ChinaTalk talks with 'Funeral AI' satirist Xianyu on why models code better but write worse
ChinaTalk · rss · 2026-10-29
ChinaTalk publishes a Mandarin conversation (with abridged English transcript) with Xianyu, author of the satirical Chinese AI blog 'Funeral AI'. Key points: models improved at coding from 2025 on but their creative writing deteriorated—coding data crowds out writing like whack-a-mole; he and friends are building a writing benchmark decomposing 'AI flavor' into hundreds of indicators; benchmarks function like journalism and even like Five-Year Plans, while AGI lacks quantitative definition; humor requires conviction and ordinary people's AI experiences matter more than lab narratives. Also covers meme-ability rankings of video models and Chinglish in AI.
More from AGI Musings
- Benjamin Todd: even AI builders keep getting surprised by progress — ben_j_todd · 2026-10-29
- "The fear isn't AI replacing you — it's realizing you only loved typing code" — feregri_no · 2026-10-29
- Employers worldwide pull back on AI amid errors, low-quality output and rising costs — GaryMarcus · 2026-10-29
- Early AI researchers made 9 figures while ETH/BTC core devs earned $150k a year — menhguin · 2026-10-29
- Why early no-coders excel with AI: they already built the frustration muscle — jacob_posel · 2026-10-29
- AI-designed drugs push lab monkey prices from $3,000 to $20,000+, sparking a 'monkey crisis' — davidad · 2026-10-29