Mollick: the most annoying part of long agentic tasks is language drift, not hallucinations
emollick · x · 2026-09-21
- Ethan Mollick observes that for long-running agentic LLM tasks, the biggest annoyance is no longer coding, errors, or hallucinations, but language quality drifting downward as the task progresses.
- He quips: "It's in the name! Just write better already!"
- A counterintuitive field observation: the bottleneck is shifting from getting tasks right to keeping expression stable over long horizons.
Related event: Ethan Mollick: Language Drift Is the Worst Problem in Long-Running Agents(2 posts)→
More from coding & agent
- Dev ditches Cursor, says he switched back to VS Code and loves it — Arindam_1729 · 2026-09-21
- Agent system layers are mostly transitional, engineer argues — rakyll · 2026-09-21
- Don't feed full HTML: how an agent fixes internal links without hallucination — cneuralnetwork · 2026-09-21
- Engineer quips: AGI is inevitable once we figure out the YAML stuff — rakyll · 2026-09-21
- One RTX 3090 ran Qwen 27B autonomously for 3 weeks — it shipped working CUDA kernels — skeole · 2026-09-21
- Dev buys a Meta coding subscription for its 'excellent model, crazy quota, low price' — intellectronica · 2026-09-21