Unlocking AI Potential: From Homeschooling to Scaling AI Engineering at Coinbase

Key Takeaways

• AI-Powered Personal Management: Learn how to leverage AI agents for homeschooling, finance management, and even physical inventory using simple photo inputs.
• Treat AI Agents Like Employees: Implement role definition, access control, and progressive trust when managing AI agents.
• Securing AI Operations: Explore physical partitioning strategies to prevent data leaks between different AI agents.
• Scaling AI in Engineering: Discover the 'speed run' tactic and other strategies Coinbase used to foster widespread AI adoption among engineers.
• Focus on Eliminating Tedious Tasks: Prioritize AI applications that remove 'soul-sucking' work to drive engineer buy-in.

The latest episode of 'How I AI' showcases the diverse applications of artificial intelligence, starting with Jesse Genet, a homeschooling parent and entrepreneur who has built a sophisticated system using five specialized OpenClaw agents. These agents, each running on its own Mac Mini, manage everything from homeschool curriculum and family finances to scheduling, development projects, and household operations. Genet's approach involves treating each agent as a new hire, with defined roles, scoped access, and a decision log to maintain accountability and prevent redundant decision-making.

Genet's innovative use of AI extends to physical inventory management. By photographing educational supplies and feeding the images to her AI agent, she has created a system that bridges the digital and physical worlds. This allows the agent to suggest specific materials for lesson plans, enhancing the learning experience.

On the other end of the spectrum, Chintan Turakhia, engineering leader at Coinbase, shares his playbook for scaling AI adoption across a large engineering organization. A key strategy was the 'speed run,' a 15-minute session where 100 engineers simultaneously used AI tools to push 75 pull requests. This demonstrated AI's effectiveness and transformed the team's perception of its capabilities.

Turakhia emphasizes the importance of leading by example. Before pushing AI tools onto his team, he personally used Cursor daily, discovering use cases and techniques to build credibility. He also focused on eliminating tedious tasks that engineers dislike, such as writing unit tests and fixing linting issues.

Coinbase established a "Cursor Wins" Slack channel to showcase AI successes. This visibility created peer learning and organic adoption as engineers saw colleagues achieving more with less effort.

Another key metric for Coinbase was the entire feedback-to-feature cycle time. By reducing PR review times and automating the conversion of user feedback into shipped features, Coinbase created a virtuous cycle of continuous improvement.

Interestingly, Turakhia advocates for engineering leaders to write *more* code in the AI era. With AI automating coordination and meetings, leaders can spend more time on hands-on technical work.

Why it matters

These examples highlight the transformative potential of AI across various domains. Whether it's streamlining personal management or revolutionizing engineering workflows, AI offers powerful tools for enhancing productivity, creativity, and efficiency. The strategies outlined in this episode provide valuable insights for anyone looking to harness the power of AI, regardless of their technical background or organizational size.