I quit my job at Amazon in May to join a startup called Icon. Best career decision I ever made, but not for the reasons you might think.
At Amazon, I was on the Amazon Q Developer team, building their AI coding assistant. You'd think being at the center of Amazon's AI developer tools would be exciting, but it was actually deeply frustrating. It was apparent to anyone outside the Amazon bubble that we were losing the AI game badly. The leadership was constantly playing catch-up because there was very little true product vision. They kept saying they wanted to move like a startup, but then had the risk tolerance of IBM.
Everything took forever. AppSec reviews, design doc reviews, architectural review boards. By the time we shipped anything, companies like Cursor and Anthropic had already iterated through ten versions. We'd spend months debating whether a feature was safe enough to release while our competitors were shipping weekly updates based on actual user feedback.
What really struck me was how Amazon's product decisions were driven by internal KPIs rather than user empathy. The most obvious example was authentication. GitHub auth is the standard for developer tools because it removes friction for the exact audience you're trying to serve. But Amazon insisted on funneling users through Builder ID, their own auth system. From an internal metrics perspective, it probably looked great (more Builder ID signups!). From a user perspective, it was just another barrier to trying the product. I watched potential users bounce off that requirement constantly.
I felt like I was reaching the ceiling of what I could learn about AI and building good products within Amazon's constraints. That's why I joined Icon. At Icon, we move at a completely different speed. We ship features in days that would have taken Amazon months to approve.
But that's not the interesting part. The interesting part is watching how my teammates work. One of them hasn't looked at actual code in weeks. Instead, he writes design documents in plain English and trusts AI to handle the implementation. When something needs fixing, he edits the document, not the code.
It made me realize something profound: we're living through the end of an era where humans translate ideas into code by hand. Within a few years, that skill will be as relevant as knowing how to shoe a horse.
What I'm Seeing Right Now
My teammate has six Claude Code terminal windows open at once, each one handling a different task or feature. He literally speaks to them one by one using Whispr Flow, and they all execute in parallel. Most of his day is spent reviewing design documents and looking at the actual web app to see the changes being made in real time. Only in very rare cases does he actually dive into the code to debug something.
This developer isn't becoming less valuable. If anything, he's becoming more valuable because he can focus on the hard problems that actually matter. Now I see him spending most of his time doing what product managers traditionally do: talking to users, understanding their problems deeply, figuring out what's actually worth building. Coding has become maybe 20% of his job, and even that 20% is mostly about understanding requirements and translating them into clear specifications. The actual implementation work that used to consume 80% of his time is now handled by machines.
... continue reading