-)