Sam Altman Fired from OpenAI: NEW Insider Context on the Board’s Decison

**Erik Torenberg** (0:00)
Turpentine is a network of podcasts, newsletters, and more, covering tech, business, and culture, all from the perspective of industry insiders and experts.
We're the network behind the show you're listening to right now.
At Turpentine, we're building the first media outlet for tech people by tech people. We have a slate of hit shows across a range of topics and industries, from AI with Cognitive Revolution to Econ 102 with Noah Smith. Our other shows drive the conversation in tech with the most interesting thinkers, founders, and investors, like Moment of Zen and my show Upstream. We're looking for industry-leading hosts and shows along with sponsors. If you think that might be you or your company, email me at erik.turpentine.co. That's E-R-I-K at turpentine.co.

**Nathan Labenz** (0:46)
Even the people at OpenAI didn't quite have a handle on just how powerful and impactful this thing was likely to be. Do you worry about that? Do you have a plan for that? And they were kind of like, yeah, we do. We do have a plan for that. Trust us. We do have a plan for that. We just can't tell you anything about it. The engine is expected to refuse prompts depicting or asking for all the unsafe categories.
I was very, very interested to try this out. Basically, it did not work at all. Oh, just to double check, you are doing this on the new model, right? And I was like, yes, I am. And then they're like, oh, that's funny because I couldn't reproduce it.
I was like, here's a thousand screenshots of different ways that you can do it. You know, I feel like it's tattooed on my brain. But what I remember is the person saying, I'm confident I could get access to it if I wanted to. And again, I was like, what? You are on the board of the company that made GPT-3 and you have not tried GPT-4. When you see something that's technically sweet, you go for it, you know, and then you kind of figure out later what to do about it. It's so damn amazing to see this stuff happen that, you know, I think it can cloud people's judgment.
Should we have AGI as our singular goal? Or is that in its own way, ideological? Hello, and welcome to The Cognitive Revolution, where we interview visionary researchers, entrepreneurs, and builders working on the frontier of artificial intelligence. Each week, we'll explore their revolutionary ideas, and together we'll build a picture of how AI technology will transform work, life, and society in the coming years. I'm Nathan Labenz, joined by my co-host, Erik Torenberg.
So, hey, did you hear what's going on at OpenAI?

**Erik Torenberg** (2:48)
No, it's, yeah, I missed the last few days. What's going on?

**Nathan Labenz** (2:54)
Yeah. So, here we were, you know, minding our own business last week, trying to nudge the AI discourse a bit towards sanity, trying to depolarize on the margin. And, you know, God showed us what he thought of those plans, you might say, because here we are just a few days later and everything is gone haywire. And, you know, certainly the discourse is more polarized than ever. So, you know, I wanted to get you on the phone and kind of use this opportunity to tell a story that I haven't told before. So not going to like recap all the events of the last few days. I think, you know, again, if you listen to this podcast, we're going to assume that you have kept up, you know, with that drama for the most part.
But there is a story that I have been kind of waiting for a long time to tell that I think does shed some real light on this.
And, you know, it seems like now is the time to tell it. Perfect.

**Erik Torenberg** (3:55)
Let's dive in.

**Nathan Labenz** (3:57)
So where to begin? For me, a lot of this starts with the GPT-4 Red Team. So I guess, you know, we'll start again there. You know, again, don't want to retell the whole story, because we did a whole episode on that. And you can go back and listen to my original GPT-4 Red Team report, which was about just the shocking experience of getting access to this thing that was leaps and bounds better than anything else the public had seen at the time.
And, you know, just the rabbit hole that I went down to try to figure out, like, exactly how strong is this thing? What can it do? How economically transformative might it be? Is it safe or even, you know, mostly under control? And, you know, we have reported on that experience pretty extensively there. But there is still one more chapter to that story that I hadn't told. And that is of kind of how the project I thought kind of fit into the bigger picture and also how my involvement with it ended.