Subhash Jha

Three environments, more internal services than I can count, and endpoints scattered across a dozen repos, wikis, and old Slack threads. Here's the first Claude Code skill I built that actually stuck, a router that figures out where an endpoint lives instead of memorizing it.

CEOs are declaring AGI and pricing in a workforce of models. But the thing they're racing to replace runs on twenty watts, adapts to every moment, and never collapses when it thinks about itself. A working engineer's case for biological intelligence.

Clean code, design patterns, readable names. Those are ergonomics for a human brain with seven slots of working memory. The AI doesn't have that brain. So why is it still writing code for it?

A Raycast extension, a folder full of selfies, and a midnight scheduled job that quietly rotates my Slack avatar. Friday gets the most photos. This was not an accident.

A new Raycast clone ships every week now that AI has collapsed the cost of building one. Most of them are solving the wrong problem. The actual moat was never the launcher, it was what users could build on top of it.

A pattern for treating unrelated types as if they shared a common interface, without making them inherit from anything. The same trick behind std::function, std::any, and a lot of modern C++ APIs.

Building software has become trivial. The moat has moved. Creative software companies that don't figure out where it went are going to find themselves outflanked by every enterprise that decides to just build their own.

The more powerful our tools become, the more they cost us in something we don't measure: the slow, frustrating, essential work of learning.

I've been a Raycast loyalist for years. Then BetterTouchTool 6.493 shipped a full SwiftUI launcher, and in one weekend I rebuilt three plugins that outperformed what I had in Raycast.