AI Can Write Code, Docs and Tests — But Not Judgment, Says Engineer
arpit_bhayani · x · 2026-09-30
Arpit Bhayani (128K followers) argues that AI can now write the code, docs, tests, and even the PR description explaining why the code is correct. What it can't reliably do: decide what should be built, ask the uncomfortable question in the room, convince people to change direction, or notice everyone is solving the wrong problem.
As the cost of producing output keeps dropping, value shifts to judgment, communication, taste, context, and knowing what not to do. Soft skills are becoming hard requirements.
Related event: As AI Makes Intelligence Cheap, Judgment Becomes the Scarce Skill(2 posts)→
More from AGI Musings
- Cambridge AI researcher: 'understand' and 'think' are hazy words, stop demanding hard definitions — dioscuri · 2026-09-30
- gmiller rebuts 'own assets and you'll be safe': AI smart enough to take your job can take your assets — ZeroStateReflex · 2026-09-30
- Naval: the best job is creative expression of a lifelong learner in a free market — gefei55 · 2026-09-30
- Anduril's Palmer Luckey: unauditable chatbot AI won't guide his missiles — victor_explore · 2026-09-30
- Shame-based writing teachers seeded today's rush to AI writing tools — StewartalsopIII · 2026-09-30
- Cloudflare: AI agent traffic up 1,700% in a year, over half of web traffic now non-human — Cloudflare Blog · 2026-09-30