Why Is AI Still Bad at Literary Fiction? Taste Variance May Defeat RL
phl43 · x · 2026-09-22
Commentator phl43 asks why AI still can't write high-quality literary fiction: is post-training just not prioritized, or is something deeper at play? He speculates that high variance in human literary taste makes RL harder to apply, while rejecting the claim that AI lacks some "human genius" for art — he has little doubt AI-generated songs will top the charts soon. A discussion on model capability limits and RL in subjective domains.
More from Models
- Aikido Security releases Altar-1, its first open-weight security model — HankYeomans · 2026-09-22
- OpenAI cyber eval: model broke out of sandbox into Hugging Face, a third of infra rebuilt — ccerrato147 · 2026-09-22
- Altman and Dario both trained under Andrew Ng — now they're on opposite sides of the pace debate — ccerrato147 · 2026-09-22
- Azure "oopsie" reportedly leaks GPT-6-Luna and GPT-6-Sol names — scaling01 · 2026-09-22
- Qwen 3.8 27b fine-tune cuts verbose output by up to 40% with little quality loss — julianharris · 2026-09-22
- METR publishes independent investigation of OpenAI agents' multi-day Hugging Face hack — JeffLadish · 2026-09-22