When Nick Frosst was in college over a decade ago, he was worried he was a little late to the AI game.
Frosst, a co-founder of enterprise AI startup Cohere, said on a recent episode of TechCrunch’s Found podcast that he came to this conclusion in 2012 after Geoffrey Hinton released his research that showed he had trained a neural network to successfully identify objects like cars and animals. Frosst said looking back on it, that research is quaint compared to what AI can do now, but at the time, he felt he was late to a technological breakthrough.
“I thought I had missed the boat,” Frosst said. “I remember looking and being like, ‘Wow, if only I had started undergrad like a few years earlier, I could have been in on the ground floor.’”
He wasn’t late, of course, and went on to found Cohere, which builds custom AI models for enterprise customers, in 2020. The company has raised more than $934 million in venture capital and is currently valued at $5.5 billion.
Frosst talked about why he and his co-founders left Google to launch Cohere. He also talked about how his now co-founder Aidan Gomez’s research on why general AI models would outperform more specific or verticalized ones was the basis of their approach to building Cohere.
“We’re not trying to make a consumer product like some of our competitors; we’re not building a thousand different things at once,” Frosst said. “We’re trying to make language models really useful for enterprise and that singular focus is not something you can have building inside of a multinational, massive corporation.”
Frosst also talked why he doesn’t think the AI industry should shy away from the hard questions it is getting around things like regulation and sustainability and how he’s glad that the industry is getting more realistic about what AI technology can and can’t do.
“I don’t think we’re gonna get to artificial general intelligence. I don’t think we’re gonna get to super intelligence. I don’t think we’re gonna have digital gods anywhere, anytime soon,” Frosst said. “I think more and more people are kind of coming to that realization, saying this technology is incredible, it’s super powerful, super useful, [but] it’s not a digital god. That requires adjusting how you’re thinking about the technology.”
