I believe the value of software comes from the outcomes it makes possible. What matters to me is whether a system can accomplish something useful and whether we can trust the result.

I start with what I want software to make possible, then work backward to understand and build what it needs. Sometimes that means building a framework; sometimes it means changing the language itself. Autonomous systems are one direction this takes me: software that can decide, act, recover, and verify what it achieved.

In my experimental CPython fork, I added None-aware attribute access, working through the language and compiler to make the behavior possible. This blog is where I share those experiments, test my assumptions, and work through what I learn.