The road to self-improving software

We’re all starting to think in systems to get better code out of agents. Maybe you’ve explored adding loops for computer use verification, auto-fixing CI failures, and adversarial code review to get

We’re all starting to think in systems to get better code out of agents. Maybe you’ve explored adding loops for computer use verification, auto-fixing CI failures, and adversarial code review to get something worthy of human eyes on the first pass. These are huge steps forward in agentic development. @poteto calls this “building trust” with agents, which I really like. But something feels unfinished. The ultimate goal when people talk about automating development (as in a software factory) isn’t