
Learning Strategic Programming

AI Skills with Matt Pocock
Topic 10 of 20
Published Sep 17, 2026 · Clip 1:01:08–1:03:07 (1:59)
Matt explains that strategic programming is hard to learn because its feedback loops are long — mistakes take nine months to surface, like a mixing desk you can't hear until later. AI speeds everything up, so strategic mistakes come back faster, and your codebase is the environment the agent operates in. The host notes staff engineers traditionally needed 10+ years of battle scars, and wonders whether fast AI-made fixes weaken learning. Matt questions whether companies would hire juniors without strategic knowledge, sharing Uncle Bob's suggestion to treat juniors like agents, which he calls an enormous waste of money given tactical work's falling value.