Coinbase CEO sets AI compliance deadline, fires employees who failed to onboard AI tools
Quick Summary – TLDR:
- Coinbase CEO Brian Armstrong fired engineers who did not start using AI tools within a week
- The company has purchased licenses for GitHub Copilot and Cursor to accelerate AI adoption
- Armstrong hosted a mandatory Saturday call to question non-compliant employees, resulting in some being let go
- Coinbase is pushing for 50% of its code to be AI-written by the end of the quarter
What Happened?
Coinbase CEO Brian Armstrong issued a strict AI usage mandate for engineers and fired employees who failed to comply. The move came after Coinbase acquired enterprise licenses for AI tools like GitHub Copilot and Cursor, and Armstrong demanded fast implementation across teams.
AI or Out: Armstrong’s One-Week Mandate
When Coinbase acquired company-wide licenses for AI coding assistants, Armstrong was told that adoption might take up to two quarters. That estimate didn’t sit well with him. Instead of waiting, he took immediate action by posting an internal message on the company’s Slack channel:
“AI is important. We need you to all learn it and at least onboard. You don’t have to use it every day yet until we do some training, but at least onboard by the end of the week.”
The message included a clear warning. Engineers who didn’t onboard by the deadline were required to join a Saturday meeting with Armstrong to explain their noncompliance. Some cited legitimate reasons like travel, but others did not. According to Armstrong:
“Some of them didn’t, and they got fired.”
This no-excuse policy shocked many, both inside and outside the company. Armstrong acknowledged the tough stance, saying, “Some people really didn’t like it, by the way, that heavy-handed approach, but I think it did set some clarity.”
Coinbase’s AI Strategy in Action
Coinbase isn’t just talking about AI adoption, it’s tracking and pushing results. According to Armstrong, 33% of the company’s code is now written by AI, and the target is to hit 50% AI-written code by the end of the quarter.
To ensure employees are not only using AI tools but mastering them, Coinbase has implemented monthly “AI Speed Runs”, where top-performing engineers share insights and workflows. This approach keeps the learning hands-on and peer-driven.
Armstrong also emphasized the need for human oversight, especially given the nature of Coinbase’s financial systems. “You don’t want people vibe-coding these systems moving money,” he explained. Code reviews and human checks remain a core part of the company’s AI policy.
Interestingly, Armstrong noted that design teams are also using AI heavily, and the company is exploring how AI could support executive-level decision-making. He personally relies on AI in his own workflow as CEO.
Industry-Wide AI Pressure
Armstrong’s approach reflects a broader industry push for AI integration. Leaders from companies like Google have similarly urged employees to embrace AI in their workflows. The pressure is growing, and Coinbase is positioning itself as a leader in this transformation.
Still, not everyone agrees on how much trust should be placed in AI-generated code. Stripe President John Collison, who hosted Armstrong on the Cheeky Pint podcast, raised concerns about long-term codebase maintenance. Armstrong agreed, signaling that AI integration is still an evolving conversation, not a solved problem.
SQ Magazine’s Takeaway
I think this story makes one thing crystal clear: AI in the workplace is no longer optional, at least at companies like Coinbase. Armstrong’s actions might seem harsh, but they reflect a real urgency in the tech world to move fast or get left behind. While it’s true that AI tools are still imperfect and need oversight, ignoring them altogether could mean falling behind your peers. Personally, I believe Armstrong’s message is loud but necessary. Either adapt to AI or risk being replaced by someone who will.
