I like your take. Altman is a Valley hustler, albeit a talented one who ran YC. But, he’s not technologist or a theorist, and I suspect he’s not that interested in attempting to de-risk AGI anymore, particularly now that he’s experienced the market hype. The staff who want to be billionaires clotted to him; the staff who are committed to the original ethical vision stuck with Ilya.
Good points all around.
There was a headline a day or two ago suggesting Microsoft is ready to put $50 billion into AI. If they are truly as all-in on AI as Meta is/was with VR, then something like this would make sense. OpenAI is of course at the front of things, but Microsoft still owns less than half and still has to answer to other investors.
But while OpenAI has valuable tech, the real value is the researchers who are building the next golden goose. Microsoft seems to understand that.
Look at that open letter. 500 out of 700 employees signed it and are threatening to leave. I think the key there is their equity compensation- many if not most of those 500 people were getting equity, and if OpenAI loses its place at the front of the pack that equity loses a lot of value. And the board just essentially killed their golden goose. So those 500 people are worried about the future of their stock. If it looks like the company is going down, the rats will dessert the sinking ship faster than you can snap your fingers as there will be no shortage of very lucrative offers.
I bet Microsoft is offering a very attractive life preserver for those who jump ship. Wouldn’t be surprised if, if a bunch of those people defect, the AI division get some sort of quasi autonomous status or perhaps even its own corporate entity for those employees to own stock in.
Point being- there’s a very good chance Microsoft will essentially buy OpenAI for mere pennies by simply hiring the majority of their employees.
However, as the situation unfolds… from the amount of effort being put in trying to get a hold of openAI by Altman/Microsoft…
It seems to me that there’s a lot of claimed real-estate (eg training data and weights, ongoing subscriptions, income streams, exclusive agreements etc) that would be nontrivial or impossible to be gotten hold of or recreated from scratch.
And as we know Altman is not involved in the nittty gritty, aka doesn’t know shit about shit in recreating what was there - so the initial bravado of “I’m just going to do my own” - was clearly a bluff that didn’t pay off.
Hence now is trying different ways to get back on the ship / get another hold of the golden goose.
To be honest the more I see his frantic manoeuvring the more I think he’s inconsequential to the endeavour and he just happened to be lucky - and these jerky moves, to me, demonstrate that he’s well aware of it.
Altman may not be an AI engineer. But he does appear to be an effective leader of an AI team, and he obviously has the loyalty of an awful lot of his employees. That could be because they are personally loyal to him, or because of their own self-interest with equity, but the situation remains. He put together a solid team once, he can probably do it again. But he also may not have to.
If Microsoft can vacuum up a majority of those 500 employees, if those people join Microsoft because they want to keep working for Altman, then whatever they are paying him is money well spent. And while Altman may not know the nitty gritty, those people do. Collectively they are probably worth more than he is.
Either way, Microsoft has successfully bet on both sides of the coin. Whether they build their own successful AI department, or whether OpenAI continues to succeed, or both, they’re going to come out on top.
The big question mark to me though why exactly the board chose to fire him in the first place. It’s been a good few days, and so far there’s not even a hint. That is strange to me. If whatever he did is awful, the board would want to justify itself. If whatever he did is benign, he would want to redeem his own image. The fact that there’s not even a leak is odd.
He does appear to have their loyalty, if reports are to be believed.
However even every single one of the OpenAI employees on board with him, they’re not taking existing IP with them. And to recreate it from scratch, it seems, would put them so far behind as to not be worth it. Otherwise why freak out about the workplace?
Re Microsoft, they do seem to have the least amount of risk, where at minimum they’ll keep what they have, while the max upside is 100% of everything.
Although even for that, there are some uncertainties that only people with inside knowledge might have. E.g. if MS were instrumental in Altman’s subterfuge that caused his firing, if the board doesn’t want MS involved if they can force them to divest their 49%. Doesn’t seem likely.
Or if the board can create a different for-profit where MS doesn’t have a part and let the other one wither.
Or the risk of massive lawsuits from OpenAI against MS if they wholesale take their employees re IP theft (we’ve seen it between apple, google and Amazon).
I like your take. Altman is a Valley hustler, albeit a talented one who ran YC. But, he’s not technologist or a theorist, and I suspect he’s not that interested in attempting to de-risk AGI anymore, particularly now that he’s experienced the market hype. The staff who want to be billionaires clotted to him; the staff who are committed to the original ethical vision stuck with Ilya.
My little theory is that he was in cahoots with Microsoft re some product roadmap, against the wishes of the board.
Microsoft was not consulted re his firing as not to alert him or because there’s no point asking the revealed adversary.
Microsoft was “furious” and desperately tried to reinstate him.
Immediately offered him a job. Effectively undermining their investment in open ai.
Good points all around. There was a headline a day or two ago suggesting Microsoft is ready to put $50 billion into AI. If they are truly as all-in on AI as Meta is/was with VR, then something like this would make sense. OpenAI is of course at the front of things, but Microsoft still owns less than half and still has to answer to other investors.
But while OpenAI has valuable tech, the real value is the researchers who are building the next golden goose. Microsoft seems to understand that.
Look at that open letter. 500 out of 700 employees signed it and are threatening to leave. I think the key there is their equity compensation- many if not most of those 500 people were getting equity, and if OpenAI loses its place at the front of the pack that equity loses a lot of value. And the board just essentially killed their golden goose. So those 500 people are worried about the future of their stock. If it looks like the company is going down, the rats will dessert the sinking ship faster than you can snap your fingers as there will be no shortage of very lucrative offers.
I bet Microsoft is offering a very attractive life preserver for those who jump ship. Wouldn’t be surprised if, if a bunch of those people defect, the AI division get some sort of quasi autonomous status or perhaps even its own corporate entity for those employees to own stock in.
Point being- there’s a very good chance Microsoft will essentially buy OpenAI for mere pennies by simply hiring the majority of their employees.
Good points for you also.
However, as the situation unfolds… from the amount of effort being put in trying to get a hold of openAI by Altman/Microsoft…
It seems to me that there’s a lot of claimed real-estate (eg training data and weights, ongoing subscriptions, income streams, exclusive agreements etc) that would be nontrivial or impossible to be gotten hold of or recreated from scratch.
And as we know Altman is not involved in the nittty gritty, aka doesn’t know shit about shit in recreating what was there - so the initial bravado of “I’m just going to do my own” - was clearly a bluff that didn’t pay off.
Hence now is trying different ways to get back on the ship / get another hold of the golden goose.
To be honest the more I see his frantic manoeuvring the more I think he’s inconsequential to the endeavour and he just happened to be lucky - and these jerky moves, to me, demonstrate that he’s well aware of it.
Altman may not be an AI engineer. But he does appear to be an effective leader of an AI team, and he obviously has the loyalty of an awful lot of his employees. That could be because they are personally loyal to him, or because of their own self-interest with equity, but the situation remains. He put together a solid team once, he can probably do it again. But he also may not have to.
If Microsoft can vacuum up a majority of those 500 employees, if those people join Microsoft because they want to keep working for Altman, then whatever they are paying him is money well spent. And while Altman may not know the nitty gritty, those people do. Collectively they are probably worth more than he is.
Either way, Microsoft has successfully bet on both sides of the coin. Whether they build their own successful AI department, or whether OpenAI continues to succeed, or both, they’re going to come out on top.
The big question mark to me though why exactly the board chose to fire him in the first place. It’s been a good few days, and so far there’s not even a hint. That is strange to me. If whatever he did is awful, the board would want to justify itself. If whatever he did is benign, he would want to redeem his own image. The fact that there’s not even a leak is odd.
He does appear to have their loyalty, if reports are to be believed.
However even every single one of the OpenAI employees on board with him, they’re not taking existing IP with them. And to recreate it from scratch, it seems, would put them so far behind as to not be worth it. Otherwise why freak out about the workplace?
Re Microsoft, they do seem to have the least amount of risk, where at minimum they’ll keep what they have, while the max upside is 100% of everything.
Although even for that, there are some uncertainties that only people with inside knowledge might have. E.g. if MS were instrumental in Altman’s subterfuge that caused his firing, if the board doesn’t want MS involved if they can force them to divest their 49%. Doesn’t seem likely.
Or if the board can create a different for-profit where MS doesn’t have a part and let the other one wither.
Or the risk of massive lawsuits from OpenAI against MS if they wholesale take their employees re IP theft (we’ve seen it between apple, google and Amazon).