As of this morning, the new board is in place and everything else at OpenAI is otherwise officially back to the way it was before. Events seem to have gone as expected. If you have read my previous two posts on the OpenAI situation, nothing here should surprise you.
The link is interesting. There are a lot of people with a completely unfounded belief that AI decisions, especially from an AGI, would necessarily be more objective than human ones.
Huh? How is the people's gender relevant? We know that fields with abstract systems statistically attract more males (this is a statement about interests, NOT abilities), and I don't see how specifically aiming to add a woman would be a good idea.
Mira Murati is a woman, the CTO of OpenAI, and was the first pick for interim CEO. Gender has nothing to do with this, and there are vastly more important issues at stake than irrelevant PC optics.
A power struggled ensued, and Altman won. He is a master power broker, and this was identified by people like Paul Graham years ago when he was at Y-Combinator. It would be less scary if he wasn't the head of the world's most advanced AI company (that is publicly known at least).
He is a classic tech CEO looking to create products, but this is a different kind of tech, which in my view changes all the rules and the boundaries of what is acceptable. Unfortunately, Moloch reigns supreme, and the race to develop AGI or something like it will never be tempered by safety concerns because every company thinks "If we don't do it, X and Z companies/nations will, and then we will be behind forever".
In this modern corporate world, image is as important as guile once you are public-facing enough to need public opinion on your side. Altman went to Congress and said the right things, and did a great job of acting magnanimous on Twitter after his firing, moving at light speed to get (even more) in bed with Microsoft, culminating in the complete demolition of the Board and his reinstatement in less than 3 days. I'm sure people are already writing the books about this, but if it wasn't so terrifying it would be a pretty gripping corporate drama.
I wouldn’t take Sam Altman’s comments to the effect of “clearly the company can run just as well without me” as actually being evidence that the company can run just as well without him. That’s just the sort of polite thing you say to be nice to your team.
In particular nobody else is going to be as good at fundraising as Sam Altman, and fundraising is probably still an important part of OpenAI’s future plans.
He may be valuable for all sorts of other things internally, too, I don’t know. If I had to guess I would say he probably is. There’s a big difference between great startup CEOs and mediocre startup CEOs.
This is not directly relevant but I would like to mention my continued confusion at why otherwise smart and contrarian people have any respect for graham, who has always seemed to me to be the maximally generic empty suit in tech.
I don't know what you think "generic" or "empty suit" mean, but it's very clear that Paul Graham isn't what those terms mean to most people. He's clearly someone who thinks for himself, has interesting thoughts & insights, cares about good outcomes in the world, and does not simply mouth standard meaningless corporate pr-speak.
Maybe if you are already steeped in startup & rationalist culture, it's possible you could think that Seinfeld Isn't Funny:
I don't think it's even rationalist related? Feels like all his quotes are things like "it's important to have a good head for business and understand the product" or similarly empty generic statements like that.
This is a bit late, but I'd like to offer up a few things:
a) Paul Graham played a pivotal behind-the-scenes role in shaping the modern tech economy. We are living in the world he built. The existence of Y Combinator, and the social community behind it, puts competitive pressure on VCs and has created a more founder-centric, innovation-centric culture. Things could have gone differently, and in the late 90s and early 00s, it looked like they might.
b) Paul Graham's essays are aggressively lacking in style. I'll compare to Zvi, who ended this post with a reference to "The Wire", and in his most recent, has at least one reference that I simply don't get at all (which probably means there are more that I didn't even notice). That's a valid choice on Zvi's part, but so is Paul Graham's decision to write in a way that a 10-year-old can comprehend. It's the opposite of "spicy takes on Twitter".
c) Paul Graham has successfully fired Sam Altman, which seems like a rare talent one would like to have on the board of OpenAI, just in case. ;-)
Your point about Paul Graham loving his kids is a brief comment, but I have often thought our chances for survival would go up if major AI companies were led by CxOs and Board members who had children. In lieu of that, monthly bring-your-kid-to-work-days? CxO talks at high schools and middle schools?
The concept of "future generations" moves from theory to practice when you have kids.
Thank you, Zvi. Your coverage on OpenAI recently tipped the scales for me and I became a paid subscriber.
The link is interesting. There are a lot of people with a completely unfounded belief that AI decisions, especially from an AGI, would necessarily be more objective than human ones.
Again, almost every women involved get the boot. Men men men men men.
Hard to imagine robots wouldn't do better.
Huh? How is the people's gender relevant? We know that fields with abstract systems statistically attract more males (this is a statement about interests, NOT abilities), and I don't see how specifically aiming to add a woman would be a good idea.
Mira Murati is a woman, the CTO of OpenAI, and was the first pick for interim CEO. Gender has nothing to do with this, and there are vastly more important issues at stake than irrelevant PC optics.
The major question is - will the changes lead to it being more or less likely humanity survives?
A power struggled ensued, and Altman won. He is a master power broker, and this was identified by people like Paul Graham years ago when he was at Y-Combinator. It would be less scary if he wasn't the head of the world's most advanced AI company (that is publicly known at least).
He is a classic tech CEO looking to create products, but this is a different kind of tech, which in my view changes all the rules and the boundaries of what is acceptable. Unfortunately, Moloch reigns supreme, and the race to develop AGI or something like it will never be tempered by safety concerns because every company thinks "If we don't do it, X and Z companies/nations will, and then we will be behind forever".
You have to give him credit for keeping it classy though!
In this modern corporate world, image is as important as guile once you are public-facing enough to need public opinion on your side. Altman went to Congress and said the right things, and did a great job of acting magnanimous on Twitter after his firing, moving at light speed to get (even more) in bed with Microsoft, culminating in the complete demolition of the Board and his reinstatement in less than 3 days. I'm sure people are already writing the books about this, but if it wasn't so terrifying it would be a pretty gripping corporate drama.
substantive roll → substantive role
I wouldn’t take Sam Altman’s comments to the effect of “clearly the company can run just as well without me” as actually being evidence that the company can run just as well without him. That’s just the sort of polite thing you say to be nice to your team.
In particular nobody else is going to be as good at fundraising as Sam Altman, and fundraising is probably still an important part of OpenAI’s future plans.
He may be valuable for all sorts of other things internally, too, I don’t know. If I had to guess I would say he probably is. There’s a big difference between great startup CEOs and mediocre startup CEOs.
"If the board had not botched its execution and had more gravitas? I think things go differently."
Sounds like the board was Experiencing a Significant Gravitas Shortfall.
This is not directly relevant but I would like to mention my continued confusion at why otherwise smart and contrarian people have any respect for graham, who has always seemed to me to be the maximally generic empty suit in tech.
Go read his essays.
I've skimmed through a couple and never found anything non generic, is there one you'd say is especially insightful or worth reading?
I don't know what you think "generic" or "empty suit" mean, but it's very clear that Paul Graham isn't what those terms mean to most people. He's clearly someone who thinks for himself, has interesting thoughts & insights, cares about good outcomes in the world, and does not simply mouth standard meaningless corporate pr-speak.
Maybe if you are already steeped in startup & rationalist culture, it's possible you could think that Seinfeld Isn't Funny:
https://tvtropes.org/pmwiki/pmwiki.php/Main/SeinfeldIsUnfunny
I don't think it's even rationalist related? Feels like all his quotes are things like "it's important to have a good head for business and understand the product" or similarly empty generic statements like that.
This is a bit late, but I'd like to offer up a few things:
a) Paul Graham played a pivotal behind-the-scenes role in shaping the modern tech economy. We are living in the world he built. The existence of Y Combinator, and the social community behind it, puts competitive pressure on VCs and has created a more founder-centric, innovation-centric culture. Things could have gone differently, and in the late 90s and early 00s, it looked like they might.
b) Paul Graham's essays are aggressively lacking in style. I'll compare to Zvi, who ended this post with a reference to "The Wire", and in his most recent, has at least one reference that I simply don't get at all (which probably means there are more that I didn't even notice). That's a valid choice on Zvi's part, but so is Paul Graham's decision to write in a way that a 10-year-old can comprehend. It's the opposite of "spicy takes on Twitter".
c) Paul Graham has successfully fired Sam Altman, which seems like a rare talent one would like to have on the board of OpenAI, just in case. ;-)
Your point about Paul Graham loving his kids is a brief comment, but I have often thought our chances for survival would go up if major AI companies were led by CxOs and Board members who had children. In lieu of that, monthly bring-your-kid-to-work-days? CxO talks at high schools and middle schools?
The concept of "future generations" moves from theory to practice when you have kids.
Thank you, Zvi. Your coverage on OpenAI recently tipped the scales for me and I became a paid subscriber.