Brian Armstrong, CEO of cryptocurrency company Coinbase, revealed in a podcast that he fired some of his engineers after they failed to onboard with the company’s AI tools. While some engineers had a valid excuse for not doing so, those who didn’t were promptly asked to leave.
The move gives the impression that AI tools aren’t a negotiable for Coinbase, and Armstrong said that the company is already investing in employee training and monthly AI meetings.
However, forcing employees that don’t want to use AI will certainly be disruptive and frustrating for them in the long run, even if it does have productivity benefits.
Coinbase Engineers Immediately Fired After Refusing AI
Coinbase CEO Brian Armstrong has revealed that he fired engineers who didn’t adopt AI a week after being told to do so. The information was brought to light when Armstrong appeared on the podcast of John Collison, co-founder and president of the payments company Stripe.
Armstrong explained that he’d sent a message to the Slack channel with all of Coinbase’s engineers and gave them a week to onboard with the new technology. Earlier, Coinbase had secured enterprise licenses for both GitHub Copilot and Cursor to cover every engineer at the company.
This just in! View
the top business tech deals for 2025 👨💻
By the end of the week, Armstrong met with all of the engineers who had failed to onboard. While some had valid excuses, such as they had been on vacation, others who didn’t were abruptly fired, a move that Armstrong described as him going “rogue”.
AI is Not Optional, Says Coinbase CEO
It’s clear from Armstrong’s actions that he sees AI adoption as necessary, rather than an option, for the employees at Coinbase. In fact, he’s even willing to lose employees if it means the technology is deployed to the extent he wishes.
While some at the company said to Armstrong that AI adoption would be slow, predicting it would take months to get even half of the engineers on board, Armstrong is already taking measures to move this along.
As Armstrong claimed, Coinbase already hosts monthly meetings where teams or individuals who have found creative ways of using AI share what they have learned with everyone else. Armstrong is also orchestrating company-wide AI training sessions.
The move is reminiscent of Duolingo CEO Luis von Ahn, who said in a recent interview that he had introduced ‘f-r-A-I-days’ in the office as part of the company’s bid to be known as an ‘AI-first’ company.
Is Coinbase’s ‘Heavy-Handed’ AI-Approach Wise?
Armstrong claimed that a third of Coinbase’s code is written by AI, and he hopes this will be 50% by the end of the quarter. However, John Collison, on the podcast with Armstrong, seemed skeptical about how much companies and engineers should be relying on AI-generated code.
“It’s clear that it is very helpful to have AI helping you write code. It’s not clear how you run an AI-coded base.” – John Collison, Stripe co-founder and president
Indeed, there is no doubt that the method could be productive. According to a GitHub study of 500 US programmers, 92% said they were already using AI tools, and 70% said this gave them an edge in their work.
However, forcing AI onto employees could feel restrictive, and while AI is a great tool it hasn’t always succeeded when being implemented en masse. Take Klarna, for instance, which replaced workers with AI only to reverse this policy at a later date. Instead, the way forward seems to be introducing the technology slowly. .
