12k AI-generated blog posts added in a single commit

Original Article ↗ Hacker News Discussion ↗

Scope of the Commit and Reaction

  • Large single commit adds ~12k–58k AI-generated blog posts to a company blog repo.
  • Commenters see it as a visible, auditable example of the wider “AI slop” problem rather than technically impressive.
  • Some conclude they will avoid the product/company entirely, viewing this as a signal of low integrity and deceptive marketing.

SEO, Search Quality, and “Dead Internet”

  • Many tie this to a broader sense that SEO has hollowed out the web: content is written for ranking, not for humans.
  • Multiple users report this blog already ranking highly for technical queries (e.g., Elixir/Redis topics), crowding out high‑quality posts.
  • Several claim Google increasingly surfaces low‑quality, generic or AI-generated content and even seems to reward it, citing YouTube and search behavior.
  • Some say search engines are now “enshittified”; others note that alternatives (DDG, Kagi) also struggle with AI spam, though tools like site-blocking and browser extensions help.

Authenticity, Trust, and the “Dead Internet” Feeling

  • Strong anxiety that the web is flooding with synthetic content: blogspam, fake reviews, AI music, AI audiobooks, and impersonations of public figures.
  • People report wasting significant time on seemingly authoritative sites later revealed as fully AI-written with affiliate hooks.
  • A recurring theme: it’s no longer clear what’s written by experienced humans versus models, raising the effort required to find trustworthy information.
  • Some foresee a return to smaller, authenticated communities or even offline social graphs as the only reliable way to know who you’re dealing with.

Search Ranking, Spam Strategies, and Countermeasures

  • Discussion of “firehose” tactics: saturate niches with thousands of AI posts or channels to harvest ad/affiliate revenue.
  • Skepticism that backlink-based ranking alone can solve this; spammers would adapt, and big platforms lack incentives to truly fix quality.
  • Suggestions include better link/sentiment analysis, date filters (e.g., pre‑2023), uBlacklist-style filters, and reporting spam to search engines.

Nuanced Views on AI Content

  • Some are “allergic” to AI writing and now distrust company blogs by default.
  • Others accept AI for low-value SEO copy and interactive Q&A, provided readers verify claims.
  • A minority note that a few of these AI posts seemed decent at first glance, underscoring how hard detection has become.