Coinbase Forces Engineers to Code with AI or Face Consequences

Coinbase CEO Brian Armstrong has revealed that over 40% of the cryptocurrency exchange's code is currently generated by artificial intelligence (AI), with an ambitious goal of increasing this figure to 50% within the next month. This revelation, made during a recent interview with John Collison on the “Cheeky Pint” podcast, underscores Coinbase’s aggressive push to integrate AI into its development processes. Armstrong emphasized that while AI-generated code is not yet used in all areas of the business—particularly in security-critical financial systems—the company is actively leveraging it for front-end development to accelerate progress and streamline operations.

According to Armstrong, the integration of AI into the codebase has more than doubled since April, with tools such as Cursor and Copilot playing a central role in this transformation. The CEO shared an internal mandate requiring engineers to adopt these AI-powered tools, even going so far as to fire those who failed to comply without valid justification. This move, while controversial, was aimed at ensuring the team embraced AI as a core part of their workflow. “Some people really didn’t like it,” Armstrong acknowledged, “but I think it did set some clarity that we need to lean into this and learn about it.”

Coinbase’s use of AI is not limited to engineering. Armstrong noted that the technology is now being applied in design, product management, and finance departments. The company has also begun testing AI as a participant in internal decision-making processes. For example, Coinbase uses a collaborative decision-making framework called RAPIDS, in which AI is treated as one of the contributing voices. This development signals a shift from using AI as an assistant to exploring its potential as a full-fledged decision-maker in certain business functions.

The increased reliance on AI has coincided with a broader push to transform Coinbase’s workforce into “AI-Natives,” as outlined in a previous company directive. Despite these changes, Armstrong clarified that AI is not being used to replace employees. Coinbase currently employs approximately 4,200 staff and is actively hiring in engineering and development roles, with a significant number of openings specifically targeting backend developers. Of the approximately 350 job openings on the company’s website, nearly half are for hard-skilled professionals, many with direct ties to AI development. This hiring strategy highlights the company’s dual focus on human expertise and AI integration.

However, the broader industry context reveals a more complex picture. While Coinbase is expanding its use of AI, the overall hiring landscape in the crypto sector has tightened since 2022. According to industry recruiters, developers and entrepreneurs are increasingly being drawn to AI-driven opportunities, with the sector absorbing significant capital and talent. This shift has led to a reduction in both job openings and large-scale layoffs across the crypto market. As Armstrong noted, AI is not simply a cost-cutting measure—it is being positioned as a productivity-boosting tool that can complement human labor rather than replace it.

Coinbase’s AI-driven approach is also influencing its customer experience and compliance strategies. The AI and Automation team is focused on automating key areas such as customer support and compliance using agentic AI systems. These systems are designed not only to enhance efficiency but to reimagine workflows in a fully AI-driven environment. The ultimate goal is to create a scalable platform that can be used across multiple domains within the company, enabling AI to enhance customer experience and internal processes alike.

While challenges remain—such as ensuring the accuracy and reliability of AI-generated code—Coinbase is investing heavily in training, code review, and human oversight to mitigate risks. The company is also experimenting with best practices, including monthly “AI Speed Runs” where engineers share their experiences and insights with the rest of the team. These initiatives reflect a broader commitment to responsibly integrating AI into the company’s core operations without compromising on quality or security.

Source:

[1] Coinbase's Brian Armstrong Says AI Writes 40% Of Its Code (https://cointelegraph.com/news/coinbase-says-ai-writes-nearly-half-of-its-code)

[2] Coinbase CEO: Developers who don't use AI will be fired - CIO (https://www.cio.com/article/4050185/fire-any-developer-who-doesnt-use-ai-why-coinbase-ceos-tough-message.html)

[3] Cursor vs Claude Code: Ultimate Comparison Guide (https://www.builder.io/blog/cursor-vs-claude-code)

[4] Senior Staff Software Engineer, Backend - (Platform - AI ... (https://www.coinbase.com/careers/positions/7171141)