Laurie Voss

EP 74 AI Codes: Product Engineers Decide What to Build | Laurie Voss, Arize

• Laurie Voss, an npm co-founder and head of developer relations at Arize, says the job is shifting from how to build to what to build. Product sense, which used to set senior developers apart, “is becoming absolutely central to the job.” He compares the coming role to the 1960s systems analyst, who translated business requirements into software requirements without writing any software: “That job is coming back, and we’re going to call it a product engineer,” someone who works with a team of agents even without writing code themselves. Episode takeaways
• Voss says the best AI code-review tools he has seen have about an 80% hit rate, which is “really good, but it’s not enough to put stuff out into production,” so AI-assisted senior review of the last 20% still matters. Agents are weaker still at operations (deploying code, choosing and connecting hardware), refactoring, architecture and scaling, work that depends on tacit information nobody has written down. He expects agents to stay bad at tacit information for the next five years. Episode takeaways
• Voss says nobody has a great answer yet for people starting out, but he would steer away from code-focused training. Reversing a linked list in Java “is not a thing anyone’s ever going to need to do ever again,” and he’s no longer sure he would advise spending 15 grand on a bootcamp. He would pick a deliberately theoretical computer science course (complexity, big O notation, proving software correct) or one that blends in humanities such as the philosophy of computation and the business of software. Episode takeaways