Pre-2022 Books
Fears about AI-generated “slop” are driving some readers to favor books, articles, and online posts created before roughly 2022, when large language models became widely accessible. Commenters describe a flood of low-effort, AI-assisted nonfiction and self-published titles—especially on platforms like Amazon—that are hard to distinguish from genuine work and often lack depth or accuracy. Others argue that low‑quality content and copy‑pasting long predate AI, expect traditional gatekeepers and curation to regain importance, and see long-term value in human craft even if post‑2022 work is increasingly met with suspicion.
Perceived Flood of AI-Generated “Slop”
- Many see post-2022 content (books, posts, docs) as increasingly AI-written, shallow, and cliché—“good-looking” but empty.
- Concern that production has been accelerated 1000x while human reading capacity is fixed, so consumers now do unpaid filtering.
- Some argue this is just an amplification of pre-existing slop, not something fundamentally new.
Pre-2022 Cutoff & “Low-Background Steel” Analogy
- Several commenters now prefer books, posts, and references dated before ~2022/23, treating them as “pre-contamination.”
- The “low-background steel” metaphor is repeatedly invoked for uncontaminated content.
- Others note AI-written books existed before 2022, so the cutoff is somewhat arbitrary.
Fiction vs Non-Fiction and Tech Books
- Multiple people report obvious AI “smell” primarily in non-fiction/technical books (e.g., programming, cybersecurity), especially from some publishers and Amazon self-publishing.
- Several say they haven’t yet detected AI in recent fiction, or believe it’s rare because “literature is hard.”
- Some technical writers explicitly refuse AI assistance despite the market’s devaluation of post‑2022 books.
Curation, Gatekeepers, and Discovery
- Expectation that traditional gatekeepers (publishers, editors, talent scouts) and reputational cues will regain importance to filter AI slop.
- Others counter that publishers have pushed mediocre work for decades and often prioritize sales over quality.
- Self-publishing is seen as becoming harder to trust, but not dead; authors will need stronger reputations.
Authenticity, Detection, and Demoralization
- People share that AI-detection tools flag human-only writing, including pre-1900 books, and are likened to polygraphs.
- Some track drafts in git or propose proof-of-work systems, but others note these can’t prove absence of AI help.
- Creators describe being demoralized both by LLMs’ existence and by reflexive accusations of AI use.
Counterarguments & Nuanced Uses of AI
- A minority dismiss the fear as overblown: most things were always bad; just keep judging by quality.
- Some use LLMs narrowly (phrase recall, sentence splitting, business boilerplate) while insisting on keeping their own “voice.”
- Others argue society should not tolerate AI in creative fields at all.
Coping Strategies and Outlook
- Strategies: buying older/used books, relying on recommendations, university presses, archives, and time as a filter.
- Several note there is already a lifetime’s worth of good pre‑2020 material; others say clinging to old works isn’t sustainable long term.
- Mixed optimism: some think we’ll adapt and build better fact-checking/curation; others foresee deep, lasting trust collapse in online text.