OpenAI Boardroom Battle: Safety First
fabricatedknowledge.com
fabricatedknowledge.com
> Microsoft, in particular, has tied its wagon closest to OpenAI. It’s a later investor, so it cannot make the ~100x return, but it invested $10 billion dollars in OpenAI, mostly in the form of cloud computing credits. The deal’s terms are not quite disclosed, but we all understand that Microsoft has the largest economic stake in OpenAI.
This explains the deal between Microsoft and OpenAI. They weren't able to find investors who were satisfied with a lower return, but Microsoft wanted something other then a financial return: Access to the models. So the restriction on the return ensured a much worse deal for them as the only potential investors were big tech companies to whom access to the models is useful.
This entire non-profit/capped-profit structure turned out to be too clever by half. Had OpenAI been a standard business, they wouldn't have had this problem.
> The current structure of a non-profit board without much else in place should not be running a critically important company like OpenAI.
This seems insane, given the alternative of running it like your typical YC startup.
> I believe that the longer-term problem of safe AI is important, but you don’t do that with sudden shakeups of the founder and an exodus of half of the employees. OpenAI has been doing something special for a long time, beating the likes of better-funded research organizations like Google or Microsoft. It’s probably in everyone’s best interest to keep the team together.
I'm so curious about all these takes about everybody leaving if Altman is out. And it's weird to support it by talking about the long-time specialness that existed before Altman was so day-to-day involved, and that has been under tension for a long time too (see all the departures that lead to Anthropic existing). They didn't oust him out of nowhere. The schism is there and either side winning doesn't magically make for unity.
I have no idea which side I'd root for, but these takes that the board massively screwed up all look very one-sided, ignoring the disagreements over the massive pivot openai has undergone in recent years. I'd bet we see lots of people on the "losing" side leave openai over the next months regardless of the outcome.
And no, I also don't believe that Sam A. has my best interests at heart, but at least he doesn't hide behind "safety" to pretend otherwise.
What's wrong with the idea that AI should be developed with the goal of being safe and for the benefit of all humanity?
As for the objection to safe being defined by a narrow group 'great men'. I'm sure Ilya would agree that it's not ideal that such an important milestone in human history being decided by such a small group of people. But what alternatives do we have right now? What if OpenAI decided to turn over their technology to the US government or the UN? Would that make you feel better?
Nothing. I just don't believe that's what will happen here. All I see is two sides trying to steer things in a direction they profit most from, with one side saying - more or less openly - that that's what they want to do, while the other side says "oh no, we don't want to do this at all. We only have your best interests at heart!" and I hate people that try to take me for a ride.
> What if OpenAI decided to turn over their technology to the US government or the UN? Would that make you feel better?
Replace it with "a UN replacement with only democratic governments" and yes, that would make me feel better immensely. Or put it out into the open, which would be the best option if they really care about humanity.
I've seen the "We are motivated by higher goals, trust us" plot play out one too many times, just with different actors. At some point I stopped assuming the best and instead opted for "assume the worst, maybe get pleasantly surprised once in a while". So far, I've been proven correct more often than not - as sad as that is.
> There actually are people out there who are motivated by higher ideals than their own self-interest.
Oh, sure there are. Just many more who aren't. And it's really hard to tell them apart, especially if both say the same things. But we live in a society which rewards the latter, so the safe option (the irony isn't lost on me) is to assume people aren't.
> I don't know if Ilya and the OpenAI board are in that category
That's kind of the problem, isn't it? If they aren't then OpenAI isn't governed by the principles in its charter in either case. And in that case: What is better, an OpenAI which is scrutinized at every step, because no one trusts Microsoft or an OpenAI which can do whatever it wants, cause "safe AI is our goal, trust us" and then wake up one day and find out they too betrayed our trust?
What could possibly make you think that Ilya and AI Notkilleveryoneism is possibly winning the PR battle? HN, Tech Twitter, VC's, and major industry are all raking them over the coals. All of those people are trying to make money and get rich. Imagine thinking you're the underdog. The scientists are trying to prevent catastrophe for everyone. This is the climate change debate all over again. You wouldn't have the cajones to think climate change was real if that view was unpopular.
That is an interesting assumption.
It might even work, if you could do it.
Actually doing that is going to be essentially impossible, though the reasons are completely different.
For example, but not limited to: there's no point having an AI you're not using, using it — even as an Oracle — involves some kind of interaction with the outside world in the form of what questions you ask and what actions you take in response.
I hate being right about this crap but OpenAI's tech will get integrated into a lot of daily tech people use and as always abuse will flourish, the 80% will suffer and the 20% will prosper.
Do you see the boards of GM and IBM slapfighting on twitter? No. It's a thing some people have taken to doing, and - in my opinion - it's not a good thing. It's incredibly immature.
This 'I'm in the office lol' behaviour is also pretty weird. It's more the sort of thing you'd expect to see on the pro-wrestling circuit than the type of behaviour you'd expect from a CEO.
Key take-aways (excerpts from article):
"The board made a blunder. OpenAI’s employees will likely get their CEO back by Monday, and Satya Nadella’s 10 billion dollars in Azure credits will have some vote in the future of OpenAI.
What’s clear is that the board is grossly mismanaged. A non-profit board should not be running a critically important company like OpenAI. Just look at the turnover and lack of transparency on re-election.
I think that Ilya [Sutskever] will leave OpenAI when Sam is reinstated. He has to be the player who initiated the power play. The entire current board will leave, and a new board with fewer AI safety people (sadly) will be reinstated.
I would not be surprised to see the OpenAI charity and capped-profit structure flipped, with a formal board at the GP that becomes the real locus of power.
The boardroom move was amateur and sudden. And as much as boards have technical legal power, so do the organizations they rule. It’s all a construct, and the people of OpenAI will get their way. And hopefully, a better governance structure.
I believe that the longer-term problem of safe AI is important, but you don’t do that with sudden shakeups of the founder and an exodus of half of the employees.
OpenAI has been doing something special for a long time, beating the likes of better-funded research organizations like Google or Microsoft. It’s probably in everyone’s best interest to keep the team together. "
A for-profit board should not be running a critically important company like OpenAI.
This is an insane thing to say. It's an argument against OpenAI existing at all, and the alternatives (for-profit boards and a military research project) are both much, much worse.
Because if they don't agree, then all their best employees, and probably also their Azure cloud credits, will leave to join Sam and Greg at a new company picking up right where they left off.
There’s no evidence that a mass resignation will happen.
Unraveling this weekend's drama is way more complicated than you suggest. Both factions actually have quite a bit of leverage.
The same goes for the Azure credits. If MS decides they want out then other big tech companies will be lining up to take their place. This makes Microsoft’s threats rather hollow.