Google Declaring War on the Web

Original Article ↗ Hacker News Discussion ↗

Perceived “War on the Web” and Google’s Motives

  • Many see AI Overviews as Google enclosing the open web behind a proprietary layer it controls, similar to AOL/CompuServe “walled garden” history.
  • Some argue this “war” started long ago via ads, AMP, SEO incentives and browser dominance; AI is just the final turn of the screw.
  • Others say Google is acting defensively: people already use ChatGPT/LLMs as search, so Google must compete or lose users.

Impact on Websites, Creators, and Publishers

  • Strong concern that AI answers siphon traffic and ad revenue; publishers report large Google traffic drops tied to AI features.
  • Small sites fear being “laundered” into AI output with only a tiny or missing attribution link, destroying incentives to create.
  • Some site owners are already putting content behind auth walls or blocking crawlers, accepting lower visibility to avoid uncompensated scraping.
  • Counter‑view: ad‑funded “SEO sludge” deserves to die; creators who monetize through products/services, not display ads, may be fine.

User Experience: Web Rot vs AI Convenience

  • Many describe the current web as hostile: popups, autoplay video, cookie banners, newsletter modals, click‑bait listicles.
  • For these users, AI summaries feel like “improved reader mode”: fast, ad‑free answers without wading through garbage.
  • Critics respond that AI is often wrong, cites spam/TikTok, or contradicts its own sources, and most users don’t click through to verify.
  • Some report Google’s traditional snippets getting worse, suspecting deliberate degradation to push AI usage (unproven but widely felt).

Search Alternatives, Decentralization, and Blocking Crawlers

  • People mention DuckDuckGo, Kagi, SearXNG, YaCy, StumbleUpon‑like discovery, RSS, and webrings as partial antidotes.
  • Decentralized search is seen as technically hard and under‑resourced; existing projects are niche.
  • Thread discusses practical blocking of AI crawlers (Cloudflare switches, custom filters, potential shared blacklists), but notes many bots ignore robots.txt.

Training Data, Sustainability, and Long-Term Endgame

  • Recurrent question: if AI kills the economic base for new human content, what will future models train on?
  • Speculated answers in the thread: closed‑door licensing deals with big publishers, synthetic data, expert‑curated datasets, or even real‑world sensors and robots.
  • Some think executives are ignoring the long‑term data problem for short‑term metrics and stock pressure.

Governance, Regulation, and Licensing Ideas

  • Proposals include:
    • A standardized machine‑readable “agent source” metadata/contract for LLMs (allowed excerpts, commercial terms, sponsor links).
    • Legally enforceable “no‑AI‑training” tags or licenses for text and images, with penalties and class‑action mechanisms.
    • Taxing AI/tech excess profits to fund working artists and creators.
  • Skeptics warn that “government‑enforced” tools can be weaponized by hostile or captured states and that current enforcement against big tech is weak or unwilling.

Broader Cultural and Economic Concerns

  • Anxiety that AI accelerates “de‑skilling”: professionals becoming mere front‑ends for LLMs; quality and expertise eroding.
  • Multiple comments sketch a split between a “Big Business + AI + speed” world and a “small, artisanal, slower” world, with fears the former will economically crush the latter.
  • Others see cycles: previous tech (assembly lines, mass production) also displaced artisans yet raised some living standards; AI might be a digital version of that, but with higher risks (centralization, surveillance, environmental cost).
  • Some push back on the “war” metaphor itself, arguing it’s hyperbolic compared to actual armed conflict, even if Google’s behavior is harmful.