Why superintelligent AI writing reads like 'a stupid person's idea of smart'
serrjoa · x · 2026-10-30
Arvind Narayanan poses a key question: are most economic tasks like chess — with a clear optimizable metric — or like writing, where quality is a multidimensional balancing act with no training signal beyond human mimicry?
Case in point: Claude's former 'neuralese' style — high information density, metaphor-stacked, coining new terms — was dialed back by Anthropic after user complaints. It sounded like 'a stupid person's idea of a smart person,' not superintelligence.
Human tolerance for density and metaphor has a tipping point: past it, the effect completely backfires. Good human writers hold back not from lack of capability but from restraint.
Related event: Anthropic dials back Claude's writing style after user backlash(3 posts)→
More from AGI Musings
- Nadella no longer believes alignment "falls out" by default, calls it a harder problem — connoraxiotes · 2026-10-30
- Blogger mocks 'pause AI' plan: humans can't even agree on basic treaties — VraserX · 2026-10-30
- Dean Ball: 21st-century institution building has much to learn from mycology — deanwball · 2026-10-30
- Why every lab is racing toward RSI: LLMs will be the ones discovering new architectures — haider1 · 2026-10-30
- The Obsoleting Project 'Is Building a Dystopia': The Bleak Reality of Job Hunting Now — GarrisonLovely · 2026-10-30
- AI-era entertainment has barely changed, argues signull — the paradigm shift is still ahead — signulll · 2026-10-30