Satya Nadella: "We have all the rights and all the capability."
stackdiary.com
stackdiary.com
So, uh, does anyone have any ideas on what the situation actually is? If OpenAI ceased to exist, can Microsoft take the code, model weights, everything of ChatGPT 4, and just continue developing it and produce a new MS-ChatGPT based on it? That seems to be too good to be true. Why would OpenAI give away all of it's IP to Microsoft for that cheap, with the only limitation being MS can only make a 100X return + 20% a year, and MS does not own any IP which OpenAI deems to be at the level of AGI. A deal like that is surely worth more than 10B.
Musk is a poster child on this front.
Lying about a material fact in a way that would be trivial to prosecute you on, especially when you have a well funded adversary it impacts, and that would be more than happy to destroy you with such a lie, is a good way to lose riches.
This is more like Ukraine announcing Putin’s pending invasion in advance.
Weird to bring Ukraine and Russia into this...
Ukraine and Russia seem quite applicable, since Mr. Nadella was basically making it clear that:
1) he knew the potential direction someone might take to harm the company he is in charge of, and
2) there were already plans/things in place to make their chances of success zero. Plans that the opponent knows about and can verify, and in Microsoft’s case they were a party to making.
It’s the ‘I see what you’re doing there’ notice.
You’d have to be pretty desperate to keep going after that kind of notice. Russia was, and it has cost them immensely. We’ll see how it plays out OpenAI wise.
If those things weren’t all provably true, then it would be a really dumb move to do as it just exposes him to tons of direct risk with no plausible gain, as the opponent would have zero compunction in hurting him with it - and tons of ability to do so.
If you don't want Microsoft to take everything, "stop hitting yourself."
So you are saying that, on “the data”, Microsoft would not be in a materially worse business position given the OpenAI chaos if the security of their rights to OpenAI models in the event of a breakdown in their existing relationship was perceived as insecure?
Because otherwise the chain of reasoning is:
The licensing is secure, because Nadella says so, and Nadella can be trusted because if he was lying it would be securities fraud, and Nadella wouldn't commit securities fraud even if it would benefit Microsoft’s and, therefore, his interests.
That alone translates into Microsoft doesn't need to commit fraud, all they need to do is sit back and watch the rest implode and come out winners. They don't even have to sue, all they need to do is do nothing and OpenAI is toast.
That press release from MS the other day is a bowshot, not a promise. 'Play ball or else' with the 'or else' bit conveniently left to the imagination of the recipient.
> That alone translates into Microsoft doesn't need to commit fraud
My prior that “the only possible benefit to securities fraud is in circumstances of imminent existential risk” is clearly weaker than yours. Also, if Microsoft's description of the licensing situation is misleading, the existential risk to OpenAI—which makes it more likely that someone else ends up with control of its assets—makes the risk to Microsoft greater, not less.
Change of control is a very standard clause in Escrow arrangements and it may have been subtly worded specifically to ensure that control of the company remained friendly to Microsoft interests given that they did not have a board seat. That's pretty much the only way they should have accepted that construct. If they didn't they were dumb and you can say a lot about Microsoft (I know I have and they're a shit company in my opinion) but they know the legal world like few other tech companies.
Satya does not sound like someone communicating from a position of strength - he's been repeatedly assuring his shareholders that no matter what happens, their AI plans moving forward are secure. They need stability, not chaos, because whatever clear lead it seems they have right now could be quickly be eclipsed in the coming months.
Committing securities fraud and accepting a fine would pale in comparison to the potential value erosion.
Unlikely doesn't mean it doesn't happen and the more high profile a case the more likely the SEC will do something about it.
https://www.nbcnews.com/business/markets/10-ceos-who-went-bo...
I don't think the average citizen will care so much. When it comes to white collar crime making some white lies to bolster shareholder confidence is not in the same orbit as the type of massive fraud for notable jail time cases over the last couple decades.
At worst it should be a moderate SEC fine, perhaps a shareholder lawsuit if things don't turn out favorable. And this is assuming a lot that has to go wrong.
The links to politicians, lobby efforts, the US economy, etc, etc... means something really egregious has to happen for jail time to be an issue w/ high level white collar crime.
EDIT: Since I can't respond further to the chain - tons of people cared about Martha Stewart. They were sick of people in high society getting a slap on the wrist for things like this. It definitely was a cathartis... esp. from someone so unlikely.
I'm trying very hard not to let my personal feelings about Microsoft cloud my judgment and if anything happened to Microsoft I wouldn't shed a tear (and might even open a bottle of bitter lemon). I don't see them committing an own goal like that.
But, I don't think we're going to agree on this.
MSFT actually is fighting for its life. GPT4 is an intriguingly useful curiosity, Later GPTs had the potential to be culturally, politically, financially and technically revolutionary.
That likely won't happen now. We'll get some kind of enshittified prohibitively expensive mediocrity instead.
Ok. That's funny. But if this turns out to be the piece of the apple that MS ends up choking on then maybe it was all worth it?
Could someone suggest Adam D'Angelo for the boards of Facebook, Google and Twitter please. Oh, wait...
Of the slate of $1-3T tech companies the question of "where can we go from here?"... this is where, and the winners will be the first $10T enterprises.
They come out smelling like roses already and it can only get better, not worse for them going forward.
I read this whole situation as putting pressure on OpenAI to come to their senses and quick.
That we are in violent agreement on, the board should negotiate to be allowed to flee indemnified for the fall-out of their action. But if they do not OpenAI is done for, Anthropic has already indicated they are not going to be there to pick up the pieces, and Google and Meta will try to poach as many employees as they can by being 'not Microsoft' (which has a fair chance of working).
So there is damage, but it isn't over yet and for the moment Microsoft looks as though they hold the most and the most valuable cards already. After all they already have 49%. Another 2% as a suing-for-peace pacifier and they're as good as kings.
He could have worded it in a far safer way and had the same benefit if he wanted to as well.
IMO, odds are low he’s lying about this. It would likely look much different if he was trying to.
To be frank, it's entirely possible that Sam could have been fired for agreeing to a contract like this without board approval because the existential risk OpenAI constantly faces dramatically increases the likelihood that OpenAI investors -- besides Microsoft -- would never be made whole in the case of an implosion. The "Adam D'Angelo spite" hypothesis is possible, too, but it seems far less likely that the board of such a valuable company -- no matter how inexperienced and feckless they might be -- would risk tanking the company on the personal vendetta of a single director.
> This is, quite obviously, a phenomenal outcome for Microsoft. The company already has a perpetual license to all OpenAI IP (short of artificial general intelligence), including source code and model weights; the question was whether it would have the talent to exploit that IP if OpenAI suffered the sort of talent drain that was threatened upon Altman and Brockman’s removal. Indeed they will, as a good portion of that talent seems likely to flow to Microsoft; you can make the case that Microsoft just acquired OpenAI for $0 and zero risk of an antitrust lawsuit.
> (Microsoft’s original agreement with OpenAI also barred Microsoft from pursuing AGI based on OpenAI tech on its own; my understanding is that this clause was removed in the most recent agreement)
> Microsoft’s gain, meanwhile, is OpenAI’s loss, which is dependent on the Redmond-based company for both money and compute: the work its employees will do on AI will either be Microsoft’s by virtue of that perpetual license, or Microsoft’s directly because said employees joined Altman’s team. OpenAI’s trump card is ChatGPT, which is well on its way to achieving the holy grail of tech — an at-scale consumer platform — but if the reporting this weekend is to be believed, OpenAI’s board may have already had second thoughts about the incentives ChapGPT placed on the company (more on this below).
Discussed at https://news.ycombinator.com/item?id=38346869
https://www.wsj.com/articles/microsoft-and-openai-forge-awkw...
Good luck with that line of argument, I wouldn't want my reputation and my fortune to hinge on that.
> Sam & Greg saw fit to leave him in place.
Bad mistake on their part given the guy's history.
> Nothing from Emmett yet.
I wonder if he'll resign...
https://www.bloomberg.com/news/articles/2023-11-21/altman-op...
As for whether or not investors have a case: Microsoft may not, but the smaller investors probably do because they can fairly easily prove that their interests were harmed by the ill advised actions of the board.
https://www.bloomberg.com/opinion/articles/2023-11-21/openai...
And until that proof surfaces I'm going on the assumption that it doesn't exist because if it does exist they would have certainly used it to bolster their case.
Note Ilya's statement of regret and that none of the board members have said a word and that the current CEO is nowhere to be seen. That does not look like a group of people that are acting boldly in the defense of their corporate charter to me.
It looks like a bunch of weasels that thought they could get away with murdering the chicken while they were unattended for five minutes and they're now all covered with blood and pretending to be innocent.
Leaks say whatever they want to say and if I found myself in this position - highly unlikely - and I would have leaked something it would have definitely been something to make me look good. Or did you think that if it made them look bad that it would have been leaked as well? In that case I can see your point.
Note this key quote by Nadella:
"“As a partner, I think it does deserve you to be consulted on the big decisions.”"
That's one way to look at what might be grounds for a suit. He could be wrong, but that won't stop them from suing to establish that fact and they have very little to lose.
Hint: the 51% are more likely to go after the board than that they are going to go after Microsoft because the board action is what precipitated this whole mess.
Microsoft is just - conveniently - picking up the pieces and periodically drops hints about what they'll do if anybody gets in their way of doing that.
Well, if OpenAI magically got erased from existence, probably.
More realistic scenarios, like if they continued to exist, but there was a legal dispute about breach by Microsoft of the agreements by which Microsoft got access to the IP, may be more complicated.
What happens when a “gutted" firm has assets whose value cannot be realized except by an actor with greater resources and time window, but those actors do, in fact, exist?
Somewhat disagree on this, since open sourcing their model and code would help a lot of people in a lot of different ways.
I would not rule that out.
>Microsoft take the code, model weights, everything of ChatGPT 4, and just continue developing it and produce a new MS-ChatGPT based on it?
They don't need Sam to do that, or his employees. With the current state of LLMs they can hire a few dozen people with some experience and that wouldn't be the hard part.
In my mind the aggressive poaching OpenAI has been doing with Meta and Google is because they have nothing special. If they did they wouldn't need to grow to 800 headcount and still hiring. They're anticipating some company will have the next breakthrough, and they wanted every important brain on their side
Having the weights to another transformer based LLM is great, but its not going to change humanity. The only thing I saw this weekend was a bunch of entrepreneurs who's new startup is 100% dependent on the GPT-4 API freaking out. Thats not ground breaking
If OpenAI leadership views it as a part of their charter to release their weights, research, etc. as open source, ten other companies would immediately catch up to Microsoft.
Isn't it the completely the opposite of that, though?
No, its more complicated, which is why they were initially actually open, then became closed as their assessment of the circumstances changed; their own inability to act as gatekeeper, which would go away with a split with Microsoft where Microsoft had unrestricted rights and was not in practice coordinating with them, could well sufficiently change the circumstances to change their stance, without any inconsistency in reasoning (not that reasoning itself can legitimately evolve, too.)
In simple terms, it plausible that they might see the broad preference as:
good gatekeeper of top of the line AI > open models > bad gatekeeper of top of the line AI
and also see Microsoft, after the current kerfuffle, as a bad gatekeeper.
Google et al have gambled the answer for most API consumers is no
This is false. Look at the chatbot arena leaderboard to see how GPT4-Turbo crushes the competition[1].
[1]: https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboar...
And between Vicuna and GPT4-Turbo There are the Claude Models and WizardML. This is as even playing field as i've ever seen for the tech world.
When a $100B company gets to 2000, and the rest stay behind, thats when we know the tech has changed. Thats what OpenAI was hiring for. Not for anything they already gave Microsoft access to
If Microsoft wanted to get to 1100, they could do that with 50 smart hires from across the industry and universities. They don't need Sams teams $1M salaries for that
On top of that, they have done an amazing job executing on getting product out to users. No other company in this space is even remotely close to this.
It's business-speak rather than technical/legal, but I don't know how else someone could reasonably interpret that.
The bigger question to me is if what Satya is saying is true, why even continue the charade with OpenAI’s board? Then what Satya is saying is probably just to protect the stock price.
Well Microsoft could have contracted in some sort of exclusivity. Barring that, I think the obvious answer is because the leftovers aren’t that valuable.
Microsoft got 49%, and full access to all models, and most profits. The remaining profits and access to the same models as your competitors isn’t worth it.
Most big companies are making their LLMs, and I’ll postulate that while the tech is amazing, it’s not going to be the future where you need the One True Model. Google seems to be doing well enough with their models, Meta released pretty good models for free, Salesforce and Amazon are building out models, even smaller companies like Databricks got into the hype. Then with slowly opening access to models from Anthropic and other startups, GPT5+ alone isn’t worth it. Oracle maybe could buy access, but are they willing to outbid Microsoft?
The only thing that could be worth it is the ability to run/own ChatGPT the consumer product and brand. And who would own that? Google has their own, Apple wouldn’t want it, Meta probably can’t buy it without government intervention, Salesforce maybe, but they’re not consumer focused. Amazon maybe? Oracle maybe, but same problem as Salesforce?
Building an LLM is one thing, operating one is another. Without a massive discount on actual operations OpenAI didn't have much of a business. They were likely getting their Azure stuff at cost, the same as an internal Microsoft org would be charged for Azure, which is cheaper than building their own infrastructure or paying retail cloud rates.
Microsoft wouldn't give them such a sweet deal without guarantees on their part. They're not going to put up billions and base products on a company that could cease to exist at any minute for a bunch of reasons.
"just" push the state of the art. unlikely, even if they have access to everything. they need the people and even then it's unlikely.
Side rant: just when the crypto madness settled down and I hoped to score a cheap graphics card, the AI craze starts up. Oh well.
A deal like that - if true - would also imply that they should have abruptly fired Sam years ago, before it got signed off
(Of course Satya's comments can also be interpreted as rather more loosely implying they're perfectly capable of maintaining what they've licensed and hiring the expertise to build their own GPT5)
If Sam was responsible for that, he smuggled one hell of a trojan horse into OpenAI, and should have been fired a long time ago.
It seems like OpenAI needed deep pockets and access to computing power to survive. That limits the potential companies they could have cut deals with.
The compute resources that are required are massive, and if not MS then you're getting into bed with some other Big Corp, i.e. Amazon, Oracle, IBM, Google, etc.
What's the likelihood that he signed onto a $10 billion investment deal from Microsoft without approval from the board?
Think about it when the deal was on the table:
1) Take the money. If OpenAI fail, MS get the code. OK, who cares, OpenAI have gone bust.
2) Don't take the money, don't become successful unless they can find another partner with such a good fit (deep pockets and resources).
It made sense to allow MS to have the code if OpenAI failed since they'd have had nothing to lose apart from whatever residual value they could sell their IP for (which they already were effectively doing for $10B, so fair enough). If this was a sticking point to secure a $10B investment, it made sense. OTOH if they granted the IP to MS not contingent on OpenAI failing, then that would have been sheer incompetence.
Clearly the possibility of OpenAI failing is not independent from the behaviour of its corporate partner. And I'm not sure Microsoft's negotiation is quite bad faith enough or their interest in OpenAI low enough to have made the entire deal contingent on the "if/when we decide to move aggressively against your company so it collapses we get exclusive free in perpetuity access to everything you ever did and no more caps to the profits from them" clause. If it's in there, it's in there because they signed a bad deal rather than because it was the only possibility.
I think Microsoft could easily spin this as just protecting their interests.
I guess if you believe OpenAI's original stated mission was ever more than a marketing pitch, there's a good argument that it's all been bad anyway (building and licensing extremely impressive tech for their corporate partners and a growing consumer base isn't obviously contributing to making the world a "safer from unfriendly AI" place)
That's based on a long time in business, I can't recall the last time I saw someone blow up an 80 billion dollar company based on ethics alone.
That, plus accounts of him being deceptive/manipulative like this (e.g. https://twitter.com/geoffreyirving/status/172675427761849141...), plus being fired precisely for deception makes it pretty consistent.
The more important point here is Satya is saying "Shareholders don't worry" and "Board of OpenAI, don't fuck about you need us more than we need you"
Sue who? And for what?
Are you proposing that they sue the board for firing the CEO? What would be the theory?
The only defense the board would have is that they absolutely needed to fire the CEO to protect their precious non-profit charter. But for that to be the case they would need to show their homework. I don't see that happening and so they may well be exposed. At least two of the remaining board members are not positioned at all to take on Microsoft in a fight like that, they'll fold like chopsticks. Microsoft can keep this up for a decade if they want and not lose money on it.
Seems like the only winners standing in this scorched earth is Satya and his shareholders.
But doesn't really explain why, I wonder if there's any substance to it in the light of Satya's comments about having the rights to "continue the innovation".
https://www.bloomberg.com/news/newsletters/2023-11-21/micros...
If Microsoft does end up poaching the entire team, everyone will be walking on eggshells with legal breathing down their neck.
For every hour that the new CEO doesn't come out to start repairing their reputation that chance goes up.
I had no idea the cataclysm my inquiry would launch!
I suppose on the other hand I'm a little spoiled by all of the free (as in beer) tools available today. There was a time I paid well over $100/month for MSDN to get equivalent stuff from MS. (I see it's just under $100/month today so I may be misremembering the pricing.)
And yet a “Person of Significant Control” in UK corps would need only a 25% share (or voting rights), OR be deemed to have “significant influence”.
Why are competition laws so backwards :(
This statement could be interpreted as a thinly-veiled threat. When viewed in conjunction with Satya's efforts to hire 700+ OpenAI team members in bulk, any attempt at intimidation may constitute an illegal practice known as a workforce raid.
Described here https://twitter.com/8teAPi/status/1726981483032977791
https://blogs.microsoft.com/blog/2023/11/19/a-statement-from...
It's not quite like waking up next to a horse's head but it says 'Sam gets the part' regardless of what you clowns will cook up next and don't stand in my way.
Especially big decisions that potentially annihilate your investment.
https://x.com/8teAPi/status/1726983321329250666?s=20
Marc or his counsel very precisely worded the tweet to tiptoe around any risk of it being construed as workforce raiding. Satya did not take this precaution, presumably because Satya has never heard of workforce raiding and is acting increasingly recklessly, as he makes his way through this sleep-deprived weekend fueled by a blinding desire to capture all of OpenAI and win it all.
https://x.com/1hopecomingtrue/status/1727169849900868017?s=2...
If a CEO is busy trying to manage the stock than that should raise some suspicion.
Maybe MSFT really got the deal of a lifetime where they (a) tricked the FTC to not consider anything they are doing to touch antitrust laws due to "only buying 49% of the business" and (b) have a way to "extract" the core assets of OpenAI to the detriment of the remaining shareholders. That's not impossible given the quaint structure OpenAI has and the unusual structuring of the deal with MSFT, but would be quite surprising.
Anybody have any insight at all as to how this ends?
* Sam goes to MS * Sam stays at OpenAI * Sam starts his own thing
His own thing feels super unlikely because he’s said he’d go to MS.
Staying at OpenAI feels super likely since the staff revolted. 50%? 75%?
Going to MS seems likely but maybe slightly less so. 25%? 50%?
Bookmark this comment because I am a genius and time will prove it.
(All said in jest)
1. Microsoft is sold OpenAI
2. OpenAI makes some sort of deal to return to normal state and eventually heads roll on the board (as in not part of the deal but people are bought out)
3. (My favorite) OpenAI destroys Microsoft/Altman power grab and releases enough tech to the public
Lawsuits.
Interminable lawsuits, making boring headlines into the 2040s.