Brian Barrett: I'm curious as to Zoë's thoughts here. My impression is that it's a bit of a revolving door. You go to Anthropic, and then you leave Anthropic, and you go somewhere else that's super … you just keep hopping over to somewhere that has your quote, unquote, “values” until you feel like they don't anymore, and then you just cash a check somewhere else. Is that fair to say?
Zoë Schiffer: I think that that's basically it. Yeah. I mean, I think researchers tend to come from academia, and I think, more so than some of the other positions that we see at these companies, they are very values-driven people. They have a lot of ideals that they come in the door with, and then I think they realize, oh shit, we're just working for a tech company and that tech company … I always thought it was very telling when people would say, “Well, Meta puts profits over people.” It's like—
Leah Feiger: So does everyone.
Zoë Schiffer: Yes, that is the definition of a for-profit enterprise.
Brian Barrett: Publicly traded.
Zoë Schiffer: Right, but it's also, I think, fair to say that with AI companies, as with social media companies a decade ago, the way that they talk about what they're doing is not in terms of just trying to make money for shareholders. It's a much loftier vision of what they're doing and why.
Leah Feiger: Well, it's a claim that it's entirely mission-driven, but even with that, didn't Anthropic have someone who left recently as well?
Zoë Schiffer: That was one of those notes that was very vague, and we were like, OK, well, I mean every time this happens, I think Brian and me are both in Slack being like, “Can someone find out what's going on here? What does it mean that you couldn't pursue your values at Anthropic?”
Brian Barrett: There's just so much drama inside these companies, both related to this and related to everything else. It is remarkable how much money they have to burn and how consequential their products are, given the levels of dysfunction that we're seeing. And that's just what we know, and we know a lot because it's so dysfunctional, but man, messy in there.
Zoë Schiffer: I know. It's interesting because reportedly OpenAI is gearing up to go public in the next year. There's a lot that needs to happen behind the scenes for the company to be ready for that, to be ready for that level of scrutiny. But I think this is also, at least in terms of advertising, this is something that Fidji Simo, who's the CEO of applications at OpenAI, has been really concerned with, because she was brought over after she was the CEO of Instacart, but before that, spent years and years at Meta, working quite closely with Mark Zuckerberg, and people were really nervous when she joined OpenAI, that she was going to kind of run the Meta playbook at their “little” AI lab. I'm putting little in quotes too. I spoke to her a few months ago. It sounded like she was trying to be quite thoughtful about this, but I think this is a point that both Fidji and OpenAI at large is very sensitive about. Being the bad AI company, being the one that's running at enshittification faster than its competitors. I think Anthropic is kind of hitting it where it hurts with that ad, and so, frankly, is this researcher who penned the letter in The New York Times.