In a significant shake-up for tech companies and startups larger companies, Coinbase chief executive Brian Armstrong indicated on the “Cheeky Pint” podcast – hosted by Stripe co-founder John Collison – that he told his engineers to use artificial intelligence coding tools immediately or be fired. This kind of ultimatum is just one example of a larger trend of corporate leaders demanding employee make major, non-negotiable changes as a result of recent developments in AI. It is an aggressive effort to get employees to adopt AI at a speed and pace they may not be open to, and to allow a little competition in the evolving landscape of AI enabled companies.
Armstrong prompted the ultimatum after stating that he learned, after buying licenses for AI assistants like GitHub Copilot and Cursor for all engineers, that they would take time to adopt and that it would probably take months. He found this prospect unacceptable and decided to take a more direct approach.
The “Heavy-Handed” Directive
Feeling that the company could not afford to wait, Armstrong “went rogue,” as he put it, and posted a direct mandate in the company’s main engineering Slack channel. He insisted that every engineer onboard with the new AI tools by the end of the week. The message was crystal clear: familiarize and comply with the policy, or one would need to talk to him directly to explain why they couldn’t understand it.
This “heavy-handed” approach, as Armstrong himself described it, was met with some internal resistance. However, it effectively signaled that AI integration was a top priority and not an optional experiment. This was not just a matter of using a new tool, it was also about a change of culture and that they were going to do software development differently. The Saturday Showdown and Aftermath On a Saturday that had been prearranged, Armstrong was meeting with the engineers who had missed the deadlines. While some had valid reasons for their non-compliance, such as being on vacation, others did not. The result was swift and uncompromising. Those who couldn’t provide a legitimate excuse were let go. Armstrong has not disclosed the exact number of engineers terminated, but he made it clear the action was necessary to send a powerful message about the seriousness of the company’s AI push.
This episode highlights a wider reality: In an industry where AI coding assistants are becoming commonplace, refusing to change is no longer an option. The fasted iterating and adapting firms will always win in tech. Coinbase’s leadership is making clear that they are committed to changing and adapting.
Cultivating a Culture of AI Integration
Now, Armstrong is focused on creating an environment for learning and innovation for AI after the firings. Currently, Coinbase has begun a monthly meeting series in which the engineering teams can come share a way in which they have used AI creatively and innovatively to enhance their approach to work and solve their problems.
The process emphasizes a bottom-up learning process and can propel continuous learning, but also demonstrates the real-life value of the new tools all at once.
Coinbase’s mission is very ambitious. By continue to build a culture that meaningfully values and embraces AI as they did before they are continuing to build a position of leadership in a space that they will unlock large gains in productivity. Their goal is to reach to a point where 50% of their new code is written with AI by the end of this quarter.
The Broader Conversation on AI Code
Armstrong’s bold move ignited a larger conversation around the future of software development. During the episode, podcaster and fellow programmer John Collison asked a good question regarding the long-term impact of using code largely based on AI programs. He had real concerns around how companies will read, manage and maintain an open-source codebase, and said that he cannot assume its quality or consistency. Armstrong acknowledged that it was a valid concern and that it is an area that needs to be looked into and “managed”. This discussion indicates the main challenge for the tech industry: leveraging the speed and efficiency of AI, while still producing solid, secure and manageable code.
