Karpathy says coding agents finally work as of December; OpenAI's Kuprel says humans need not code
Kuprel · x · 2026-09-10
Quoting Karpathy's assessment that programming changed qualitatively in December: coding agents "basically didn't work before December and basically work since," with far better quality, long-term coherence and tenacity to power through large tasks — extremely disruptive to the default programming workflow. He built a home camera video-analysis dashboard over a weekend with minimal effort.
OpenAI's Kuprel goes further: with the latest models there's no reason for people to code again, and those models should be used to clean up the mountains of buggy human code, however genius its authors were.
More from coding & agent
- Five Things You Shouldn't Vibe Code: Auth, Payments, Time Zones and More — rseroter · 2026-09-10
- Complete 2026 Guide: How to Actually Build an App with AI Without It Hurting Your Product — davemccollough · 2026-09-10
- Claude Code 2.1.267 adds maxEffortLevel cap and system prompt snapshot toggle — ClaudeCodeLog · 2026-09-10
- Claude Code 2.1.267 adds maxEffortLevel cap and per-request system prompt re-rendering — ClaudeCodeLog · 2026-09-10
- Claude Code 2.1.267 Ships 53 CLI Changes, Adds maxEffortLevel Cap Across Providers — ClaudeCodeLog · 2026-09-10
- Claude Code 2.1.267 Full Changelog: maxEffortLevel, System Prompt Snapshot and More Fixes — mark_k · 2026-09-10