The Exchanges

Q All right, Francois, thank you so much for this. A wonderful time to bring in your co-founder, Mike, who was previously the co-founder of Zapier. First of all, how did you guys connect? How did you guys start working together?

A It was a mutual friend who introduced us. Lucas, the Boits & Weiss CEO. I, uh, my background, like you mentioned, co-founded Zapier, been basically running that company for the last, you know, uh, over 10 years and, uh, started getting into AI more deeply in the beginning of 20, 22. Um, the chain of thought paper that came out in like January that year was one that really sort of shocked me. I've been paying attention, loosely curious about AI for, for basically since the beginning of the company. Um, but, you know, Zapier's an automation company, not, not really, uh, in necessarily investing in sort of the frontier of like deep learning and machine learning. Um, and, uh, I gave a whole presentation of the company on GPT-III and saw this chain of thought paper that came out, um, you know, over a year later and was sort of really shocked that like all these sort of, uh, reasoning benchmarks that we had at the time, these language model benchmarks were showing these step function increases in score using this, like, you know, step by step, let's think step by step method. And, um, so I, I was running half the company at the time. I gave it all back to my, my co-founder Wade, uh, who is a CEO and, uh, basically just went all in on doing our research, uh, it's Appior. And I think that was one of the things that led Zapri to being a really early adopter of AI and deployer of AI into…

AI assessment note: “It was a mutual friend who introduced us. Lucas, the Boits & Weiss CEO.”

Q And, uh, let's get into the specifics of how that actually works. So, uh, you know, starting with 24, but I guess the structure is the same today. So there's, there's, uh, there's a private version, there's a public version, there is a paper prize, like, how does that all work?

A I think one of the, Goal. Some of the goals that we have for dark prize are to one initially was raise awareness of the fact that there was an unbeaten, really important AI benchmark out there that wasn't saturating, um, sort of resisted this, you know, 50,000 X scale up in pure language model systems. And this was sort of to like counter and provide sort of some public education of the fact that like, Pre-training alone is like not enough for these pure language models. Um, and this was sort of in contrast to like, again, all of the hype, all of the sort of dogma that I think you saw online. And just to put a really clear point on this, like, if you're not, if you don't really remember this era as maybe, uh, remember like, uh, last summer, uh, in California, there was that big SB, 10, 47 bill, that AI bill that was primarily regulated and introduced on the thesis that like this scaling was going to continue and lead to like really, really bad like situations. And we have to impose regulation right now, because if we don't right now, it's going to be sort of too late.

AI assessment note: “Some of the goals that we have for dark prize are to one initially”