conv.

All stories
AIQuiet 23d · day 28

Tech blogger argues writing is AI's safest job target

A programmer contends LLMs have plateaued on prose quality while dominating code, making human writers less vulnerable than coders to AI displacement.

What to know

  • A programmer argues writing is more AI-resistant than code because LLMs plateau on prose quality while excelling at formally verifiable tasks like programming.
  • Hacker News commenters counter that writers are already losing jobs in practice, and many organizations prioritize cost over prose quality.
  • Disagreement centers on whether AI limitations in writing are fundamental (as the essayist claims) or architectural (fixable with different training or design choices).
  • Critics note that market incentives and audience indifference to AI-written content may leave writers vulnerable regardless of their work's intrinsic superiority.

The dispute Whether LLM limitations in writing are fundamental and insurmountable (essay view) versus architectural and fixable (commenter skeptics), and whether the market's willingness to pay for human prose quality is sufficient to protect writers from displacement. · positions read across 200 posts and comments

some voices

Writing is more resistant to AI than code because it requires subjective judgment and Theory of Mind that LLMs structurally struggle with.

  • “It's just that the models are inefficient at it right now, due to training and/or architecture.”

    docheinestages · Hacker News
many voices

Writers are already losing jobs to AI regardless of quality differences; organizational cost-cutting and audience indifference make prose quality less valuable than the essay assumes.

  • “There may well be some writing jobs that are safe for the reasons given in this post, but that doesn't help all the writers I know who have already lost their jobs and are struggling to find work.”

    muvlon · Hacker News
some voices

AI limitations in writing stem from fixable architectural and training choices, not fundamental constraints, so improvements should be expected.

  • “It's just that the models are inefficient at it right now, due to training and/or architecture.”

    docheinestages · Hacker News
some voices

AI writing cannot solve the core problem that good writing requires authorial intent; even perfect AI cannot read a writer's mind.

  • “A prompt complex enough where it won't get any wrong ideas has to contain as much information as the output would have.. so just write it.”

    unleaded · Hacker News

Murat Yildirimoglu Tech blogger and programmer

The record 1 articles and posts · last 28 days

  1. The Safest Job from AI May Be Writing press · HN Frontpage · ilreb · 27d ago

What people are saying 24 voices from 1 site · best of 200 · verbatim