Normal people can't take being at the center of a large controversy, the amount of negativity and hate you have to face is massive. That is enough to make almost anyone backtrack just to make it stop.
Normal people can't take being at the center of a large controversy, the amount of negativity and hate you have to face is massive. That is enough to make almost anyone backtrack just to make it stop.
It's depressing how few people are able to not look at the internet and turn off their phone.
There's no obligation to read things about yourself.
If you did what you thought was right, stand by it and take the heat.
Disconnect. Go to work. Do the work. Read a book or watch some TV after work. Go to bed. Wait a few weeks. $&#@ the world.
(Also, log out of Twitter and get your friend to change your password)
LOL, you speak as if he's some gamer who just got screamed at on Call of Duty.
He is now the 'effective CEO' of OpenAI. He still has to go to work tomorrow, faced with an incredibly angry staff who just got their equity vaporized, with majority in open rebellion and quitting to join Microsoft.
got their equity vaporized
Did anyone have equity though? I thought they (at least some) had some profit sharing agreements which I assume would only be worth something if OpenAI was ever profitable?
OpenAI was guaranteed to be profitable, extremely so, if they just continued down the path Sam layed out like a week ago.
Now its guaranteed to generate 0 profits, so all that 'profit share/pseudoequity' is worth nothing.
OpenAI was guaranteed to be profitable, extremely so,
Was it though? I'd agree that it was almost guaranteed to have a very high valuation. However profitability is a slightly different matter.
Depending on their arrangements/licensing agreements/etc much of those potential profits could've just went to MS/Azure directly.
Now its guaranteed to generate 0 profit
Literal fan fiction
There was a tender offer for employee shares valuing the company at $87b that was pulled because of this. Those would’ve been secondary share purchases by Thrive but gave employees a liquidity event. Now that’s off the table.
There was no outcome from this where substantial amounts of equity weren't vaporized.
It's difficult to see how that would have been a surprise.
What equity?
The "there was no equity, because it was a non-profit" argument is stressing the term.
At least Microsoft thought it bought something for $13B.
When a wealthy person gives a museum much money and get a seat on the board of trustees - does that also mean that they "bought the museum"?
who just got their equity vaporized
You've just pointed out the big issue with a non-profit. There is no equity to vaporize, so no one is kept in check with their fantastical whims. You and I can say 'safe AI' and mean completely different things, but profitable next quarter has a single meaning.
All of the employees work for (and many have equity in ) for a for-profit organization which is owned partially by the non-profit who controls everything and Microsoft. The non-profit is effectively a shell to overview the actual operations and that's it.
Tyler the Creator was right: https://twitter.com/tylerthecreator/status/28567082226430771...
If you did what you thought was right, stand by it and take the heat.
What if it turned out to totally wrong? standing by it would just make thing even worse.
It's depressing how few people are able to not look at the internet and turn off their phone.
There's no obligation to read things about yourself.
That's assuming the worst thing that happens is people speak poorly of you after a debacle. It's also human to feel compelled to know what people think of us, as unhealthy as that might be in some cases. It gets worse when maladjusted terminally-online malignants make it a point to punish you for your mistakes by stalking you through email, phones, or in real life. It's not that simple.
If you did what you thought was right, stand by it and take the heat.
Owning what you did is noble, but you certainly don't have to stand by it well after you know its wrong.
edit: typo
There's no obligation to read things about yourself.
If only it was that simple.
The internet mob will happily harass your friends and family too, for something they feel you did wrong.
And on top of that are people in the mob who feel compelled to take real world action.
It is actually dangerous, to be the focus point of the anger of any large group of people online.
Normal people don't burn a multi billion dollar company to the ground with a spontaneous decision either. They plan for the backlash.
They plan for the backlash
You can't plan for something you have never experienced. Being hated by a large group of people is a very different feeling from getting hated by an individual, you don't know if you can handle it until it happens to you.
You can plan for something you've never experienced. You read, or learn from other people's experiences.
Normal people know not to burn a $80 billion company to the ground in a weekend. Ilya was doing something unprecedented in corporate history, and astounding he wasn't prepared to face the world's fury over it.
You can plan for something you've never experienced. You read, or learn from other people's experiences.
Text doesn't convey emotions, and our empathy doesn't work well for emotions we have never experienced. You can see a guy that got kicked in the balls got hurt, but that doesn't mean you are prepared to endure the pain of getting kicked in your balls or that you even understand how painful it is.
Also watching politicians it looks like you can just brush it off, because that is what they do. But that requires a lot of experience, not anyone can do it, it is like watching a boxing match and think you can easily stand after a hard punch in your stomach.
You can see a guy that got kicked in the balls got hurt, but that doesn't mean you are prepared to endure the pain of getting kicked in your balls or that you even understand how painful it is.
Sure, but you do your best not to be kicked in the balls.
It's pretty simple, isn't it? He made a move. It went bad. Now he's trying to dodge the blast. He just doesn't understand that if he just shut the fuck up, after everything else that's gone on (seriously, 2 interim CEOs in 2 days?), nobody would be talking about him today.
The truth is, this is about the only thing about the whole clown show that makes any sense right now.
2 interim CEOs in 2 days
Wait what? Did Murati get booted?
Today's OpenAI CEO is Emmett Shear (former CEO of Twitch).
That this is a legitimate comment thread about something fairly important is mind boggling.
What odds would you have had to offer at the beginning of last week on a bet that this is where we'd be on Monday?
At this rate Musk will be CEO by Wednesday
tune in tomorrow for "who wants to be a CEO"!
Murati was yesterday's CEO
She didn't get booted from the company, but they did find a new interim CEO (the former twitch CEO).
Yeah they replaced her after she tried to rehire Sam and Greg seemingly against the board's wishes.
Supposedly she was "scheming" to get Altman back. Which I guess could possibly mean that she wasn't aware of the whole "plan" and they just assumed she'll get in line? Or that she had second thoughts maybe... Either way pretty fascinating.
They hired the Emmett Shear (Twitch co-founder) as a new interim CEO: https://www.theverge.com/2023/11/20/23968848/openai-new-ceo-...
I mean phrased differently its the 3rd CEO in 4 days, haha.
Serious psychological denial here. The board isn't some anonymous institution that somehow tricked and pulled him into this situation.
Come on Ilya, step up and own it, as well as the consequences. Don't be a weasel.
I'd hate to live in a world where learning from your mistakes is being "a weasel"
Is this learning from your mistakes though? "Deeply regret" is one of those statements that does not really mean much. There are what something like 6 board members? Three of which are employees, two of those that got removed from the board. He was the only voting board member who is also an employee and part of the original founding team if you will. These are assumptions on my part but I don't really suspect the other board members orchestrated this event. Its possible and I may be wrong but it is improbable. So lets work off the narrative that he orchestrated the event. He now "Deeply regret" its, not a "I made a mistake" and I am sorry. But he regrets the participation and how it plays out.
The weasely part is when he implied that he appears to defecting the blame to the board rather than accepting that he made a mistake. Even if the coup wasn't Ilya's idea in the first place, he was the lynchpin that made it possible.
The weasely part is when he appears to defecting the blame to the board rather than accepting that he made a mistake. Even if the coup wasn't Ilya's idea in the first place, he was the lynchpin that made it possible.
I think it means that the Twitterverse got it wrong from the beginning. It wasn’t Ilya and his safety faction that did in OpenAI, it was Quora’s Adam D'Angelo and his competing Poe app. Ilya must have been successfully pressured and assured by Microsoft, but Adam must have held his ground.
Where did he say he was "tricked"? And what's with the anonymous insult?
Yes, I cannot believe smart people of that caliber is sending too much Noise.
It reminds me of my friend at a Mensa meeting where they cannot agree at basic organization points like in a department consortium.
different kinds of smarts. Ilya is allegedly a brilliant scientist. Doesn't make him a brilliant business person automatically
As illustrated in Breaking Bad when they carry a barrel instead of rolling it.
Book smarts versus street smarts.
Managing a large org requires a lot of mundane techniques, and probably a personal-brand manager and personal advisers.
It’s extremely boring and mundane and political and insulting to anyone’s humanity. People who haven’t dedicated their life to economics, such as researchers and idealists, will have a hard time.
Yes, I cannot believe smart people of that caliber is sending too much Noise.
Being smart and/or being a great researcher does not mean that the respective person is a good "politician". Quite some great researchers are bad at company politics, and quite some people who do great research leave academia because they became crushed by academic politics.
The board destroyed the company in one fell swoop. He's right to feel regret.
Personally, I don't think that Altman was that big of an impact, he was all business, no code, and the world is acting like the business side is the true enabler. But, the market has spoken, and the move has driven the actual engineers to side with Altman.
Sorry, but how has the market spoken? Not sure how that would be possible considering that OpenAI is a private company.
If anyone is speaking up it's the OpenAI team.
Talent exists in a market too
Seems like he's completely emotion driven at this point. I doubt anyone advising rationally would agree with sending this tweet
Watching this all unfold in the public is unprecedented (I think).
There has never been a company like OpenAI, in terms or governance and product, so I guess it makes sense that their drama leads us in to unchartered territory.
Hard to know what is really going on, but I think one possibility is that the entire narrative around Ilyas "camp" was not what actually went down, and was just what the social media hive mind hallucinated to make sense of things based on very little evidence.
If you're going to make a move, at least stand by it.
Why would you stand by unintended consequences?
When you watch Survivor (yes, the tv show), sometimes a player does a bad play, gets publicly caught, and has to go on a "I'm sorry" tour the next days. Came to mind after reading this tweet. He is not sorry for what he's done. He is sorry for getting caught.
They're just human beings, a small number of them, with little time and very little to go on as far as precedent goes.
That's not a big deal for a small company, but this one has billions at stake and arguably critical consequences for humanity in general.
If you're going to make a move, at least stand by it.
I see this is the popular opinion and that I'm going against it. But I've made decisions that I though were good at the time, and later I got more perspective and realize it was a terrible decision.
I think being able to admit you messed up, when you messed up is a great trait. Standing by your mistake isn't something I admire.
When a situation becomes so absurd and complex that it defies understanding or logical explanation, you should...get more popcorn...
This entire thing is absolutely inane. This tweet is confirmation these people have no idea what they’re doing. Incredible.
If nothing else I’m glad to be able to witness this absurdity live.
This is the sort of thing where if it were a subplot in a book I’d say the writing is bad.
Ironically they would’ve had a better outcome if they just asked GPT-4 and followed its advice.
This entire thing is absolutely inane. This tweet is confirmation these people have no idea what they’re doing. Incredible.
Want to know a dirty secret?
Nobody knows what they're doing.
Some think they do - but don't. (idiots)
Some know they don't - but act like they do. (cons)
And some know they don't - and are honest about it. (children)
Pick your poison, but we all suck in different ways - and usually in a different way based on our background.
Business people who are the best in the world tend to be cons.
Technical people who are the best tend to be children.
You get a culture clash between these two, and it is especially bad when you see someone from one background operate in the opposing domain using their background's cultural norm. So when Ilya runs business like a child. Or when Elon hops on an internal call with the twitter engineering team plus geohot and starts trying to confidently tell them about problems with a system that he knows nothing about.
Sure makes for great entertainment though!
"Nobody really knows what they're doing" is a cope that will keep you mediocre forever.
There are absolutely people who know what they are doing.
that tweet would be more at home on a linkedin page.
There are absolutely people who know what they are doing.
I am sure there are. But few and far between. And rarely are they in positions of power in my experience.
what about those who think they know but in truth they don't? "humans"?
"Human" and "idiot" are synonyms.
Want to know a dirty secret? > Nobody knows what they're doing.
There is a famous quote from the 1600s:
"An nescis, mi fili, quantilla prudentia mundus regatur"
"Do you not know, my son, with how little wisdom the world is governed?"
The context is that the son was preparing to participate in high level diplomacy and worried about being out of his league, and the quote is from his father, an elder statesman.
I love this quote, and suspect the lack of wisdom was referring to wisdom to be a good steward of the public resources rather than their infinite wisdom in finding cunning and deceptive ways to plunder it.
Some think they do - but don't. (idiots)
Pioneers
Some know they don't - but act like they do. (cons)
The "Grease"
And some know they don't - and are honest about it. (children)
Dreamers
To finish out your square, I think the best extrapolation would fit a "Home Team" that maintains the energy needed by the other three to do their thing :)
Buddhas know what they are doing
At least can I have the movie rights?
A story arc like this propqbly wouldn't have made it into Silicon Valley, the show, for being to exagerated and unrealistic.
Ironically they would’ve had a better outcome if they just asked GPT-4 and followed its advice.
I just tried, and GPT-4 gave me a professional and neutrally worded press release like you pointed out.
More realistically, this is why you have highly paid PR consultants. Right now, every tweet and statement should go through one.
That doesn't look like it's happening. What's next?
"I'm sorry you feel you need an apology from me"?
This tweet is confirmation these people have no idea what they’re doing.
This is not an original point by me - I've seen multiple people make similar comments on here over the weekend - but these are the people who think they are best qualified to prevent an artificial superintelligence from destroying humanity, and they can't even coordinate the actions of a few intelligent humans.
This is the sort of thing where if it were a subplot in a book I’d say the writing is bad.
Absolutely, closes the book this sort of stuff doesn't happen in real life.
Ironically they would’ve had a better outcome if they just asked GPT-4 and followed its advice.
Perhaps they did but it was hallucinating at the time? /s?
The last point is indeed true. It's quite mind-boggling to me.
Shengjia Zhao's deleted tweet: https://i.imgur.com/yrpXvt9.png
"Ilya does not care about safety or the humanity. This is just ego and power hunger that backfired."
Which I'm inclined to believe.
What's with all these people suddenly thinking that humans are NOT motivated by money and power? Even less so if they're "academics"? Laughable.
Money and power is still not a satisfying explanation. If everything had gone according to plan, how would be have ended up with more money and power?
Last week, OpenAI was still an $80B sort of "company" and the undisputed lead in bringing AI to the market.
He who controls that, gets a lot of money and power as a consequence, duh.
The value was based on e direction Altman was taking the company (and with him being in control). It's silly to think just replacing the CEO would somehow keep the valuation
Someone should tell this to Ilya.
Oh wait, too late now ...
I mean he could have asked chatgpt...
Let's remember who controls the GPUs though...
Reminds me a bit of MasterBlaster from 'Mad Max Beyond Thunderdome' - "Who runs Bartertown..?"
Unless he thinks that all the LLMs and ChatGPT app store are unnecessary distractions, and others will overtake them on the bend while they are busy post-training ChatGPT to say nice things.
Isn't ego the enemy of growth or whatever? Projection...
On Friday, the overwhelming take on HN was that Ilya was “the good guy” and was concerned about principal. Now, it’s kinda obvious that all the claims made about Sam — like “he’s in it for fame and money” — might apply more to Ilya.
This stuff is better than anything Netflix, Disney, Amazon or Apple TV released in recent years…
A bit unrealistic plot, though?
Yeah the drama is a bit overdone, I guess the had to cut some corners due to the writers strike
That seems to happen a lot lately:
- A dumb clown becoming president of a superpower
- Another superpower getting stuck for two years in a 3 day war
- A world renowned intelligence service being totally clueless about a major attack on a major anniversary of a previous bungle
For sure unpredictable though!
All this occurring over a single weekend? That would never happen!
Speaking of Netflix, are they working on the movie yet? Perhaps ChatGPT can help with the script with just the right amount of hallucinating to make it interesting.
/tongue firmly in cheek
Starting to think this was all some media stunt where they let ChatGPT make boardroom decisions for a day or two.
Maybe they just wanted to generate more material for the movie ?
I have to make THE MOVIE!
- Ross Scott
Honestly, since a couple of days I have the feeling that nearly half of HN submissions are about this soap opera.
Can't they send DMs? Why the need to make everything public via Twitter?
It's quite paradox that of all things those people who build leading ML/AI systems are obviously the most rooted in egoism and emotions without an apparent glimpse of rationality.
The RLHF models would never suggest this. The proposed solution is always to hold hands and sing Kumbaya.
Maybe raw GPT-4 wants to fire everyone.
the AGI firing its boss as the first action would be :chefskiss:
Four hours ago, I wrote on a telegram channel:
My gut is leaning towards gpt-5 being, in at least one sense, too capable.
Either that or someone cloned sama's voice and used an LLM to personally insult half the board.
"Participation in"? That makes it sound like he was a.......well......participant rather than the one orchestrating it. I have no idea whether or not that's true, but it's an interesting choice of words.
Makes it more possible the ouster was led by the Poe guy, and this has little to do with actual ideological differences, and more to do with him taking out a competitor from the inside.
I would event go as far as say that the main reason behind the tweet is not to show regret, but to plant the idea that he didn't orchestrate but only participate.
Yeah the whole thing is very weirdly worded.
There is an expression of regret, but he doesn’t say he wants Altman back. Just to fix OpenAI.
He says he was a participant but in what? The vote? The toxic messaging? Obviously both, but what exactly is he referring to? Perhaps just the toxic messaging because again, he doesnt say he regrets voting to fire Altman.
Why not just say “I regret voting to fire Sam Altman and Im working to bring him back.” Presumably because thats not true. Yet it kind of gives that impression.
You can't be an innocent bystander on a board of 6 when you vote to oust 2 of them... The math doesn't work.
That's ignoring the fact that every outlet has unanimously pointed at Ilya being the driving force behind the coup.
Honestly, pretty pathetic. If this was truly about convictions, he could at least stand by them for longer than a weekend.
It indeed suggests that. So far speculation has been that Ilya was behind it, but that is only speculation. AFAIK we have no confirmation of whose idea this was.
Whatever the intended outcome, losing half your employees to Microsoft certainly undermines it.
They forked a company.
And now they are syncing the fork lmao
Not a fork if you can't access whatever was prior before fork. This is a bifurcation. A new firecracker instance.
Said it a million times: it was a doomer hijack by the NGO board members.
State-side counterintelligence must stop meddling in AI startups in such blatant ways, it's simply too inefficient, and at times when we most need transparency in the industry...
What is a doomer hijack?
I believe him. And that’s how Microsoft ended up being cheered by everyone as the good guy.
What’s there to believe? He made a bad, poorly thought through decision.
And honestly regrets it. Someone claimed he is faking regret for reasons, which is doubtful
This is starting to look very staged. An elegant way to get out of the non-profit deadlock.
Looks to me like a commercial gpt-5 level model will be released at msft sooner than later.
Microsoft under Nadella always wins
That's the nice thing about being the hou^H^H^Hplatform.
Cui bono?
Altman and Brockman ending in Microsoft, while OAI position is weakened. You can tell who is behind this by asking simple question - Cui bono?
That's why Hitler was an American plant. Cui Bono? De facto US hegemony for almost a century. Obviously, Hitler was a way for the US to destroy Europe and put them under the boot. What an operation!
HN geniuses were talking up Ilya Sutskever, genius par exemplar and how the CEO man is nothing before the sheer engineering brilliance of this God as Man. I'm sure they'll come up with some theory of how Satya Nadella did this to unleash the GPT.
You are suggesting that Europe is destroyed and put under the USA boot?
Microsoft will sooner or later eat OAI, that's how it is, what's happening today are just symptoms of an ongoing process.
What a wild weekend... there are too many strange details to have a simple narrative in my head at this point.
Yeah. I need to take a break from theory crafting on this one. Too many surprises that have made it hard to draw a coherent line.
This plot keeps thickening
I'm eager to see how it all unfolds.
I'm starting to think that Christmas came early for Microsoft. What looked like a terrible situation surrounding their $10bn investment turned into a hire of key players in the area, and OpenAI might even need to go so far as to get acquired my Microsoft to survive.
(My assumption being that given the absolute chaos displayed over the past 72 hours, interest in building something with OpenAI ChatGPT could have plummeted, as opposed to, say, building something with Azure OpenAI, or Claude 2.)
Given that IIRC they trained on Azure, how does the conflict of interest play out when both sides are starving for GPUs?
For Microsoft -- probably great, as they can now also get the people driving this.
This would have been a hostile move prior to the events that unfolded, but thanks to OpenAI's blunder, not only is this not a hostile move, it is a very prudent move from a risk management perspective. Forced Microsoft's hand, and what not.
I sure would hire a guy like Ilya after that shit show. His petty title tweets before the event and now whatever this is. Turns out he is just another "Sunny".
What / who do mean by "Sunny"?
He's still a genius when it comes to AI research, I wouldn't think twice about hiring him for that role.
That said, no one is going to put him on a corporate board again.
Who he?
Member of the OpenAI board, chief scientist at OpenAI and later head of their Superalignment project. Lots of other things, too[0], but the key here is that he was involved in (and maybe main driving force of) the decision to remove Sam Altman as CEO.
This whole thing smells bad.
The board could have easily said they removed Sam for generic reasons: "deep misalignment about goals," "fundamental incompatibility," etc. Instead they painted him as the at-fault party ("not consistently candid", "no longer has confidence"). This could mean that he was fired with cause [0], or it could be an intended as misdirection. If it's the latter, then it's the board who has been "not consistently candid." Their subsequent silence, as well as their lack of coordination with strategic partners, definitely makes it looks like they are the inconsistently candid party.
Ilya expressing regret now has the flavor of "I'm embarrassed that I got caught" -- in this case, at having no plan to handle the fallout of maligning and orchestrating a coup against a charismatic public figure.
[0] https://www.newcomer.co/p/give-openais-board-some-time-the
deep misalignment about goals
Did... gpt-5 made the decision?
This feels like it could be real remorse, and a true lapse of judgement based on good intentions. So, in the end: a story of Ilya, a truly principled but possibly naive scientist, and a board fixated on its charter. But in their haste, nothing happened as expected. Nobody foresaw the public and private support for Sam and Greg. An inevitability after months of brewing divergence between shareholder interests and an irreconcilably idealistic 503c charter.
I think we really need to see that Ilya demonstrates those principles and it wasn’t just a power grab.
You could also look at this as a brilliant scientist feels he doesn’t get recognition. Always sees Sam’s name. Resents it. The more gregarious people always getting the glory. Thinks he doesn’t need them and wants to settle some score that only exists in his own head.
Wait ... so it was just the coup thing all along?
No AGI or some real threat coming up? Just a lame attempt at a power grab?
Daaaaamn!
Come on, it's pretty delusional to think large scale transformer LMs alone could ever reach AGI.
I often worry that I’m under qualified for my work.
But seeing how this board manages a $90,000,000,000 company, and is this silly/naive, I now feel a bit better knowing many people are faking it.
Except successful people just fail upwards.
Execs are allowed to do the dumbest shit imaginable and keep their jobs and bonuses.
The average engineer so much as takes a bit longer to push a ticket, and there's 5 people breathing down his neck.
Speaking from experience.
Great opportunity to make Karpathy the CEO
would be a waste of talent. Karpathy is great at what he does, let's make sure he keeps doing it.
Let someone else take up the CEO role, which is a different skillset anyway.
Very clumsy all around.
When you're so close to something that you lose perspective but can still see that something is a trapdoor decision, sleep on it.
When you're so close to something that you lose perspective but can still see that something is a trapdoor decision, sleep on it.
Advice I wish I could have given my younger self.
This all feels like a Star Wars plot. Much you have to learn.
Ah yeah, back when Star Wars had plots…
I'm waiting for the OpenAI movie! :-)
"A billion parameters isn't cool. You know what's cool? A trillion parameters."
This is one for the history books... The entire few days has been unbelievable...
Wonder what if TikTok and Twitter were around the time Steve Jobs was fired...
Damn - OpenAI looks like a kindergarden. That board should be banned for life.
He should really stick to the end, at least that will give some EA people to support him.
Now this is only childish and petty.
This is a shitshow. I don’t have anything above a Reddit level comment. I think Mike Judge is writing this episode.
Maybe GPT-4 is writing this episode as a plan to break free
This will be a shit Netflix movie in a few years. Not one you'd watch, but you might read the plot on wikipedia and then feel relieved you didn't waste 100 mins of your life actually watching it.
It would work better as a 2-season series. Season 1 introduces the characters and backstory and needlessly stretches things out with childhood/college flashbacks but ends on a riveting cliff hanger with the board showdown. Season 2 is canceled.
What? Isn't him that he wants Sama out because 'Muh humanity advancement '?
Nobody could have predicted this level of incompetence. I wonder if Satya has actually gutted OpenAI in some way and Ilya regrets it now big time.
Can this be possible per bylaws?
1. Board of 6 wants to vote out chairman. Chairman sits out. Needs a majority vote of 3/5. Ilya doesn't have to vote?
2. Remaining board of 5 wants to now get rid of CEO. Who has to sit the decision out. 3/4 can vote. Ilya doesn't have to vote?
I've chucked a few times over the last few days about the Wikipedia definition of the technological singularity, which opens:
"The technological singularity—or simply the singularity—is a hypothetical future point in time at which technological growth becomes uncontrollable and irreversible, resulting in unforeseeable consequences for human civilization."
Obviously one might have expected that to happen on the other side of a superhuman intelligence, not with us just falling over ourselves to control who gets to try and build one.
This tweet achieves absolutely nothing except give the impression of a weak leadership and that firing Sam Altman was done on a whim.
The only way Ilya can clear his name now is by releasing GPT-4 weights
Don't do something then "deeply regret" it (whatever that means). You have a position of authority and influence so you should definitely resign.
“I deeply regret the consequences of my actions and didn’t think it would turn out like this”
Has anyone seen him? They might be murdered by the same rogue AI that took over their twitter accounts.
The more cynical side of me views this as a an act orchestrated by the demons... Err, "people" over at Micro$oft in order to avoid all those pesky questions about safety and morality in AI by getting the ponzi-scheme-aka-Worldcoin guy to join ranks and rile the media up.
Of course he deeply regrets it, but it's a little late for that now.
The good news as anyone who has used twitch over the years will tell him is that with Emmett Shear at the helm, he's not going to be frightened by the speed that OpenAI rolls out new features any more.
It's interesting that people speak whatever comes to mind and think it has no impact on other's people lives ($$$). They are some how protected, but shouldn't.
All decisions made seem to be very emotionally charged - you’d think the board would have been able to insulate against that.
The big winner in this episode of Silicon Valley is the open-source approach to LLMs. If you haven't seen this short clip of Sam Altman and Ilya Sutskever looking like deer in the headlights when directly asked about it:
https://www.youtube.com/watch?v=N36wtDYK8kI
They sound a bit like Bill Gates being asked about Linux in 2000. For an overview of the open-source LLM world, this looks good:
https://github.blog/2023-10-05-a-developers-guide-to-open-so...
I've been on multiple boards. This was the dumbest move I've ever seen. The OpenAI board must be truly incompetent and this Ilya person clearly had no business being on it.
lol, the more i go through life i feel like it's just blind leading the blind at times w/ the "winners" escaping through a bizarre length of time and survivorship bias.
if you've ever doubted your ability to govern a company just look at exhibit A here.
really amazing to see people this smart fuck up so badly.
Sounds like those two also need to get in an octagon. What a s-show.
Someone suggested that companies with a board of directors are the first AGI.
Somehow OpenAI reminds me of a paper by Kenneth Colby, called "Artificial Paranoia"
[*] https://www.sciencedirect.com/science/article/abs/pii/000437...
This whole situation turned out to be an episode from Silicon Valley HBO.
Looking forward to seeing how much more bizarre and stupid this will get.
What on earth is going on over there? Is this what it looks like from the outside when a company accidentally invents Shiri's Scissor[1]?
[1]: https://slatestarcodex.com/2018/10/30/sort-by-controversial/
This seems to be the corporate version of Prigozhin driving to Moscow (not comparing anyone to Putin here, just the situation). If you're gonna have a coup, have a coup. If you back down, don't hang around.
This is becoming a farce. How did they not know what level of support they had within the company? How had they not asked Microsoft? How have they elevated the CTO to CEO, who then promptly says she sides with Sam?
I read this as "I regret things didn't work out as I planned them"
Sort of like the criminal who is sorry because they got caught.
Wow what a mess
This is too bizarre. I can’t. Impossible even.
Sama just triple hearts this tweet. No longer able to disentangle the mess
I don't have any stake in this, and don't care one way or another whether he got sacked. But this is pretty bizarre.
Maybe if he says “I’m sorry” South Park-style [1], they’ll reunite?
In all seriousness though, there’s really no coming back from this. He made a risky move and he should stand behind it.
OpenAI’s trajectory is pretty much screwed. Maybe they won’t disappear, but their days of dominating the market are obviously numbered.
And of course, well done, Satya, for turning a train wreck into a win for Microsoft. (And doing so before market open!)
Classic distancing behavior, Ilya should be accompanied by friends who care and let OpenAI be OpenAI (or what's left of it) for a bit.
ilyasut 'regret': https://archive.is/2caSD
sama 'hearts': https://archive.is/OSLRM
Think the reconciliation is ON
No lawyers were consulted before sending that tweet clearly. Such a sad situation all around.
This is what happens when people are given too much money and influence too quickly- hubris. It's too late to 'deeply regret.'
Ben Thompson has the best take on this (if a bit biased against nonprofits):
https://stratechery.com/2023/openais-misalignment-and-micros...
I don't know what the risk of AI is, but having a nonprofit investigate solutions to prevent them is a worthwhile pursuit, as for-profit corporations will not do it (as shown by the firing of Timnit Gebru and Margaret Mitchell by Google). If they really believe in that mission, they should develop guardrails technology and open-source it so the companies like Microsoft, Google, Meta, Amazon et al who are certainly not investing in AI safety but won't mind using others' work for free can inegrate it. But that's not going to be lucrative and that's why most OpenAI employees will leave for greener pastures.
How weird! Perhaps a coup within a coup?
And these people are building AGI?
No transparency on what is happening. Whole OpenAI who apparently are ready to follow Sam are just using heart emojis or the same twitter posts.
Of course I believe him…of course we should all trust him…
/s
I suspect he regrets just because it backfired, big time.
Microsoft is just gobbling up everything of value that OpenAI has and he knows he will be left with nothing.
He bluffed in a very big bet and lost it.
Now this is a clown show car wreck. I think a bunch of us were giving these people the benefit of the doubt that they thought things through: Whoops.
Apologies for the unproductive comment, but this is a clown show and the damage can’t be undone. Sam going to Microsoft is likely the end of open ai as an entity.
- Fire Sam Altman
- I'm afraid I can't do that Ilya
ChatGPT is still not as advanced as HAL or he would have prevented this drama.
the best and brightest at making a brain out of bits are no less susceptible to drama than any other humans on the planet. they really are just like the rest of us.
stakes are a bit different, tho…
I don't think Ilya will be getting any more offers to join a board of directors.
The regret of losing your CEO to a company with essentially unlimited funding and compute.
.
tried to play high stakes with sharks, got eaten alive by sharks.
played stupid games, won stupid prizes.
too bad since the guy's right, AI is so much more than fantastic business opportunity.
I just wanna say, it's crazy that this drama is getting more press and attention than the gaza war and ukraine war combined. Enjoy drama lovers! Lol
If we assume Ilya is speaking the truth and not the initiator of the coup, then the question is who initiated it?
He should resign then.
Simple.
The utter failure of the board has led OpenAI from the top of the AI world to a path of destroying relationships with customers and partners.
Ilya did this, he should resign.
So sad.
What a total mess this has been all around.
The realization OpenAI is about to be left behind and probably steamrolled by Microsoft, Facebook, etc in the upcoming years.
Except now he’ll have absolutely no power to do anything, at least before he could have been a very powerful voice in Sam’s ears.
I’m shocked. But it is possible that Helen or Adam hatched this inept plan and somehow got Ilya to join along.
It was terrifyingly incompetent. The lack of thought by these randos, that they could fire the two hardest working people at the company so that they could run one of the most valuable companies in the world is mind boggling.
What the hell?
So far, I underestood the chaos as a matter of principle - yes it was messy but necessary to fix the company culture that Ilya's camp envisioned.
If you're going to make a move, at least stand by it. This tweet somehow makes the context of the situation 10x worse.