Silicon Valley’s top A.I. companies are locked in a cutthroat race to build artificial general intelligence (AGI). But Cohere, a quieter Canadian rival, is charting a difference course. Instead of chasing AGI, the company is focused on using A.I. to streamline business operations—and its leadership isn’t impressed with competitors touting imminent breakthroughs.
“I don’t think Sam Altman has done a service to the world by talking about how close AGI is,” Nick Frosst, co-founder of Cohere, said on a recent episode of the 20VC podcast. “There’s a lot of hype around it, there’s a lot of misleading rhetoric, there’s a lot of misinformation,” he added, referring to claims about AGI’s timeline.
Cohere’s low-key approach also defines its business model. Rather than building consumer-facing products, the startup develops A.I. tools specifically for enterprises, avoiding the friendly, chatbot-like traits popularized by OpenAI’s ChatGPT. That strategy has resonated with investors: Cohere was valued at $6.8 billion last month after raising $500 million from backers including Nvidia, Salesforce and Radical Ventures.
Altman, the CEO of OpenAI, has been among the most vocal executives regarding about A.I.’s future. Earlier this year, he declared that OpenAI is “now confident” it knows how to build AGI, telling Bloomberg that the milestone will likely be reached during Donald Trump’s presidency.
Frosst called such hype “damaging and confusing.” “I think [Altman] has made several predictions now that are wrong, and that were obviously wrong at the time he made them,” he said.
He also criticized Altman’s public warnings about A.I.’s potential existential risks, describing them as “incorrect and not helpful” in conversations about the technology’s actual dangers. Altman’s claims were “academically disingenuous, and I think it did a real disservice to the technology he loves,” Frosst said.
Frosst, who formerly worked under A.I. pioneer Geoffrey Hinton at Google, launched Cohere in 2019 alongside fellow former Google researchers Aidan Gomez and Ivan Zhang. Outside of his role at the company, he’s also known as a member of the indie rock band Good Kid, which he formed at the University of Toronto. The group, made up entirely of programmers, played Lollapalooza last year.
At Cohere, though, Frosst is all business. The company partners with enterprises like Fujitsu, Bell and Oracle, and emphasizes security and privacy across its A.I. products such as North, a platform for agentic A.I. To better train its systems for business use, Cohere recently began utilizing synthetic datasets. The goal, he said, is to keep products focused on augmenting workforces rather than entertaining consumers. “When we train our model, we’re not training it to be an amazing conversationist with you,” Frosst said, arguing that A.I.’s business potential far outweighs its everyday uses.
“If I look at my personal life, there’s not a ton that I want to automate,” he added. “I actually don’t want to respond to text messages from my mom faster—I want to do it more often, but I want to be writing those.”