01What happened

The story, straight

A Hacker News post by user @trms struck a nerve by confessing a subconscious bias: they gravitate toward books published before 2022, discounting newer titles—especially from unfamiliar authors. The reasoning is simple but loaded. Before the generative AI era, every word in a published book was 'typed manually, checked manually, edited manually, proofread manually.' Even though the author uses LLMs daily for coding and acknowledges that output quality is what matters, the knowledge of human effort changes how they weight a book's authority. The post drew engagement on Hacker News, surfacing a tension many readers share but rarely articulate.

a hacker news user named @trms posted something that's been bouncing around a lot of readers' heads: they unconsciously prefer books from 2022 or earlier. the logic? before generative AI went mainstream, every word in a book was typed, checked, edited, and proofread by humans. they know that's irrational—they use LLMs themselves for coding—but the knowledge of that manual effort changes how much weight they give a book. the post resonated because it names a bias a lot of people feel but haven't said out loud.

02Spread timeline

Where it actually started

Jun 20, 2026Origin
@trms publishes essay on subconscious preference for pre-2022 books.@trms posts essay about preferring books from before AI went mainstream.
source

03Source receipts

Every claim, linked

04What's solid, what isn't

What's solid and what isn't

Confirmed
  • @trms uses LLMs regularly for coding work and does not oppose AI tools in principle.
  • The post identifies 2022 as the dividing line—roughly aligning with ChatGPT's public launch—between books felt to be reliably human-authored and those that might not be.
Disputed
  • Whether this sentiment is widespread or limited to a niche of tech-savvy readers.
  • How publishers and authors are responding to this kind of reader skepticism.
Developing
  • Growing discourse around 'pre-AI authenticity' as a market signal in publishing and other creative fields.

05Why it matters

The editorial take

This post captures a growing cultural anxiety around the integrity of published work in the AI era. It's not an anti-AI screed—it's something more unsettling: a reader who embraces AI tools but still can't trust post-2022 books the same way. As generative AI gets better at producing passable prose, this kind of subtle skepticism could reshape how readers discover, trust, and value books—particularly from lesser-known authors who lack a pre-AI track record.

this isn't anti-AI doomposting. it's a person who literally uses LLMs every day admitting they can't help but trust older books more. that's the interesting part—the bias isn't ideological, it's instinctual. and if enough readers quietly share it, the publishing world has a trust problem that no amount of 'but the output is good' fixes.