LinkedIn cofounder Reid Hoffman had a blunt message for managers leading their company’s AI strategy: Your approach is probably outdated.
Loading audio narration…
On Monday at the Semafor World Economy Summit 2026 in Washington, DC, Hoffman — now a partner at Greylock — said too many executives are treating AI like a traditional software rollout. Companies are built to test new software with a small team, polish it into a proof of concept, and then scale it across the business.
Instead, he said companies should be quickly experimenting with the new tools across every area of the business.
“What you want is people who are using AI tokens to be explaining how it is that they’re actually exploring things that are useful to the company,” Hoffman said.
His method also includes a standing weekly habit.
“We say we should have a weekly check-in,” Hoffman said. “It doesn’t have to be everyone, all the time, with each other — but a group check-in about what we tried to do new this week to use AI for both personal and group and company productivity. And what did you learn?”
That puts Hoffman somewhere in the middle of a growing divide in Silicon Valley over how aggressively to push AI.
On one side, tech leaders are stomping on the gas by pushing aggressive, top-down mandates for AI use. Shopify CEO Tobi Lütke called AI use “a baseline expectation,” while Nvidia CEO Jensen Huang has said he wants his engineers to spend half their salaries’ worth on AI tokens. Last week, a startup CEO went viral on LinkedIn after praising his four-person team for spending over $100,000 on AI tokens in a month.
Other leaders are pumping the brakes. Companies inside Chamath Palihapitiya’s 8090 startup incubator are pulling back on AI spending after the monthly bills have quickly skyrocketed, the investor said.
Hoffman’s take threads the needle, saying that teams need to move quickly into AI use, but can’t pretend like the technology is magic or ready for all prime-time uses.
For example, he isn’t trusting current models to make his financial decisions. He called AI’s tepid investing advice “business school mediocre.”
“It gives you a smart business professor answer,” he told the audience on Monday. “That’s not actually in fact how you make money in venture investing.”
Hoffman didn’t immediately respond to a request for comment from Business Insider.
