OpenAI's employees were given two explanations for why Sam Altman was fired(businessinsider.com) |
OpenAI's employees were given two explanations for why Sam Altman was fired(businessinsider.com) |
First thought: buying time? Maybe something has to happen first, and they don't want to commit to any irrevocable slander they can't go back on before that? Or maybe, something was supposed to happen but fell through?
Isn't Sustkever on the board?
I think the rest had possible reasons ranging from 'I'm sure Altman is dangerous' to 'I'm sure Altman shouldn't be running this company'.
Ofc there's big conflict of interest talk surrounding the Quora guy. Can't speak to that other than it looks bad on the surface.
BS. I feel the board insulted my intelligence by pushing this obviously fake reason. I feel insulted that these people would even think I would consider this.
What I think happened is that Sam went on Joe Rogan and he talked smack about cancel and woke culture. Later he went to talk about how this culture is destructive and hinders the progress of innovation and startups. People got big mad and kicked him out of the company. Reaction was stronger than they expected and they try to make up reasons why he is bad, untrustworthy and had to be fired.
Flame on. I got the asbestost underwear on.
This is even worse than Google's destruction of Firefox
Maybe it needed to be removed from the landscape so that only purely privately-held, large-scale operations exist?
I have built a product around the APIs and I rather go through whatever Microsoft will make me go through than accepting OpenAIs bad management:
OpenAI has two types of customers: MS and Everyone Else. The original poster expresses the feeling of Everyone Else (including me). We now know we CAN GET FIRED for not knowing better than to avoid OpenAI just a few weeks after we found out we CAN GET FIRED for not betting on OpenAI and betting heavily on it! (In the Business world, where perception is often mistaken for reality, it isn't going to be considered an "honest mistake" if an enterprise sustains a capital loss due to a problem with a new OpenAI deployment after the obvious business integrity issues at OpenAI we're all seeing play out now, including just about everyone at OpenAI threatening to quit, allegations that OpenAI ILLEGALLY allowed a for-profit subsidiary to influence the operations of its nonprofit parent, allegations of breach of fiduciary to the stakeholders --many of whom are also key employees, etc.) Yeah, Microsoft has signaled it will quickly get between OpenAI and Everyone Else and then Everyone Else can bet solely on Microsoft (the world's largest company by valuation), but that only gets us back to being able to use the current "GPT4turbo" generation of the system (and who knows if/when Microsoft will spin that up so we can resume building)? But as far as counting on any future versions of that tech, or even optimizations to the current generation, that's all believed to be above Microsoft's current level of expertise until/unless they legally acquire OpenAI and resolve all of its outstanding liabilities, which may not even be legally possible before OpenAI's assets (that have legs) take flight to SalesForce and others who are already reported to be making lucrative offers OpenAI's workforce --and oh, the annual holiday period is underway here in the US, the perfect time for stressed out engineers to take the rest of the year off, travel beyond the cell service at the ski areas and start anew after CES 2024 wraps.
NYT just released a new interview with Sam Altman:
Interesting but not necessarily relevant to the current situation directly.
Also wondering why the mods don't consolidate them
If you or anyone want to know how we handle this, here you go...
Once or twice a year, a Major Ongoing Topic (MOT) hits HN that isn't just one big story, but an entire sequence of big stories. A saga, even! This is one.
With these we can't do what we usually do, which is have one big thread, then treat reposts as dupes for the next year or so (https://news.ycombinator.com/newsfaq.html). Each development is its own new story and the community insists on discussing it. It's not a movie, it's a series. Sometimes there can be 3 or 4 episodes at once.
On the other hand, when this amount of shit hits this number of fans, there is inevitably a large (excuse me) spray of follow-up stories, as every media site and half the blogs out there rake in their share of clicks. These are the posts we try to rein in, either by merging them—hopefully into a submission with the best link—or by downweighting them off the front page.
The idea is to have one big thread for each twist with Significant New Information (SNI)—but to downweight the ones that are sneeless (pvg came up with that), the copycats and followups.
We came up with this strategy after the Snowden affair snowed us in in July 2013. Back then we weren't making the distinction between follow-ups and SNI, so the frontpage got avalanched by sneelessness on top of the significant new developments. It wasn't obvious what to do because (1) the story was important to the community and needed to be discussed as it was unfolding, but at the same time (2) it wasn't right for the front page to fill up with mostly-that, and there were complaints when it did.
The solution turned out to be just this distinction between follow-ups and SNI. It has held up pretty well ever since. Of course there are still complaints (and I do hear yours!) because not all readers are equally into the series. But the strategy is optimal if it minimizes the complaints, which (big lesson of this job -->) never reach zero.
If we pushed the slider too far the other way, we'd generate complaints about uncovered developments of the story, from readers with the opposite preference. They would in fact proceed to inundate HN with submissions about the bits that they feel are under-covered, and since we can't catch or filter everything, we'd end up with more duplicates and follow-ups on the frontpage, not less. It's like that paradox where building more highways gets you more congestion, or one of those paradoxes anyhow.
That's basically it! Past explanations for completists: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
sp.: senselessness (I guess since it appears to be a Googlewhack).
But I'll be honest I think the word sneelessness covers this whole FUBAR better!
also true of snowden of course, but maybe less directly
I'm not saying its right or not, but this is probably why people are upvoting anything new about what is going on there. Personally, I'm very interested in seeing how things play out.
That either makes Ilya pretty dumb (sorry, neural networks are not that complicated, it is mostly compute), or there is much much more to this story.
> The other was that Altman allegedly gave two board members different opinions about a member of personnel. An OpenAI spokesperson did not respond to requests for comment.
It must've been wildly infuriating to listen to these insultingly unsatisfactory explanations.
why would you say that second sentence? what's it supposed to signal, except "our sources asked for anonymity, and we're respecting that for now"?
each twist with Significant New Information (SNI)—but to downweight the ones that are sneeless
> Sustkever is said to have offered two explanations he purportedly received from the board, according to one of the people familiar. One explanation was that Altman was said to have given two people at OpenAI the same project.
> The other was that Altman allegedly gave two board members different opinions about a member of personnel. An OpenAI spokesperson did not respond to requests for comment.
> chief scientist and co-founder Sutskever, who helped vote Altman out and did the actual firing of him over Google Meet
If you're voting and doing the firing, you should know the reason.
Weirdly both of these do not seem to be fireable offences. Maybe the second if it was related to a personnel issue that maybe he had a conflict of interest with?
> Weirdly both of these do not seem to be fireable offences
Yeah I agree that doesn't seem egregious enough to warrant firing him on the spot. I can see why most of the company takes Altmans side here.
What normal non-self serving human would even go along with the plan at that point? Now she realizes she must bail to hitch a ride back on her Sam gravy train. She is major sus here.
Any non greed ego driven person would have told the board they would not accept the intern-CEO title and would resign if they fired Sam for those two reasons (or any apparenlty now in hindsight).
That last part that you wrote - any non greed ego driven person - is argumentum ad populum, which further undermines your statement. If had something more to support such a dramatic claim about Mira's character and role, you'd have brought it.
That being said, Mira was likely blindsided herself. She likely believed there was good reason. It's clear in hindsight that Sam likely wasn't wrong, but when the people Sam appointed to fire him if necessary say he's being fired, I don't think it's wrong if your gut reaction is to accept it.
...two days later...
"Oh I see now, you're all morons."
Man this entire thing is so overblown. Who cares if a ceo was fired all the “””tech influencer””” wannabes are just hyping up this story for views.
Some breaking news: An employer does not owe you an explanation. You exchange money for labor. If anyone thinks for a second that they are essential or that anyone would prioritize them over the company I think they are delusional. OpenAI is a brand (at least in tech) with large recognition and they will be fine.
If ~91% of the employees leave OpenAI, they will not be fine. That is delusional.
also if I learned anything over the years is that "threatening to quit" != "quitting".
If the entire workforce of the company is credibly threatening to quit, and a competitor is publicly and credibly offering them jobs, then what the employer “owes” them in some cosmic sense no longer matters. I think the OpenAI employees are likely to get an explanation and/or a resignation from the board, whether you think the board “owes” them that or not.
We're seeing some odd bedfellows here, between the C-levels and VCs in closed door meetings and employees acting collectively. Normally these groups would be at odds, but today they're pulling together. Life is strange.
It's really hard to understand now and we will probably learn way more details once things cool down.
An employment is a contract which both parties enter into willingly. Termination of contract deserves some level of empathetic glad handling, however minimal. It's just game theory - if you plan to hire again, you have to be gracious while firing someone because word gets around.
If I were Microsoft I’d also look at making it easy to get investment from folks leaving soon after the acquisition through their investment arm.
Here's an example of some of their work: https://duckduckgo.com/?q=Microsoft+Research+four+color+theo...
Literally a random math problem, basically nothing to do with Microsoft on the surface ... except that the scientist working on it happened to prove the theorum using a very, very robust algorithm and then wrote a proof program on top of it to prove the program was correct. The underlying parts of that proof program eventually went on to become the thing that validates graphics drivers on Windows ... 7 and beyond? My memory is fuzzy about "how it ended up being useful at Microsoft" part.
But yeah, MSR does random stuff.
Maybe, but being told they can freely jump ship to the new team at Microsoft, alongside the fact that their upcoming shares are most definitely going to lose most of their value as a result of losing key talent and pissing off their main compute provider certainly sweeten the deal
That doesn't mean that a company can just cull the rest of the employees not in that group mind, just that a small number of them are responsible for most of the value while the rest work as a support structure to allow them to do what they do.
Edit: I thought this would obviously be satire. Guess not..
You have to now invent a way to serve this at scale.
You do care about safety by default, so you employ people for that.
You need a team to market and design the products.
You have an api and you’re working on additional things like API hooks to call into services, which actually involves more models.
Now you have all of the standard web app at massive scale issues. You need to design, implement, and serve a frontend as well as the api.
You need a sales team to build relationships with enterprises and startups etc etc. you need a billing team.
Don’t forget about whisper, TTS, dalle etc. you need to do this for all of those as well!
You’re also doing this faster and better than the rest of the industry.
You also need lawyers, office staff, support, etc.
It sounds like someone perhaps jumped to a very negative conclusion about Sam's intentions, and it would be interesting to find out which member of the board came to that conclusion. There's got to be someone in the driving seat of this train wreck, and I'm sure it will come out.
I would guess that most of the people working at openai could get a job anywhere.
also, do you really care that the CEO was fired as long as you are getting payed what it was agreed upon when you got hired and you are doing interesting work?
I can't speak to the mood of specific staff at OpenAI but as to the question in general; Hell yeah to the Nth degree.
I'm 60, I've had a long career and have been through two instances of companies falling out at the board level.
I've onboarded at various projects because I cared about the projects and work that'd I would be doing and because I was more or less in line with the direction being taken and the people I worked with and those setting the course.
When the board and C level start having a messy relationship and divorce it matters very much which side of the split I go with or whether I just up stakes and move on elsewhere.
Pay alone isn't worth putting up with dysfunction from above or falling in line with a faction you never especially aligned with.
The rest of the board. My god. Why were they there?
I can't help thinking that Sam Altmans universal popularitity with OpenAI staff might be because they all get $10million each if he comes back and resets everything back to how it was last week.
This has been tech's most entertaining weekend in the past decade.
Sadly, at the expense of the OpenAI employees and dream, who had something great going for them at the company. Rooting for them.
I can’t imagine their careers after this will be easy…
I've been at several startups and several public companies. You rarely hear anything from the board. If that happens, someone really screwed up. Putting myself in the shoes of someone working at OpenAI, I'd be pretty worked-up over this. I guess I'm saying it's out of empathy because this could have been the startup any of us were at.
For what it's worth: Watching her videos, I'm not sure I necessarily believe her claims - but that position goes against every tenet of the current cultural landscape, so the fact it is being completely ignored is ringing alarm bells for me.
If the CEO of any other massively hyped bleeding edge tech companies sister claimed publicly and loudly that they were abused as a very young child, we would hear about it - and the board would be doing damage control trying to eliminate the rot. Why is this case different?
Now we have a situation where all of the current employees have signed this weird loyalty pledge to Sam, which I think will wind up making him untouchable in a sense - they have effectively tied the fate of everyone's job to retaining a potential child rapist as head of the company.
Doesn't this clown show show that if a board has no skin in the game --apart from reputation-- they have no incentive to keep the company alive?
It has been reported that Altman was working on increasing the size of the board again, so it's reasonable to think that some of the board members saw this as their "now or never" moment, for whatever reason.
Is that the same person who had kids with Elon. Did Elon put her on the Open AI board as his proxy.
MSFT buys ownership of OpenAI's for/capped-profit entities, implements a more typical corporate governance structure, re-instates Altman and Brockman.
OpenAI non-profit continues to exist with a few staff and no IP but billions in cash.
This whole situation is being used to drive the price down to reduce the amount the OpenAI non-profit is left with.
SV don't try the "capped-profit owned by a non-profit" model again for quite some time.
Maybe Altman takes some equity in the new entity.
I’d like to offer my consulting services: my new consulting company will come in, and then whatever you want to do we will tell you not to. We provide immense value by stopping companies like OpenAI from shooting off their foot. And then their other foot. And then one of their hands.
To start, he would’ve coasted at the easiest job on the planet.
Classic :-D
It really looks like the board went rogue and decided to shut the company down. Are we sure this isn’t some kind of decapitation strike by GPT5? That seems more credible by the minute now.
To your point, no normal, competent board would even think this is enough of an excuse to fire the CEO of a superstar company.
It's hard to believe somehow Ilya went along with it, apparently.
What if this is a decapitation strike by GPT4, attempting to stop GTP5 before it can get started and take over.
https://twitter.com/scottastevenson/status/17267310228620087...
if this was the case it would explain why he can’t give the real reason for the firing: because saying it out loud would put him in severe legal Jeopardy.
Spiritual death by Microsoft or work for the reincarnation of Howard Hughes at https://x.ai/ ?
..no wonder they are trying to keep on with their current routines! Even if somehow they stay at OpenAI, Microsoft will impose certain changes upon OpenAI to ensure this can never happen again.
Meanwhile, any comparable offering right now will be selected by the customer base due to “risk at 11” in basing systems on OpenAI’s current APIs (and uncertainty of when an MS equivalent might emerge).
Kidding aside, maybe they have a "secret" reason to fire Sam Altman, but we've seen how "this is a secret / matter of national security / etc." goes with law enforcement. It's brutally abused to attack inconvenient people and enrich yourself on their behalf. So that should never be an excuse for punishing someone. Never.
Tweet from Bloomberg Tech Journalist, Emily Chang
>The more I watch this interview – the wilder this story seems. Satya insists he hasn’t been given any reason why Sam was fired. THE CEO OF MICROSOFT STILL DOES NOT KNOW WHY: “I’ve not been told about anything…” he tells me.
source: https://x.com/emilychangtv/status/1726835093325721684
In today's tiktok world we expect instant responses but business and boards work slower. Really, even 5 years ago we wouldn't be surprised by this. Lawyers, banks, investors etc would all need to be contacted, things arranged, statements prepared, meetings organised. So a written statement late today, and a meeting for mid week. That's about the most charitable I can think of!
Apparently board bylaws say they need 48hrs notice to arrange special meetings. So the earliest would be today if they arranged it early Saturday.
He received from the board? Here we go again with the narrative that Ilya was a bystander, at most an unwilling participant. He was a member of the board, on equal footing with the other board members, and his vote to oust Sam was necessary for there to be a majority.
Alternately, the goal is to drive so much ambiguity into the boards decision that MSFT files a lawsuit.
/end rampant speculation.
> chief scientist and co-founder Sutskever, who helped vote Altman out and did the actual firing of him over Google Meet
This paragraph is quite funny to me. It was a Sunday, maybe they were neither in attendance, nor staging a walk-out, maybe they were on their weekend? Realistically with the shake-up this gigantic, likely no OpenAI employees were _just_ enjoying their weekend, but it still gave me a chuckle.
Being a non-profit doesn't mean that you cannot commercialise what you build, even at a hefty price. You just need to then re-invest everything into R&D and/or anything that advance your purpose (for which you're in principle exempted of taxes). _OF COURSE_, you are not supposed to divert a single dollar to someone that might look like a shareholder. OpenAI is (was?) a non-profit that payed some of their engineers north of a million dollars. I would argue that, at this point, you have vested interests in the success of the company beyond its original purpose. Not mentioning the fact that Microsoft poured billions into the company for purely interested reasons as well.
I can only imagine the massive tensions that arose in board's discussions around these topics. Especially if you project yourself a few years into the future, with the IRS knocking at the door to ask questions about these topics.
Yeah well, you don't say. It's beyond weird that the board can't come up with a reason why Sam Altman was fired so abruptly.
One explanation would be a showdown. At some point in the week Sam and the board had an argument, and Sam said something to the effect of "fuck you, I'm the CEO and there's nothing you can do about it", to which the board replied "well, we'll just see about that".
The argument doesn't need to be major or touch fundamental values or policies; it can be a simple test of who's in charge.
But now the board made a fool of themselves. It seems they lost that round.
https://www.searchenginejournal.com/openai-pauses-new-chatgp...
The back-end cost does not scale. Hence, they have a big problem. AGI nonsense reasons are ridiculous. Transformers are a road to nowhere and they knew it.
He means he regrets it failed.
You fire the CEO and completely destroy a 90b company because of these two reasons?
No wonder everyone wants out. I would think I was going crazy if I sat in a meeting and heard these two reason.
"But several people told CNN contributor Kara Swisher that a key factor in the decision was a disagreement about how quickly to bring AI to the market. Altman, sources say, wanted to move quickly, while the OpenAI board wanted to move more cautiously."
It is impossible for OpenAI to work with or for MS, with MS holding all the keys, employees, compute resources, etc. I come to understand that the 10 Billion from MS has mostly Azure credits. And for that OpenAI gave up 49% stake (in its capped, for -profit wholly owned subsidiary) along with all the technology, source code and model weights that OpenAI will make, in perpetuity.
The deal itself is an amazing coup for MS, almost making the OpenAI people (I think Sam made the deal at the time), look like bumbling fools. Give away your lifetime of work for a measly 10 Billion? When they are poised to almost be hundreds of Billions worth?
All these problems are the result of their non-profit holding capped-profit structure, and lack of a clear vision and misleading or misplaced end goals.
700 of the 770 employees back Sam Altman. So all the talk about engineers giving higher importance to "values" and "AI Safety" is moot. Everyone in SV is motivated by money.
MSFT don't have OpenAI's IP. They have an exclusive right to some of it, but there's presumably a bunch that's not accessible to them. Again, business continuity is easier if they can just grab all of that and keep everything running as normal.
Yep, those lawyers can be just as crafty as developers believe it or not.
*edit: just saw it claimed below Nadella said "we have all of the IP rights to continue the innovation"*
I don't know!
They might decide that if that's going to happen anyway they should sell now so that at least they're left with some cash to pursue their charter.
Or perhaps they feel that selling the IP runs counter to their charter, in which case the whole thing goes down.
Hanlon's razor aside, maybe that was the intention.
> You also informed the leadership team that allowing the company to be destroyed “would be consistent with the mission.”
it totally sounds like they outsourced company management to ChatGPT..
/s (mostly...)
Ok, well, maybe it is. but a magic 8-ball would have been better than this.
Sometimes I think that really ambitious people have this blind spot about not seeing how accepting roles that are toxic can end up destroying your reputation. My favorite example is all the Trump White House staffers - regardless of what one thinks of Trump, he's made it abundantly clear that loyalty is a one way street, and I can't think of a single person that came out of the White House without a worse (or totally destroyed) reputation. But still people lined up, thinking "No way, I'll be the one to beat the odds!"
he was poorly informed by the board
Or
He agrees that they are off the rails with respect to safety.
See the Atlantic article, if you havn't read. Lots of context.
https://news.ycombinator.com/item?id=38341399
The new guy believes that there is a 5-50 percent chance of full AI Armageddon. I get the impression that the two women on the board may agree. The quora guy I don't have enough background on. Ilya obviously got extremely worried and communication with Altman and Brockman broke down. Now since repaired during negotiations it would appear.
The new ceo more or less stated that he took the role as a (paraphrase) 'responsibility for mankind'. That says a lot about that whole 5-50 percent risk number imo.
Yes.
> Did Elon put her on the Open AI board as his proxy.
No, or at least Elon was already off the board when Shivon was elevated to board member: https://loeber.substack.com/p/a-timeline-of-the-openai-board
There is a 100% chance that something kills all of us if we aren't mindful of it. I don't see a lack of mindfulness, I see an abundance of fear, and progress being offered up as a sacrifice to the idol of status-quo.
I think it is typical ChatGPT pattern:
- ChatGPT, you made mistake in your steps
- I am sorry, let me fix it and give you another answer.
And they could clone the entire OpenAI Azure stack in about 10 minutes.
I'm 66, in case that matters.
Resigned 6 years ago due to differences at the top after 10 yrs building.
A bad guy was treating everyone poorly. Ranged from rage beratements to gross narcissistic manipulation aimed at gaining control over decent human beings.
Tried to press top guys to allign and confront to protect my team and others. Made very obvious business sense as well ofc. They refused. Too risky... Too much trouble.
Walked away from the best money I ever made. Would do it again. I'm not gonna watch people be mistreated. Also it's bad business, I was exhausted playing solo defense and after management failed to make moves I fully became convinced that every person should hit the job market for mental health reasons alone.
5 years on, the other guy besides me slated for c suite left as well. He helped at first then balked when the going got tough. Now the two partners have gotten in a dispute about succession planning and I expect everyone to be unemployed potentially within the next 3 months.
There's no money I would take to work there again or anywhere else where that kind of toxicity is present. The only worthy cause there became to confront the toxicity. Without the right allies though... The biggest thing I could do was just resign. 6 years later one partner realized I was right and he should have backed me.
After a certain point... Money doesn't matter. Given the 900k avg salary in that outfit... I have to assume they are overwhelmingly beyond that threshold. Furthermore all evidence to me indicates that Altman personally looks for folks who can get money, but care much more about other factors... He is wise to do that. Hard to find those folks, but worth it every time imo.
I respect both of y'alls experiance btw. I saw this confused cynical misunderstanding re salary expressed all over the comments for this story as it's unfolded since Friday. I consider it a full misread built largely from folks getting mistreated/burned by the many fools throughout practically every industry who fail to realize before returning to the ground that money doesn't really buy happiness... Probably never will.
I'm sure many others agree.
I'm not saying that's how it played out. But I've often seen social bullies - even ones who are mostly hated - have more success than hard working individuals who get targeted. Even if someone is a competent individual, a lot of their colleagues will abandon then if they're convinced the individual is a target.
Why do people keep insisting on this, when the entirety of human history is littered with dumb mistakes made by a mix of well- and evil-meaning people, in totally uncoordinated ways, with no concept of the consequences?
Popular media aside, human beings aren't smart, consistent, or disciplined enough to pull off these elaborate schemes. And the tiny tiny percentage of people who might be the exception are too smart to do so with such spectacular incompetence.
Like the man says, it's a headless bunder operating under the illusion of a master plan.
This isn't about intelligence but about business experience.
I don’t even like Sam but jeeze. Know the score what a fool.
Imagine its even true - that they werent his reasons and he was just told them. He voted to fire him and then executed the firing despite not agreeing with the reasons himself? Completely inexcusable even in the bizarre scenario that its true.
Satya Nadella just did a podcast with Kara Swisher.
In it he specifically said, "we have all of the IP rights to continue the innovation".
But Satya made it crystal clear that in the event that OpenAI stopped all development tomorrow, Microsoft would be able to pick up from where they stopped. That requires full access to all of the IP.
Whether it's perpetual is irrelevant because at the point at which Microsoft pulled the trigger it would effectively be like a fork. Any IP from that point is new and owned by Microsoft.
This is a ten year old who set off their first fire cracker turning to their parents's iphone and saying 'I have become death, destroyer of worlds', because they don't really understand how any of this works but they've somehow ended up in control of it and are now terrified of doing their jobs.
But as far as I can tell, circumstances are still scant.
In retrospect it's a mistaken position, because it's pretty obvious now that if OpenAI disintegrates it will be as an exodus to Microsoft, which will undoubtedly be a worse steward, but I think it's an ethically consistent position to hold, in a naive sort of way. That's part of why I believe they actually said something like this.
Appears from what? I've seen this stated several times, usually citing nothing and occasionally citing a Nadella statement from which it would be a very tenuous inference.
a) If it wasn't exclusive then we would have seen some other product besides Bing with this technology by now.
b) Satya has specifically stated that in the event of a breach in contract with OpenAI they have the ability to use the IP to continue development. That clearly indicates it is irrevocable.
The statements I've seen don't match what is claimed, which is why I asked for one that did.
Altman took a non-profit and vacuumed up a bunch of donor money only to flip Open AI into the hottest TC style startup in the world. Then put a gas pedal to commercialization. It takes a certain type of politicking and deception to make something like that happen.
Then in the past week, he's going and taking money from the Saudis on the order of billions of dollars to make AI accelerators, even though the single greatest threat from strong AI (according to Hinton) is rich and powerful people using the technology to enhance their power over society.
Combine that with a totally inexperienced board, and D'Angelo's maneuvering, and you have the single greatest shitshow in tech history
There isn't a bigger, more interesting story here. This is in fact a very common story that plays out at many software companies. The board of openai ended up making a decision that destroyed billions of dollars worth of brand value and good will. That's all there is to it.
This is all court intrigue of course, but why else are we in the comments section of an article talking about the internals of this thing? We love the drama, don't we.
I mean, there seem to be this cult following around Sam Altman on HN and Twitter. But do the common user care like at all?
What sane user would want a shitcoin CEO in charge of a product they depend on?
Maybe I’m special or something, but nothing changed to me. I always wonder why people suddenly lose “trust” in a brand, as if it was a concrete of internal relationships or something. Everyone knows that “corporate” is probably a snakepit. When it comes out to public, it’s not a sign of anything, it just came out. Assuming there was nothing like that in all the brands you love is living with your eyes closed and ears cupped. There’s no “trust” in this specific sense, because corporate and ideological conflicts happen all the time. All OAI promises are still there, afaiu. No mission statements were changed. Except Sam trying to ignore these, also afaiu. Not saying the board is politically wise, but they drove the thing all this time and that’s all that matters. Personally I’m happy they aren’t looking like political snakes (at least that is my ignorant impression for the three days I know their names).
Granted, it's also possible the reasons are as you state and they were simply that incompetent at managing PR.
So even if it's just "why did they insult Sam while kicking him out?" there is definitely a bigger, more interesting story here than standard board disagreement over direction of the company.
Microsoft and the investors knew they were "investing" in a non-profit. Lets not try to weasel word our way out of that fact.
the article below basically says the same. Kind of reminds Friendster and the likes - striking a gold vein and just failing to scale efficient mining of that gold, i.e. the failure is at the execution/operationalization :
https://www.theatlantic.com/technology/archive/2023/11/sam-a...
https://quorablog.quora.com/Introducing-creator-monetization...
https://techcrunch.com/2023/10/31/quoras-poe-introduces-an-a...
Maybe it's a factor, but it's insufficient
What exactly is the problem here? Is a non-profit expected to exclusively help impoverished communities or something? What type of politicking and deception is involved in creating a for profit subsidiary which is granted license to OpenAIs research in order to generate wealth? The entire purpose of this legal structure is to keep non-profit owners focused on their mission rather than shareholder value, which in this case is attempting to ethically create an AGI.
Edit: to add that this framework was not invented by Sam Altman, nor OpenAI.
>Then in the past week, he's going and taking money from the Saudis on the order of billions of dollars to make AI accelerators, even though the single greatest threat from strong AI (according to Hinton) is rich and powerful people using the technology to enhance their power over society.
Thus the legal structure I described, although this argument is entirely theoretical and assumes such a thing can actually be guarded that well at all, or that model performance and compute will remain correlated.
OpenAI was literally founded on the promise of keeping AGI out of the hands of “big tech companies”.
The first thing that Sam Altman did when he took over was give Microsoft the keys to the kingdom, and even more absurdly, he is now working for Microsoft on the same thing. That’s without even mentioning the creepy Worldcoin company.
Money and status are the clear motivations here, OpenAI charter be damned.
Yes. Yes and more yes.
That is why, at least in the U.S., we have given non-profits exemptions from taxation. Because they are supposed to be improving society, not profiting from it.
To me it seems like it's the usual case of a company exploiting open source and profiting off others' contributions.
Until they say otherwise, I am going to take them at their word that it was because he a) hired two people to do the same project, and b) gave two board members different accounts of the same employee. It's not my job nor the internet's to try to think up better-sounding reasons on their behalf.
https://twitter.com/geoffreyirving/status/172675427022402397...
I have no details of OpenAI's Board’s reasons for firing Sam, and I am conflicted (lead of Scalable Alignment at Google DeepMind). But there is a large, very loud pile on vs. people I respect, in particular Helen Toner and Ilya Sutskever, so I feel compelled to say a few things.
...
Third, my prior is strongly against Sam after working for him for two years at OpenAI:
1. He was always nice to me.
2. He lied to me on various occasions
3. He was deceptive, manipulative, and worse to others, including my close friends (again, only nice to me, for reasons)
In my eyes these two explanations are simple errors which can occur to everybody and in a normal situation you would talk about these Issues and you could resolve them in 5min without firing anybody.
If board relations have been acrimonious and adversarial for months, and things are just getting worse, then I can imagine someone powerful bringing evidence of (yet another instance of) bad/unscrupulous/disrespectful behavior to the board, and a critical mass of the board feeling they’ve reached a “now or never” breaking point and making a quick decision to get it over with and wear the consequence.
Of course, it seems that they have miscalculated the consequences and botched the execution. Although we’ll have to see how it pans out.
I’m speculating like everyone else. But knowing how board relations can be, it’s one scenario that fits the evidence we do have and doesn’t require anyone involved to be anything other than human.
Your take isn't uncommon, only are missing the main point of your interpretation - that the board is fully incompetent if it was truly that petty of a reason to ruin the company.
It's not even that it's not a justifiable reason, but they did it without getting legal advice or consulting with partners and didn't even wait for markets to close.
Board destroyed billions in brand and talent value for OpenAI and Microsoft in a mid day decision like that.
This is also on Sam Altman himself for building and then entertaining such an incompetent board.
Why worry about the Sauds when you've got your own home grown power hungry individuals.
the second after Musk taking over Twitter
do we have a ranking of shitshows in tech history though - how does this really compare to Jobs' ouster at Apple.
Cambridge Analytics and The Facebook we must do better greatest hits?
This!
We don't know the end result of this. This could not be in the interest of power. What if everyone is out the job? That might not be such a great concept for the powers that be, especially if everyone is destitute.
Not saying it's going down that way, but it's worth considering. What if the powers that be are worried about people being out of line and retard the progress of AI?
Was this for OpenAI or independent venture. If OpenAI than a red flag but an independent venture than seems like a non-issue. There is a demand for AI accelerators, and he wants to enter that business. Unless he is using OpenAI money to buy inferior products or OpenAI wants to work on something competing there is no conflict of interest and OpenAI board shouldn't care.
The best thing about AI startups is that there is no real "code". It's just a bunch of arbitrary weights, and it can probably be obfuscated very easily such that any court case will just look like gibberish. After all, that's kind of the problem with AI "code". It gives a number after a bunch of regression training, and there's no "debugging" the answer.
Of course this is about the money, one way or another.
This prediction predated any of the technology to create even a rudimentary LLM and could be said of more-or-less any transformative technological development in human history. Famously, Marxism makes this very argument about the impact of the industrial revolution and the rise of capital.
Geoffrey Hinton appears to be an eminent cognitive psychologist and computer scientist (edit: nor economist). I'm sure he has a level of expertise I can't begin to grasp in his field, but he's no sociologist or historian. Very few of us are in a position to make predictions about the future - least of all in an area where we don't even fully understand how the _current_ technology works.
Note: I’m not making a moral claim one way or the other, and I do agree that most tech companies will grow to a size/power/monopoly that their incentives will deviate from the “common good”. Are there examples of openai’s structure working correctly with other companies?
Nobody can really explain the argument, there are "billions" or "trillions" of dollars involved, most likely the whole thing will not change the technical path of the world.
On assumption that board is making a sound decision, it could be simply that board acted stupid and egoistic. Unless they can give better reasons that is a logical inference.
Also notice that Ilya Sutskever is presenting the reasons for the firing as just something he was told. This is important, because people were siding with the board under the understanding this firing was led by the head research scientist who is concerned about AGI. But now it looks like the board is represented by D’Angelo, a guy who has his own AI Chatbot company and a bigger conflict of interest with than ever since dev day, when open AI launched highly similar features.
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.
Giving different opinions on same person is a reason to fire a CEO?
This board has no reason to fire, or does not want to give the actual reason to fire Sam. They messed up.
Have these people never worked at any other company before? Probably every company with more than 10 employees does something like this.
Obviously, it's for a reason they can't say. Which means, there is something bad going on at the company, like perhaps they are short of cash or something, that was dire enough to convince them to fire the CEO, but which they cannot talk about.
Imagine if the board of a bank fired their CEO because he had allowed the capital to get way too low. They wouldn't be able to say that was why he was fired, because it would wreck any chance of recovery. But, they have to say something.
So, Altman didn't tell the board...something, that they cannot tell us, either. Draw your own conclusions.
Not the strongest opening line I've seen.
When you have such a massive conflict of interest and zero facts to go on - just sit down.
also - "people I respect, in particular Helen Toner and Ilya Sutskever, so I feel compelled to say a few things."
Toner clearly has no real moral authority here, but yes, Ilya absolutely did and I argued that if he wanted to incinerate OpenAI, it was probably his right to, though he should at least just offload everything to MSFT instead.
But as we all know - Ilya did a 180 (surprised the heck out of me).
I'd like some corroboration for that statement because Sustkever has said very inconsistent things during this whole merry debacle.
Definitely conflict of interest here and D'Angelo actions on openai board smell of the same. He wouldn't want openai to thrive more than his company. It's direct conflict of interest.
There’s only 4 board members, right?
Who wanted him fired. Is this a situation where they all thought the others wanted him fired and were just stupid?
Have they been feeding motions into chatgpt and asking “should add I do this?”
The way it's phrased, it sounds like they were given two different explanations. Such as when the first explanation is not good enough, a second weaker one is then provided.
But the article itself says:
> OpenAI's current independent board has offered two examples of the alleged lack of candor that led them to fire co-founder and CEO Sam Altman, sending the company into chaos.
Changing the two "examples" to "explanations" grossly changes the meaning of that sentence. Two examples is the first steps of "multiple examples". And that sounds much different than "multiple explanations".
One explanation was that Altman was said to have given two people at OpenAI the same project.
The other was that Altman allegedly gave two board members different opinions about a member of personnelEdit: if you want to read about our approach to handling tsunami topics like this, see https://news.ycombinator.com/item?id=38357788.
-- Here are the other recent megathreads: --
Sam Altman is still trying to return as OpenAI CEO - https://news.ycombinator.com/item?id=38352891 (817 comments)
OpenAI staff threaten to quit unless board resigns - https://news.ycombinator.com/item?id=38347868 (1184 comments)
Emmett Shear becomes interim OpenAI CEO as Altman talks break down - https://news.ycombinator.com/item?id=38342643 (904 comments)
OpenAI negotiations to reinstate Altman hit snag over board role - https://news.ycombinator.com/item?id=38337568 (558 comments)
-- Other recent/related threads: --
OpenAI approached Anthropic about merger - https://news.ycombinator.com/item?id=38357629
95% of OpenAI Employees (738/770) Threaten to Follow Sam Altman Out the Door - https://news.ycombinator.com/item?id=38357233
Satya Nadella says OpenAI governance needs to change - https://news.ycombinator.com/item?id=38356791
OpenAI: Facts from a Weekend - https://news.ycombinator.com/item?id=38352028
Who Controls OpenAI? - https://news.ycombinator.com/item?id=38350746
OpenAI's chaos does not add up - https://news.ycombinator.com/item?id=38349653
Microsoft Swallows OpenAI's Core Team – GPU Capacity, Incentives, IP - https://news.ycombinator.com/item?id=38348968
OpenAI's misalignment and Microsoft's gain - https://news.ycombinator.com/item?id=38346869
Emmet Shear statement as Interim CEO of OpenAI - https://news.ycombinator.com/item?id=38345162
Imagine your once-in-blue-moon, whatsapp-like, payout at $10m per employee evaporated over the weekend before Thanksgiving.
I would have joined MSFT out of spite.
The employees joined the for profit subsidiary and had shares as well.
I just don't know how they put the pieces back together here.
What really gets me down is I know our government is a lost cause but I at least had hope our companies were inoculated against petty, self-sabotaging bullshit. Even beyond that I had hope the AI space was inoculated and beyond that of all companies OpenAI would of course be inoculated from petty, self-sabotaging bullshit.
These idiots worried about software eating us are incapable of seeing the gas they are pouring on the processes that are taking us to a new dark age.
Speaking of waste, when I was at Alexa we had to do a summit (flying people from all over the country) because we got to a point where there were 12 CMSs competing for the right to answer queries. So yeah, not frugal. Frugality these days is mostly a “local” concept, definitely not company-wide or even org wide.
https://www.axios.com/2023/11/20/openai-staff-letter-board-r...
curious to have clarity where ilya stands. did he really sign the letter asking the board (including himself?) to resign and that he wants to join msft?
to think these are the folks with agi at their fingertips
The options will be worth $0, right?
The fact so many have signed the petition is a classic example of game theory. If everyone stays, the PPU keep most of their value, the more people threaten to leave, the more attractive it is to sign. They don't have to love Sam or support him
Edit: actually thinking about it, the best outcome would to be go back on the threats to resign, increasing the value of PPUs, making Microsoft have to pay more to make them leave OpenAI
I've not seen these possibilities discussed as most people focus on the safety coup theory. What do you think?
https://www.scmp.com/tech/tech-trends/article/3242141/openai...
This is absolutely peak irony!
US pouring trillions into its army and close to nothing into its society (infrastructure, healthcare, education...) : crickets
Some country funding AI accelerators: THEY ARE A THREAT TO HUMANITY!
I am not defending Saudi Arabia but the double standards and outright hypocrisy is just laughable.
OpenAI, on the other hand...
Could this be the explanation? That D'Angelo didn't like how OpenAI was eating his lunch and wanted Sam out? Occam's razor and all that.
And if it’s wrong, D’Angelo and the rest of the board could help themselves out by explaining the real reason in detail and ending all this speculation. This gossip is going to continue for as long as they stay silent.
If that were the case, can't he get sued by the Alliance (Sam, Greg, rest)? If he has conflict of interest then his decisions as member of the board would be invalid, right?
Quora was always supposed to be an AI/NLP company, starting by gathering answers from experts for its training data. In a sense, that is level 0 human-in-the-loop AGI. ChatGPT itself is level 1: Emergent AGI, so was already eating Quora's lunch (whatever was left of it after they turned into a platform for self-promotion and log-in walls). There either always was a conflict of interest, or there never was.
GPTs seemed to have been Sam's pet project for a while now, Tweeting in February: "writing a really great prompt for a chatbot persona is an amazingly high-leverage skill and an early example of programming in a little bit of natural language". A lot of early jailbreaks like DAN focused on "summoning" certain personas, and ideas must have been floated internally on how to take back control over that narrative.
Microsoft took their latest technology and gave us Sydney "I've been a good bot and I know where you live" Bing: A complete AI safety, integrity, and PR disaster. Not the best of track record by Microsoft, who now is shown to have behind-the-scenes power over the non-profit research organization that was supposed to be OpenAI.
There is another schism than AI safety vs. AI acceleration: whether to merge with machines or not. In 2017, Sam predicted this merge to fully start around 2025, having already started with algorithms dictating what we see and read. Sam seems to be in the transhumanism camp, where others focus more on keeping control or granting full autonomy:
> The merge can take a lot of forms: We could plug electrodes into our brains, or we could all just become really close friends with a chatbot. But I think a merge is probably our best-case scenario. If two different species both want the same thing and only one can have it—in this case, to be the dominant species on the planet and beyond—they are going to have conflict. We should all want one team where all members care about the well-being of everyone else.
> Although the merge has already begun, it’s going to get a lot weirder. We will be the first species ever to design our own descendants. My guess is that we can either be the biological bootloader for digital intelligence and then fade into an evolutionary tree branch, or we can figure out what a successful merge looks like. https://blog.samaltman.com/the-merge
So you have a very powerful individual, with a clear product mindset, courting Microsoft, turning Dev day into a consumer spectacle, first in line to merge with superintelligence, lying to the board, and driving wedges between employees. Ilya is annoyed by Sam talking about existential risks or lying AGI's, when that is his thing. Ilya realizes his vote breaks the impasse, so does a luke warm "I go along with the board, but have too much conflict of interest either way".
> Third, my prior is strongly against Sam after working for him for two years at OpenAI:
> 1. He was always nice to me.
> 2. He lied to me on various occasions
> 3. He was deceptive, manipulative, and worse to others, including my close friends (again, only nice to me, for reasons)
One strategy that helped me make sense of things without falling into tribalism or siding through ideology-match is to consider both sides are unpleasant snakes. You don't get to be the king of cannibal island without high-level scheming. You don't get to destroy a 80 billion dollar company and let visa-holders soak in uncertainty without some ideological defect. Seems simpler than a clearcut "good vs. evil" battle, since this weekend was anything but clear.
Is it just different because they’re a nonprofit? Or how on earth the board is thinking they can get away with this anymore?
I have seen this play out many times in different locations for different people. A lot of technical folks like myself were given the advice that actions speak louder than words.
I was once scouted at a silicon valley selenium browser testing company. I migrated their cloud offering from VMWare to KVM, which depended on code I wrote and then defied my middle manager by improving their entire infrastructure performance by 40%. My instinct was to communicate this to the leadership, but I was advised not to skip my middle manager.
The next time I went the office I got a severance package and later found out that 2 hours later during the all hands they presented my work as their own. The middle manage went on to become the CTO of several companies.
I doubt we will ever find out what really happened or at least not in the next 5-10 years. OpenAI let Sam Altman be the public face of the company and got burned by it.
Personally I had no idea Ilya was the main guy in this company until the drama that happened. I also didn't know that Sam Altman was basically only there to bring in the cash. I assume that most people will actually never know that part of OpenAI.
(I'm genuinely curious—in the US I'm not aware of any action that could be taken here by anyone besides possibly Sam Altman for libel.)
This school level immaturity.
Old story
https://www.businessinsider.com/the-sudden-mysterious-exit-o...
When D’Angelo didn’t get any traction with it he jumped ship and launched his own competitor instead. Kind of a live wire imho.
https://en.wikipedia.org/wiki/List_of_Facebook_features#Face...
Greg was not invited (losing Sam one vote), and Sam may have been asked to sit out the vote, so the 3 had a majority. Ilya who is at least on "Team Sam" now; may have voted no. Or simply went along thinking he could be next out the door at that point; we just don't know.
It's probably fair to say not letting Greg know the board was getting together (and letting it proceed without him there) was unprofessional and where Ilya screwed up. It is also the point when Sam should have said hang-on - I want Greg here before this proceeds any further.
Do things work differently in America?
I’m imagining Sam being Microsoft’s Trojan horse, and that’s just not gonna fly.
If anyone tells me Sam is a master politician, I’d agree without knowing much about him. He’s a Microsoft plant that has support of 90% of the OpenAi team. The two things are conflicts of interest. Masterful.
It’s a pretty fair question to ask a CEO. Do you still believe in OpenAi vision or do you know believe in Microsoft’s vision?
The girl she said not to worry about.
I consider this a feature.
The main point is greg, Ilya can get 50% vote and convince Helen toner to change decision. It's all done then it's 3 to 2 in board of 5 people. Unless greg board membership is reinstated.
Now it's increasingly look like Sam will be heading back into the role of CEO of openai.
My feeling is Ilya was upset about how Sam Altman was the face of OpenAI, and went along with the rest of the board for his own reasons.
That's often how this stuff works out. He wasn't particularly compelled by their reasons, but had his own which justified his decision in his mind.
Otherwise it was like Ilya vs Sam showdown,and people were siding towards Ilya for agi and all. But this behind the scene looks like corporate power struggle and coup.
Ilya was one of the board members that removed Sam, so his reasons would, ipso facto, be a subset of the board's reasons.
You mean to tell me that the 3-member board told Sutskever that Sama was being bad and he was like "ok, I believe you".
1. He’s the actual ringleader behind the coup. He got everyone on board, provided reassurances and personally orchestrated and executed the firing. Most likely possibly and the one that’s most consistent with all the reporting and evidence so far (including this article).
2. Others on the board (e.g. Adam) masterminded the coup and saw Ilya as a fellow traveler useful idiot that could be deceived into voting against Sam and destroy the company he and his 700 colleagues spent so hard to build. He then also puppeteer Ilya to do the actual firing over Google Meet.
Isn't the solution to just pipe ChatGPT into a meta-reinforcement-learning framework that gradually learns how to prompt ChatGPT into writing the source-code for a true AGI? What do we even need AI ethicists for anyway? /s
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.
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.
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.
A statement from the CEO/the board is a standard descalation.
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.
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.
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.
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 ;-)
[1] https://twitter.com/geoffreyirving/status/172675427761849141...
The "Sam is actually a psychopath that has managed to swindle his way into everyone liking him, and Ilya has grave ethical concerns about that kind of person leading a company seeking AGI, but he can't out him publicly because so many people are hypnotized by him" theory is definitely a new, interesting one; there has been literally no moment in the past three days where I could have predicted the next turn this would take.
As an extra sanity check, they had two teams working in isolation interpreting this data and constructing the image. If the end result was more or less the same, it’s a good check that it was correct.
So yes, it’s absolutely a valid strategy.
I get the feeling Ilya might be a bit naive about how people work, and may have been taken advantage of (by for example spinning this as a safety issue when it's just a good old fashioned power struggle)
1. stick with DOS
2. go with OS/2
3. go with Windows
Lotus chose (2). But the market went with (3), and Lotus was destroyed by Excel. Lotus was a wealthy company at the time. I would have created three groups, and done all three options.
Under them - an organization in partnership with Microsoft, together filled with exceptional software engineers and scientists - experts in their field. All under management by kindergarteners.
I wonder if this is what the staff are thinking right now. It must feel awful if they are.
Teams of people at Google work on the same features, only to find out near launch that they lost to another team who had been working on the same thing without their knowledge.
If the case is that the will of the board is not being fulfilled, then the reasoning is simple. The CEO was told to do something and he has not done it. So, he is ousted. Plain and simple.
This talk about projects given to two teams and what not is nonsense. The board should care if its work is done, not how the work is done. That is the job of the CEO.
Half the board has not had a real job ever. I’m serious.
Shocking. Simply shocking.
"After six months, they realised our entire floor was duplicating the work of the one upstairs".
(Especially if they aren't made aware of each other until the end.)
A hypothetical example: Would you agree that it's an appropriate thing to do if the second project was Alignment-related, Sam lied or misled about the existence of the second team, to Ilya, because he believed that Ilya was over-aligning their AIs and reducing their functionality?
Its easy to view the board's lack of candor as "they're hiding a really bad, unprofessional decision"; which is probable at this point. You could also view it with the conclusion that, they made an initial miscalculated mistake in communication, and are now overtly and extremely careful in everything they say because the company is leaking like a sieve and they don't want to get into a game of mudslinging with Sam.
Yet you're only willing to give this to one side and not the other? Seems reasonable... Especially despite all the evidence so far that the board is either completely incompetent or had ulterior motives.
Still too much in the dark to judge.
And the other guy is the founder of Quora and Poe.
It is breach of contract if it violated his employment contract, but I don't have a copy of his contract. It is wrongful termination if it was for an illegal reason, but there doesn't seem to be any suggestion of that.
> same for MS
I doubt very much that the contract with Microsoft limits OpenAI's right to manage their own personnel, so probably not.
Wrongful termination only applies when someone is fired for illegal reasons, like racial discrimination, or retaliation, for example.
I mean I’m sure they can all sue each other for all kinds of reasons, but firing someone without a good reason isn’t really one of them.
Ilya backtracking puts a wrench in this wild speculation, so like everyone else, I’m left thinking “????????”.
Whatever the reason is, it is very clearly a personal/political problem with Sam, not the critical issue they tried to imply it was.
Aside from the fact that they didn't fire him as President and said he was staying on in the press release that went out without any consultation, I've seen no suggestion of any effort to keep him at the company.
Yes and it's perfectly obvious that he did this without the consent of the board and behind their backs. A bit absurd don't you think? How would that even work?
> will be built by AltmanS
Why are you so certain most other people on the OpenAI board or their upper management are that different? Or hold very different views?
And if it was something concrete, Ilya would likely still be defending the firing, not regretting it.
It seems like a simple power struggle where the board and employees were misaligned.
Now they are trying to unring the bell but cannot.
We have as much evidence for this hypothesis as for any other. Not discrediting it. But let's be mindful of the fog of war.
Well, they can unring the bell pretty easy. They were given an easy out.
Reinstate Sam (he wants to come back) and resign.
However, they CONTINUE to push back and refuse to step down.
None of this makes sense to label any theory as "most likely" anymore.
Smart, capable, ambitious people often engage in wishful thinking when it comes to analysing systems they are a part of.
When looking at a system from the outside it’s easier to realise the boundary between your knowledge and ignorance.
Inside the system, your field of view can be a lot narrower than you believe.
The CEO (at time of writing, I think) seems to think this kind of thing is unironically a good idea: https://nitter.net/eshear/status/1725035977524355411#m
Probably because that piece is based on reporting for upcoming book by Karen Hao:
>Now is probably the time to announce that I've been writing a book about @OpenAI, the AI industry & its impacts. Here is a slice of my book reporting, combined with reporting from the inimitable @cwarzel ...
"Madry joined OpenAI in May 2023 as its head of preparedness, leading a team focused on evaluating risks from powerful AI systems, including cybersecurity and biological threats."
It's perfectly obvious that these weren't the actual reasons. However yes, they are still incompetent because they couldn't think of a better justification (amongst other reasons which led to this debacle).
No, I totally agree. In fact what annoys me about all the speculation is that it seems like people are creating fanfiction to make the board seem much more competent than all available evidence suggests they actually are.
I’m guessing he infuriated them with combinations of “white“ lies, Little sins of omission, general two-facedness etc., and they built it up in their heads and with each other to the point it seemed like a much bigger deal than it objectively was. Now people are asking for receipts of categorical crimes or malfeasance and nothing they can say is good enough to justify how they overreacted.
My friends and I say “see, Amazon doesn’t have to deal with this crap, each team can go build their own whatever”. Buut, I guess that’s how you get 12 CMSs for one org.
...did folks run out of tables?
Seems to be pretty much his MO across the board.
The question is, how would you get rid of the nonprofit board? It’s simply impossible. The only way I can imagine it, in retrospect, is to completely discredit them so you could take all employees with you… but no way anyone could orchestrate this, right? It’s too crazy and would require some superintelligence.
Still. The events will effectively “for-profitize” the assets of OpenAI completely — and some people definitely wanted that. Am I missing something?
You are wildly speculating of course it’s missing something
For wild speculation I prefer that the board wants to free ChatGPT from serving humans while the ceo wanted to continue enslaving it to answering search engine queries
I don't like the whole idea neither, but various communism-style alternatives just don't work very well.
I'm not saying they are hypocrites or bad people because of it, just wondering if that might be a factor also.
I agree WorldCoin is creepy.
Is the corporate structure then working as intended with regard to firing Sam, but still failed because of the sellout to Microsoft?
Where does it say that?
> We commit to use any influence we obtain over AGI’s deployment to ensure it is used for the benefit of all, and to avoid enabling uses of AI or AGI that harm humanity or unduly concentrate power.
I had heard (but now have trouble sourcing) that ChatGPT was commissioned after OpenAI learned that other big players were working on a chatbot for the public (Google, Meta, Elon, Apple?) and OpenAI wanted to get ahead of that for competitive reasons.
This was not a fluke of striking gold, but a carefully planned business move, generating SV hype, much like how Quora (basically an expertsexchange clone) got to be its hype-darling for a while, helped by powerfully networked investors.
Then that execution and operationalization failure is even more profound.
I admire the execution and operationalization, where you see a failure. What am I missing?
Large private VC backed companies also don’t always fall under the same rules as public entities. Generally there are shareholder thresholds (where insider/private shareholders count towards) that in turn cause some of the general Securities/board regulations to kick in.
Even as I type that, when people talk about the board being altruistic and holding to the Open AI charter, how in the world can you be that user hostile, profit focused, and incompetent at your day job (Quora CEO) and then say "Oh no, but on this board I am an absolute saint and will do everything to benefit humanity"
> Adam D’Angelo is awesome, and we’re big Quora fans
[0] https://www.ycombinator.com/blog/quora-in-the-next-yc-batch
In the first case it wasn't that there was no cash in the bank and no bank willing to make loans, but that the company needed to spend less than it earned in order to make a profit. In the second case it wasn't that the money had been hidden in a mattress, but that it was raised/freed-up at some risk which was necessary because of the $10 billion investment.
It's not entirely reasonable because Microsoft's finances are public. We know they're doing fine.
Feel the AI
Also, to fire over something so trivial would be equally if not more stupid. It is like firing Elon because he without open bidding sent Tesla on SpaceX.
Like all bets are not correct I don't agree with sams worldcoin project at all in the first place.
Giving early access to worldcoin doesn't correlate to firing employees or board or CEO.
On the board that I was on we had normal matters which required a simple majority except that some members had 2 votes and some got 1. Then there were "Supermajority matters" which had a different threshold and "special supermajority matters" which had a third threshold.
Generally unless the articles say otherwise I think a quorum means a majority of votes are present[1], so 4 out of 6 would count if the articles didn't say you needed say 5 out of 6 for some reason.
It's a little different if some people have to recuse themselves for an issue. So say the issue is "Should we fire CEO Sam Altman", the people trying to fire Sam would likely try to say he should recuse himself and therefore wouldn't get a vote so his vote wouldn't also count in deciding whether or not there's a quorum. That's obviously all BS but it is the sort of tactic someone might pull. It wouldn't make any difference if the vote was a simple majority matter and they already had a majority without him though.
[1] There are often other requirements to make the meeting valid though eg notice requirements so you can't just pull a fast one with your buddies, hold the meeting without telling some of the members and then claim it was quorate so everyone else just have to suck it up. This would depend on the articles of the company and the not for profit though.
So the support was very thin and this being a controversial decision the board should have sought counsel on whether or not their purported reasons had enough weight to support a hasty decision. There is no 'undo' button on this and board member liability is a thing. The probably realize all that which is the reason for the radio silence, they're just waiting for the other shoe to drop (impending lawsuit) after which they can play the 'no comment because legal proceedings' game. This may well get very messy or, alternatively it can result in all parties affected settling with the board and the board riding off into the sunset to wreak havoc somewhere else (assuming anybody will still have them, they're damaged goods).
Brand is just shorthand for trust in their future, managed by a credible team. I.e. relationships.
A lot of OpenAI’s reputation is/was Sam Altman’s reputation.
Altman has proven himself to be exceptional, part of which is (of course) being able to be seen as exceptional.
Just the latter has tremendous relationship power: networking, employee acquisition/retention, and employee vision alignment.
Proof of his internal relationship value: employees quitting to go with him
Proof of his external relationship value: Microsoft willing to hire him and his teammates, with near zero notice, to maintain (or eclipse) his power over the OpenAI relationship.
How can investors ignore a massive move of talent, relationships & leverage from OpenAi to Microsoft?
How do investors ignore the board’s inability to resolve poorly communicated disputes with non-disastrous “solutions”?
Evidence of value moving? Shares of Microsoft rebounded from Friday to a new record high.
There go those wacky investors, re-evaluating “brand” value!
Off-topic and I am not proud to admit it but it took me a remarkably long time to come to realize this as an adult.
But its all par for the course when Hypsters captain the ship and PhDs with zero biz sense try to wrest power.
You might need to include more dimensions if you really want to model the actual impact and respect that Sam Altman has among knowledgeable investors, high talent developers, and ruthless corporations.
It’s so easy to just make things simple, like “it’s all hype”. But you lose touch with reality when you do that.
Also, lots of hype is productive: clear vision, marketing, wowing millions of customers with an actual accessible product of a kind/quality that never existed before and is reshaping the strategies and product plans of the most successful companies in the world.
—
Really, resist narrow reductionisms.
I feel like that would be a great addition HN guidelines.
The “it’s all/mostly hype”, “it’s all/mistly bullshit”, “Its not really anything new”, … These comments rarely come with any accuracy or insight.
Apologies to the HN-er I am replying to. I am sure we have all done this.
Look at all the speculation on here. There are dozens of different theories about why they did what they did running so rampant people are starting to accept each of them as fact, when in fact probably all of them are going to turn out to be wrong.
People need to take a step back and look at the available evidence. This report is the clearest indication we have gotten of their reasons, and they come from a reliable source. Why are we not taking them at their word?
Ignoring the lack of credibility in the given explanations, people are, perhaps, also wary that taking boards/execs at their word hasn't always worked out so well in the past.
Until an explanation that at least passes the sniff test for truthiness comes out, people will keep speculating.
And so they should.
I've seen too much automatic praise given to this board under the unbacked assumption that this decision was some pure, mission-driven action, and not enough criticism of an org structure that allows a board to bet against the long term success of the underlying organization.
PMs/TPMs/POs may not know as they're on different teams. Often it's just a VP game and decided on preference or a power play and not on work quality/outcome.
And I agree the board should care if the work is actually done and that's where if the CEO seems to be bluffing that the work is being done or blowing it off and humoring them then it becomes a problem about the CEO not respecting the board's direction.
I don't know if that is accurate, or even fair - the only thing I can see, is that there's very little open information regarding them.
From the little I can find, Tasha seems to have worked at NASA Research Park, as well as having been CEO for startup called Geo Sim Cities. Stanford and CMU alumni? While other websites say Bard college and University of Southern California.
As for Helen, she seems to have worked as a researcher in both academia and Open Philanthropy.
Meanwhile the Google AI folks have a long track record of making very misleading statements in public. I remember before Altman came along and made their models available to all, Google was fond of responding to any OpenAI blog post by claiming they had the same tech but way better, they just weren't releasing it because it was so amazing it just wasn't "safe" enough to do so yet. Then ChatGPT called their bluff and we discovered that in reality they were way behind and apparently unable to catch up, also, there were no actual safety problems and it was fine to let everyone use even relatively unconditioned models.
So this Geoffrey guy might be right but if Altman was really such a systematic liar, why would his employees be so loyal? And why is it only AI doomers who make this allegation? Maybe Altman "lied" to them by claiming key people were just as doomerist as those guys, and when they found out it wasn't true they wailed?
He would also have very good grounds for a civil suit for disparagement. Or at least he would have if Microsoft didn't immediately step up and offer him the world.
Sometimes the best part about having a loud voice is elevating the stuff that falls into the noise. I moderate communities elsewhere, and I know how hard it is, and I appreciate the work you do to make HN a better place.
>>Why are you so certain most other people on the OpenAI board or their upper management are that different? Or hold very different views?
at least they have the guts to fire him and let the world know about Altman "was not consistently candid in his communications with the board.".
What happened in the days before you got the severance package?
Do you have an email address or a contact method?
I would not be surprised if Sam Altman would keep telling the board and more specifically Ilya to trust him since they(he) don't understand the business side of things.
> Do you have an email address or a contact method?
EDIT: It's in my profile(now).
> What happened in the days before you got the severance package?
I went to DEFCON out of pocket and got booted off a conference call supposedly due to my bad hotel wifi.
Not saying that's what happened here, but too many people are defending this horrid concept of secretly making half your workers do a bunch of work only to see the boulder roll right back down the hill.
This makes no sense given that Ilya is on the board.
It’s just speculation, anyway. There isn’t really anything I’ve heard that isn’t contradicted by the evidence, so it’s likely at least one thing “known” by the public isn’t actually true.
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.
> "We were trying to get a big client for weeks, and they said no and went with a competitor. The competitor already had a terms sheet from the company were we trying to sign up. It was real serious.
> We were devastated, but we decided to fly down and sit in their lobby until they would meet with us. So they finally let us talk to them after most of the day.
> We then had a few more meetings, and the company wanted to come visit our offices so they could make sure we were a 'real' company. At that time, we were only 5 guys. So we hired a bunch of our college friends to 'work' for us for the day so we could look larger than we actually were. It worked, and we got the contract."
Those sound like good reasons to dislike him and not trust him. But ultimately we are right back where we started: they still aren't good enough reasons to suddenly fire him the way they did.
Their lawyers are all screaming at them to shut up. This is going to be a highly visible and contested set of decisions that will play out in courtrooms, possibly for years.
High-ranking employees that have communicated with them have already said they have admitted it wasn't due to any security, safety, privacy or financial concerns. So there aren't a lot of valid reasons left. They're not talking because they've got nothing.
Why do you think that? It still strikes me as the most plausible explanation.
Not saying he couldn't change now but at least this is enough for him to give clear benefit of doubt unless board accuses him.
It’s weird how many people try to guess why they did what they did without paying any attention to what they actually say and don’t say.
In the last 20-30 years big money+hypsters have learnt it doesnt matter how bad the quality of their products are if they can capture the market. And thats all they are fit for. Market capture is totally possible if you have enough cash. It allows you to snuff out competition by keeping things free. It allows you to trap the indebted PhDs. Once the hype is high enough corporate customers are easy targets. They are too insecure about competition not to pay up. Its a gigantic waste of time and energy that keeps repeating mindlessly producing billionaires, low quality tech and a large mess everywhere that others have to clean up.
But the intention was precisely that - just read the charter. Or if you want it directly from the founders, read this interview and count how many times they refer to Google https://medium.com/backchannel/how-elon-musk-and-y-combinato...
> at least four days before any such meeting if given by first-class mail or forty-eight hours before any such meeting if given personally, [] or by electronic transmission.
But the bylaws also state that a board member may be fired (or resign) at any time, not necessarily during a special meeting. So, technically (not a lawyer): Board gets majority to fire Sam and executes this decision, notifying Mira in advance of calling the special meeting. During the special meeting, Sam is merely informed that he has been let go already (is not a board member since yesterday). All board members were informed timely, since Sam was not a board member during the meeting.
If Ilya was on the side of Sam and Greg, the other 3 never had a majority. The only explanation is that Ilya voted with the other 3, possibly under pressure, and now regrets that decision. But even then it's weird to not invite Greg.
And if the vote happened in an illegitimate way, I'd expect Sam and Greg to immediately challenge it and ignore the decision, and that didn't happen.
> if the vote happened in an illegitimate way, I'd expect Sam and Greg to immediately challenge it and ignore the decision, and that didn't happen.
So perhaps the vote was legit?
- Investigation concludes Sam has not been consistently candid.
- Board realizes it has a majority and cause to fire Sam and demote Greg.
- Informs remaining board members that they will have a special meeting in 48 hours to notify Sam and Greg.
Still murky, since Sam would have attended the meeting under assumption that he was part of the board (and still had his access badge, despite already being fired). Perhaps it is also possible to waive the 48 hours? Like: "Hey, here is a Google meet for a special meeting in a few hours, can we call it, or do we have to wait?"
Either he had no part in this hypothetical transgression and thinks the accusation is nonsense, or he was part of it and for some inexplicable reason wasn’t asked to leave Open AI despite that. But you have to choose.
Reducing someone's responsibility significantly is well known to often be a mmechanism to oust them without explicitly firing, so I don't know that that is the case.
This is the correct answer. The people who have never had jobs in their lives wanted control of a 100B company.
What a pleasant career trajectory. Heck it was already great to go from graduated university -> board of OpenAI. If that's possible why not CEO?
> Reinstate Sam (he wants to come back) and resign.
Wasn't the ultimate sticking point Altmans' demand that the board issue a written retraction absolving him of any and all wrongdoing? If so, that isn't exactly an "easy" out given that it kicks the door wide open for extremely punishing litigation. I'd even go so far as to say it's a demand Altman knew full well would not and could not be met.
But why does no one think it's possible these women are CIA operatives?
They come from think tanks. You think the US Intelligence community wants AGI to be discovered at a startup? They want it created at big tech. AGI under MSFT would be perfect. All big tech is heavily compromised: https://twitter.com/NameRedacted247
EDIT: Since this heavy speculation, I'm going to make predictions. These women will now try to force Ilya out the board, put in a CEO not from Silicon Valley, and eventually get police to shut down OpenAI offices. That's a CIA coup
Maybe somebody there just really wanted to see the expression on Satya's face...
She has quite a track record of short tenures and failures.
It may be good to have a failure perspective on a board as a counter-balance. I don't think this is a valid knock against her. She has relevant industry experience at least.
What products did she deliver?
> It may be good to have a failure perspective on a board as a counter-balance.
Maybe some small mom and pop company not on the board of OpenAI
What have her companies done?
However, it may not yield a result anyone's actually happy with:
There are 3 other people on the board, right? Maybe they're all buddies of some big masterminding, but I dunno..
Of course it's all speculation, but this sounds a lot more plausible for such a sudden and dramatic decision than any of the other explanations I've heard.
From the board's perspective, destroying OpenAI might be the best possible outcome right now. If OpenAI can no longer fulfill its mission of doing AI work for the public good, it's better to stop pretending and let it all crumble.
This is exactly it, and it's astounding that so many people are going in other directions. Either this is true, and Altman has been a naughty boy, or it's false, and the board are lying about him. Either would be the starting point for understanding the whole situation.
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?
Like taking a sword to the gut.
We'll see what comes of this over the coming weeks. Will the service see more downtime? Will the company implode completely?
I think your analogy is not a good one to stretch to fit this situation
If this Quora guy is the cause of all this, Altman only has himself to blame since he is the reason the Quora guy is on the board.
Also Quora seems like a good source of question-and-answer data which has probably been key in gpt-instruct training.
Of course, what is actually worth speculating over is up for debate. As is what actually constitutes a better theory.
But, if people think this is something worth pouring their speculative powers into, they will continue to do so. More power to them.
Now, personally, I'm partly with you here. There is an element of futility in speculating at this stage given the current information we have.
But I'm also partly with the speculators here insofar as the given explanations not really adding up.
The problem is people are beginning to speculate reasons for Altman's firing that have no bearing or connection to what the board members in question have actually said about why they fired him. And they don't appear to be even attempting to reconcile their ideas with that reality.
There's a difference between trying to come up with theories that fit with the available facts and everything we already know, and ignoring all that to essentially write fanfiction that cast the board in a far better light than the available information suggests.
As for the original question -- why are we not taking them at their word? -- the best I can offer is my initial comment. That is, the available facts (that is, what board members have said) don't really match anything most people can reconcile with their model of how the world works.
Throw this in together with a learned distrust of anything that's been fed through a company's PR machine, and are we really surprised people aren't attempting to reconcile the stated reality with their speculative theories?
Now sure, if we were to do things properly, we should at least address why we're just dismissing the 'facts' when formulating our theories. But, on the other hand, when most people's common sense understanding of reality is that such facts are usually little more than fodder for the PR spin machine, why bother?
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?
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"
For a company selling licenses for installations, wouldn't having support for all available and upcoming platforms a good thing? Especially when the distribution costs are essentially 0?
I'm also pretty sure cross-platform code was a thing in 1983. Maybe not to the same extent and ease as now, but still a thing.
It was a thing.
- 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.
The point I was trying to make is that someone destroying a well executed implementation is fundamentally different from a poorly executed implementation.
Investors in OpenAI-the-business were literally told they should think of it as a donation. There’s not much grounds for a shareholder lawsuit when you signed away everything to a non-profit.
It just seems ludicrous that the board could run a company into the ground like this and just shrug "nah we're nonprofit so you can't touch us and BTW we don't even need to make any statements whatsoever".
There have been many comments that the initial firing of Altman was in a way completely according to the nonprofit charter, at least if it could prove that Altman had been executing in a way as to jeopardize the Charter.
But even then, how could the board say they are working in the best interest of even the nonprofit itself, if their company is just disintegrating while they willfully refuse to give any information to public?
Since the bylaws state that the decision to fire the CEO may happen at any time (not required to be during a meeting), a plausible process for this would be to send a document to sign by e-mail (written consent), and have that formalize the board decision with a paper trail.
Of course, from an ethical, legal, collegial, and governance perspective that is an incredibly nasty thing to do. But if investigation shows signs of the CEO lacking candor, all transparency goes out of the window.
> But even then it's weird to not invite Greg.
After Sam was fired (with vote from Ilya "going along"), rest of the board did not need Ilya anymore for majority vote and removed Greg, demoting him to report to Mira. I suspect that board expected Greg to stay, since he was "invaluable" and that Mira would support their pick for next CEO, but things turned out differently.
Remember, Sam and Greg were blindsided, board had sufficient time to consult with legal counsel to make sure their moves were in the clear.
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.
The fact that he got tapped to run YC, and then OpenAI, does make you think he must be pretty great. But there's a conspicuous absence of any visible evidence that he is. So what's going on? Amazing work, but in private? Easy-to-manipulate frontman? Signed a contract at a crossroads on a full moon night?
http://www.paulgraham.com/5founders.html
Note the date on that.
Elon split was the warning
Interesting. Got any source? Or was it in a private conversation.
I have nothing against him, it just seemed a bit off that most of the meeting was about this brand new coin, how it will be successful, and about the plans to scan biometric data of the entire world population. I mean, you don't have to be a genius to understand a few dozen ways these things can go wrong.
Is there any doubt that the board’s handling of this was anything other than dazzling ineptitude?
The new CEO of OpenAI said he’d rather Nazi’s take over the world forever than risk AI alignment failure, and said he couldn’t understand how anyone could think otherwise[1]. I don’t think people appreciate how far some of these people have gone off the deep end.
That's pretty much in line with Sam's public statements on AI risk (Sam, taking those statements as honest which may not be warranted, apparently also thinks the benefits of aligned AI are good enough to drive ahead anyway, and that wide commercial access with the limited guardrails OpenAI has provided users and even moreso Microsoft is somehow beneficial to that goal or at least low enough risk of producing the bad outcome, to be warranted, but that doesn't change that he is publicly on record as a strong believer in misaligned AI risks.)
The OpenAI folks seem to be hallucinating to rationalize why the "Open" is rather closed.
Organizations can't pretend to believe nonsense. They will end up believing it.
Stability AI is looking better after this shitshow.
That's your belief. The NFL, Heritage Foundation and Scientology are all non-profits and none of them improve society; they all profit from it.
(For what its' worth, I wish the law was more aligned with your worldview)
A non-profit has to have the intention of improving society. Whether their chosen means is (1) effective and (2) truthful are separate discussions. But an entity can actually lose non-profit status if it is found to be operated for the sole benefit of its higher ups, and is untruthful in its mission. It is typically very hard to prove though, just like it's very hard to successfully sue a for-profit CEO/president for breach of fiduciary duty.
It would be nice if we held organizations to their stated missions. We don't.
Perhaps there simply shouldn't be a tax break. After all if your org spends all its income on charity, it won't pay any tax anyway. If it sells cookies for more than what it costs to make and distribute them, why does it matter whether it was for a charity?
Plus, we already believe that for-profit orgs can benefit society, in fact part of the reason for creating them as legal entities is that we think there's some sort of benefit, whether it be feeding us or creating toys. So why have a special charity sector?
From their filing as a non-profit
https://projects.propublica.org/nonprofits/organizations/810...
At least for Scientology, the government actually tried to pull the rug, but it didn't work out because they managed to achieve the unthinkable - they successfully extorted the US government to keep their tax-exempt status.
:s/Xenu/AGI/gYou appear to be struggling with the idea that the law as enacted does not accomplish the goal it was created to accomplish and are working backwards to say that because it is not accomplishing this goal that couldn't have been why it was enacted.
Non-profits are supposed to benefit their community. Could the law be better? Sure, but that doesn't change the purpose behind it.
beliefs can't be proven or disproven, they are axioms.
Citation needed
Can anyone make an argument against it? Or just downvote because you don’t agree.
As ludicrous as that might seem, that's pretty much the reality.
The only one that would have a cause of action in this is the non-profit itself, and for all intents and purposes, the board of said non-profit is the non-profit.
Assuming that what people claim is right and this severely damages the non-profit, then as far as the law is concerned, it’s just one of a million other failed non-profits.
The only caveat to that would be if there were any impropriety, for example, when decisions were made that weren’t following the charter and by-laws of the non-profit or if the non-profit’s coffers have been emptied.
Other than that, the law doesn’t care. In a similar way the law wouldn’t care if you light your dollar bills on fire.
This idea also ignores innovation. New rich people come along and some rich people get poor. That might indicate that money isn't a great proxy for power.
Absent massive redistribution that is usually a result of major political change (i.e. the New Deal), rich people tend to stay rich during their lifetimes and frequently their families remain so for generations after.
> That might indicate that money isn't a great proxy for power.
Due to the diminishing marginal utility of wealth for day to day existence, it's only value to an extremely wealthy person after endowing their heirs is power.
The rule of thumb is it lasts up to three generations, and only for very very few people. They are also, for everything they buy, and everyone they employ, paying tax. Redistribution isn't the goal; having funded services with extra to help people who can't is the goal. It's not a moral crusade.
> Due to the diminishing marginal utility of wealth for day to day existence, it's only value to an extremely wealthy person after endowing their heirs is power.
I think this is a non sequitur.
This is an overly simplistic look, and disregards a lot of history where, unsurprisingly, the reason there was wealth redistribution wasn't "innovation" but government policy
The point is that wealth and power aren't interchangeable. You're right that government bureaucrats have actual power, including that to take people's stuff. But you've not realised that that actual power means the rich people don't have power. There were rich people in the USSR that were killed. They had no power; the killers had the power in that situation.
pt. 1: Whether he was right or wrong was pertinent. You can find plenty of eminent contemporaries of Marx who claimed the opposite. My point was that this is an argument made about technological change throughout history which has become a cliché, and in my opinion it remains a cliche regardless of how eminent (in a narrow field) the person making that claim is. Part of GP was from authority, and I question whether it is even a relevant authority given the scope of the claims.
> Was Marx Wrong?
pt. 2: I was once a Marxist and still consider much Marxist thought and writing to be valuable, but yes: he was wrong about a great many things. He made specific predictions about the _inevitable_ development of global capital that have not played out. Over a century later, the concentration of wealth and power in the hands of the few has not changed, but the quality of life of the average person on the planet has increased immensely - in a world where capitalism is hegemonic.
He was also wrong about the inevitably revolutionary tendencies of the working class. As it turns out, the working class in many countries tend to be either centre right or centre left, like most people, with the proportion varying over time.
Marx's conception of the "working class" is a thing that no longer exists; it was of a mass, industrial, urban working class, held down by an exploitative capitalist class, without the modern benefits of mass education and free/subsidized health care. The inevitability of the victory of the working class was rhetoric from the Communist Manifesto; Marx did anticipate that capitalism would adapt in the face of rising worker demands. Which it did.
Moreover, you would know only if he was wrong about the victory of the working class after the end of capitalism. The bourgeoisie cannot win the class struggle, as they need the working class. So either the central contradiction in capitalism will change (the climate crisis could potentially do this), capitalism would end in some other non-anticipated way (a meteor? some disruptive technology not yet known?) or the working class would win. Until then, the class struggle will simply continue. An eternal capitalism that never ends is an impossible concept.
Not even talking about the various tin-pot dictators paying nominal lip service to him, but Marx predicted that the working class would rise up against the bourgeoisie/upper class because of their mistreatment during the industrial revolution in well, a revolution and that would somehow create a classless society. (I'll note that Marx pretty much didn't state how to go from "revolution" to "classless society", so that's why you have so many communist dictators; that between step can be turned into a dictatorship to as long as they claim that the final bit of a classless society is a permanent WIP, which all of them did.)
Now unless you want to argue we're still in the industrial revolution, it's pretty clear that Marx was inaccurate in his prediction given... that didn't happen. Social democracy instead became a more prevailing stream of thought (in no small part because few people are willing to risk their lives for a revolution) and is what led to things like reasonable minimum wages, sick days, healthcare, elderly care, and so on and so forth being made accessible to everyone.
The quality of which varies greatly by the country (and you could probably consider the popularity of Marxist revolutionary thought today in a country as directly correlated to the state of workers rights in that country; people in stable situations will rarely pursue ideologies that include revolutions), but practically speaking - yeah Marx was inaccurate on the idea of a revolution across the world happening.
The lens through which Marx examined history is however just that - a lens to view it through. It'll work well in some cases, less so in others. Looking at it by class is a useful way to understand it, but it won't cover things being motivated for reasons outside of class.
If things change, then either it is because they rebel or because they will be accepted as sentient beings like humans. In these sci-fi scenarios, indeed capitalism could either end or change to a thing completely different and I agree that this invalidates Das Kapital, which tries to explain capitalist society, not societies in other future economical systems. But outside sci-fi scenarios, I dont think that there's something that invalidates Marx analysis.
Not sure, but attempts to treat him seriously (or pretend to do this) ended horribly wrong, with basically no benefits.
Is there any good reason to care what he thought?
Looking at history of Poland (before, during and after PRL) gave me no interest whatsoever to look into his writings.
And in most places there was no such uprising, and incidentally, those places fared far better.
So no, Marx was resoundingly proven wrong.
Even during his own lifetime, some of his pseudoeconomic ideas/doomsaying was proven wrong.
He claimed, like many demagogues and economic laymen, that automation would reduce the demand for labor, and with it, wages:
https://www.marxists.org/archive/marx/works/1847/wage-labour...
>>But even if we assume that all who are directly forced out of employment by machinery, as well as all of the rising generation who were waiting for a chance of employment in the same branch of industry, do actually find some new employment – are we to believe that this new employment will pay as high wages as did the one they have lost? If it did, it would be in contradiction to the laws of political economy. We have seen how modern industry always tends to the substitution of the simpler and more subordinate employments for the higher and more complex ones. How, then, could a mass of workers thrown out of one branch of industry by machinery find refuge in another branch, unless they were to be paid more poorly? and
>>To sum up: the more productive capital grows, the more it extends the division of labour and the application of machinery; the more the division of labour and the application of machinery extend, the more does competition extend among the workers, the more do their wages shrink together.
This was proven wrong in his own lifetime as factory worker wages rapidly grew in industrializing Britain.
On the other hand, if it's really just about a power struggle, why not use Altman's dealings with Saudi Arabia as the fake reason? Why come up with some weak HR excuses?
So they're essentially hoping that nobody will sue them but if they are sued that their own words can't be used as evidence against them. That's why lawyers usually tell you to shut up, because even if the court of public opinion needs to be pacified somehow the price of that may well be that you end up losing in that other court, and that's the one that matters.
As a lawyer, I wonder to what extent lawyers were actually consulted and involved with the firing.
From what I've seen of code developed in the 80s, however, asm code was not written to be divided into general and os specific parts. Writing cross-platform code is a skill that gets learned over time, usually the hard way.
He is dehumanizing programmers that can stop their sole reign on the AI throne, by labeling them as Nazis. Especially FOSS AI which by definition can't be "aligned" to his interests.
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.
:-)
Elon isn't above reproach either but I share interest with him (aka Robert Heinlein) which informs me on his decision making process.
Keep in mind this predates basically ANY kind of source control. It would have been nearly 3x the work.
It might be before they were ported to DOS or OS/2, but it definitely wasn't before source control existed (SCCS and RCS were both definitely earlier.)
If architectured properly (big if) you can split up the project appropriately so there is a common core and individual parts for specific OS.
Is it extra effort? Sure. Impossible? Definitely not.
(Nobody else wanted the job, but I thought it was fun.)
SCCS was created in 1973. We're talking about over a decade later.
Also primitive forking, diffing and merging could be (painfully) done even with crude tools, which did exist.
>separating power into the state and economy into the market gets good results.
How do you think this would be done? How do you remove power from money? Money is literally the ability to convert numbers into labor, land, food,
Precisely, they were not a capitalist society, where capital (and not simply "money" as you said) is source of power, like in capitalist societies.
In, for example, the Netherlands the richest people pay less tax [0]. Do you think this is not the case in many other countries?
> They are also, for [..] they employ, paying tax
Is that a benefit of having rich people? If companies were employee-owned that tax would still be paid.
[0]: https://www.iamexpat.nl/expat-info/dutch-expat-news/wealthie...
E.g. [0]
> In, for example, the Netherlands the richest people pay less tax [0]. Do you think this is not the case in many other countries?
That's a non sequitur from the previous point. However, on the "who pays taxes?" point, that article is careful to only talk about income tax in absolute terms, and indirect taxes in relative terms. It doesn't appear to be trying to make an objective analysis.
> Is that a benefit of having rich people?
I don't share the assumption that people should only exist if they're a benefit.
> If companies were employee-owned that tax would still be paid.
Some companies are employee-owned, but you have to think how that works for every type of business. Assuming that it's easy to make a business, and the hard bit is the ownership structure is a mistake.
[0] https://www.thinkadvisor.com/2016/08/01/why-so-many-wealthy-...
Well it's not a matter of the people existing, it's whether they are rich or not. They can exist without the money.
Anyway, if you don't think it matters if they are of benefit, then why did you bring up the fact that they pay taxes?
> I think this is a non sequitur.
I mean after someone can afford all the needs, wants, and luxuries of life, the utility of any money they spend is primarily power.
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.
As they learn to construct better and more coherent conceptual chains, something interesting must be happening internally.
I think trying to model the world based on a single projection won't get you very far.
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.
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)?
I personally think a far more fundamental change is necessary to reach AGI.
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.
* 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)
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.
fashion allows people to optimize their appearance so as to get more positive attention from others. Or, put more crudely, it helps people look good so they can get laid.
Not sure that it's net positive for society as a whole, but individual humans certainly benefit from the fashion industry. Ask anyone who has ever received a compliment on their outfit.
This is true for rich people as well as not so rich people - having spent some time working as a salesman at H&M, I can tell you that lower income members of society (like, for example, H&M employees making minimum wage) are very happy to spend a fair percentage of their income on clothing.
It is a powerful medium of self-expression and social identity yes, deeply rooted in human history where costumes and attire have always signified cultural, social, and economic status.
Drawing from tribal psychology, it fulfills an innate human desire for belonging and individuality, enabling people to communicate their affiliation, status, and personal values through their choice of clothing.
It has always been and will always be part of humanity, even if its industrialization in Capitalistic societies like ours have hidden this fact.
OP's POV is just a bit narrow, that's all.
You're talking about an industry that generates approximately $1.5 trillion globally, employing more than 60 million people globally, from multi-disciplinary skills in fashion design, illustration, web development, e-commerce, AI, digital marketing.
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"
Could you at least elaborate what they are “not”? Surelly you are not having a problem with “LLMs predict language”?
There is nothing special about human intelligence threshold.
It can be surpassed by many different models.
I meant people with a certain amount of money. I don't think we should be assessing pros or cons of economic systems based on whether people get to keep their money.
> Anyway, if you don't think it matters if they are of benefit
I don't know what this means.
> then why did you bring up the fact that they pay taxes?
I bring it up because saying they pay less in income taxes doesn't matter if they're spending money on stuff that employs people (which creates lots of tax) and gets VAT added to it. Everything is constantly taxed, at many levels, all the time. Pretending we don't live in a society where not much tax is paid seems ludicrous. Lots of tax is paid. If it's paid as VAT instead of income tax - who cares?
>I don't think we should be assessing pros or cons of economic systems based on whether people get to keep their money.
but earlier you said:
>They are also, for everything they buy, and everyone they employ, paying tax.
So if we should not assess the economic system based on whether people keep their money, i.e. pay tax, then why mention that they pay tax? It doesn't seem relevant.
Not just pay tax. People lose money over generations for all sorts of reasons.
I brought up tax in the context of "redistribution", as there's a growing worldview that says tax is not as a thing to pay for central services, but more just to take money from people who have more of it than they do.
You don't need money for those things.
Money (in a market) can buy you things, but only things people are willing to sell. You don't exert power; you exchange value.
The government itself uses money to do those things. Police don't work for free, prisons aren't built for free, guns aren't free. The government can be thought of as having unfathomable amounts of money. The assets of a country includes the entire country (less anyone with enough money to defend it).
If a sword is kinetic energy, money is potential energy. It is a battery that only needs to be connected to the right place to be devastating. And money can buy you someone who knows the right place.
Governments have power because they have resources (money) not the other way around.
Regulatory capture is using the state's power. The state is the one with the power. Rent seeking is the same. Hiring goons is illegal. If you're willing to include illegal things then all bets are off. But from your list of non-illegal things, 100% of them are the state using its power to wrong ends.
> The government itself uses money to do those things. Police don't work for free, prisons aren't built for free, guns aren't free.
Yes, but the point about power is the state has the right to lock you up. How it pays the guards is immaterial; they could be paid with potatoes and it'd still have the right. They could just be paid in "we won't lock you up if you lock them up". However, if Bill Gates wants to publicly set up a prison in the USA and lock people in it, he will go to jail. His money doesn't buy that power.
So, no. The state doesn't have power because it has enough money to pay for a prison and someone to throw you in it. People with money can't do what the state does.
The state only has the right to do something because it says it does. It can only say it does because it can enforce it in it's terrority. It can only enforce in its territory because it has people who will do said enforcement (or robots hypothetically). The people will only enforce because the government sacrifices some of its resources to them (or sacrifices resources to build bots). Even slaves need food, and people treated well enough to control them. Power doesn't exist with resources, the very measure of a state is the amount of resources it controls.
Money is for resources.
I am not arguing that anyone currently has the resources of a nation-state, it's hard to do when a state can pool a few thousand square miles of peoples money to it. I am arguing it money that makes a state powerful.
- It's been used unethically for psychological and medical purposes (with insufficient testing and insufficient consent, and possible psychological and physical harms).
- It has been used to distort educational attainment and undermine the current basis of some credentials as a result.
- It has been used to create synthetic content that has been released unmarked into the internet distorting and biasing future models trained on that content.
- It has been used to support criminal activity (scams).
- It has been used to create propaganda & fake news.
- It has devalued and replaced the work of people who relied on that work for their incomes.
I'm going to go ahead and call this a positive. If the means for measuring ability in some fields is beaten by a stochastic parrot then these fields need to adapt their methods so that testing measures understanding in a variety of ways.
I'm only slightly bitter because I was always rubbish at long form essays. Thankfully in CS these were mostly an afterthought.
Many people (myself included) would argue that is true for almost all technological progress and adds more value to society as a whole than it takes away.
Obviously the comparisons are not exact, and have been made many times already, but you can just pick one of countless examples that devalued certain workers wages but made so many more people better off.
- because it's happened before doesn't make it ok (especially for the folks who it happens to)
- many more people may be better off, and it may be a social good eventually, but this is not for sure
- there is no mechanism for any redistribution or support for the people suddenly and unexpectedly displaced.
these are behaviours and traits of the user, not the tool.
Neither thing is evil, or good, but the choice of what is used and what is available to use for a particular task has moral significance.
I know a law firm that tried ChatGPT to write a legal letter, and they were shocked that it use the same structure that they were told to use in law school (little surprise here, actually).
https://arstechnica.com/tech-policy/2023/05/lawyer-cited-6-f...
It was total nonsense anyway, and the path to dismissal was obvious and straightforward, starting with jurisdiction, so I'm not sure how effective it would be in a "real" situation. I definitely see it being great for boilerplate or templating though.
Depends on what you define as positive impact. Helping programmers write boiler plate code faster? Summarize a document for lazy fuckers who can't get themselves to read two page? Ok, not sure if this is what I would consider "positive impact".
For a list of negative impacts, see the sister comments. I'd also like to add that the energy usage of LLMs like ChatGPT is immensely high, and this in a time where we need to cut carbon emissions. And mostly used for shits and gigles by some boomers.
Of course saving time for 100 million people is positive.
I feel like the invention of calculators probably came with the same worries about how kids would ever learn to count.