I spent a full day vibe coding with Brian Jang, an investment manager at Kakao Ventures. It was a lot of fun, and I came away with a few reflections.
-
"Developer onboarding is over." When I was leading engineering, one thing I cared about was making sure a new developer could deploy safely to production and improve the product on their very first day. That mattered because it improved a key leading indicator, engineering velocity, which in turn drives lagging indicators like business impact. Now AI makes that surprisingly easy. "Find a small improvement -> make the change -> review the code -> merge and deploy / monitor -> move to the next task." Brian had never done traditional engineering work before, and it took less than an hour for this loop to start running.
-
"Teams whose bottlenecks are meetings and communication are exactly the teams that need AI." We barely discussed what or how to build. We just said, "What if we each build the version we imagine in 30 minutes and then merge only the best parts?" Brian used Replit, I used Figma, and then we took the agreed shape and had Claude Code generate production-ready code from it. After three sessions like that, an MVP was already visible. It makes me think we may need to redefine what Lean and Agile even mean.
-
"Developers, become founders." Development is no longer as hard as it used to be, and developers are no longer the bottleneck or the scarce resource they once were. If your job title describes the work you do rather than the impact the work creates, like data scientist or software engineer, this may be a good time to rethink your role. Those titles used to be leading indicators that reliably improved lagging indicators. Now the real advantage may be the ability to contribute directly to business outcomes. Being a developer was always a hard job. What if the people who broke through and created impact simply switched tracks and became founders? Brian and I even joked that we should start a "one-person unicorn school."
