Emacs and AI: A programmer's journey into the 'Mathocalypse'

A software developer reflects on the profound shift in programming brought on by powerful AI coding assistants like Claude Code, which they've used for the past year. They describe the initial experience as so transformative they had to 'pinch' themselves, and note that the tools have grown an 'order of magnitude' more capable. The piece situates this within a broader technological revolution, referencing the recent 'Mathocalypse' where AI models solved thousands of research problems in hours.

The article contrasts the traditional, deterministic 'digital age' of precise transistors and rigid logic with the new, stochastic era of AI. It argues that our old, sequential conception of software couldn't handle the unbounded complexity of tasks like human dialogue, but that AI's rule-based, diagnostic approach—akin to how Richard Stallman once described analyzing a circuit 'as a human would'—is now proving powerful. The takeaway is a thoughtful observation on a paradigm shift in how we build and think about software.