DuckDuckGo search saw 28% more visits after Google said people love AI mode

Original Article ↗ Hacker News Discussion ↗

Google’s rollout of AI-generated “Overviews” in Search is polarizing users, with many complaining that the feature is inaccurate, intrusive, and makes it harder to get straightforward web results. Some people report switching to alternatives like DuckDuckGo, Kagi, or smaller indie engines—and DDG’s AI‑free endpoint briefly saw a ~28% jump in visits—while others say Google’s AI mode is genuinely useful for quick, simple questions. Commenters also argue this shift reflects Google’s deeper incentives around ads, data collection, and controlling how information is presented, raising concerns about search quality, user choice, and the future of the open web.

Reactions to Google’s AI Mode

  • Strong split: some find AI Mode genuinely useful and time‑saving, especially for quick “how do I…” or “what is…” questions; others see it as intrusive, lower‑quality, and hard to disable.
  • Several report Gemini / Google AI as more restrictive than other LLMs (e.g., refusing harmless topics), and often factually wrong or shallow.
  • Many dislike that AI overviews are inserted by default at the top of results; some use URL parameters (e.g., udm=14) or browser tricks to avoid them.

Shift Toward DuckDuckGo and Other Alternatives

  • A noticeable subset changed defaults to DuckDuckGo, often specifically to noai.duckduckgo.com or lite.duckduckgo.com to avoid AI overlays.
  • Others already used DDG and only fall back to Google with !g when needed.
  • Some argue DDG results are weaker (especially for non‑English or niche queries) and end up using !g most of the time; others claim DDG is now comparable or better than Google’s increasingly SEO‑heavy results.
  • Kagi, Brave Search, Marginalia, and Bing are mentioned as alternatives; several praise Kagi’s “AI on demand” (question mark or button) rather than by default.

Search Quality, SEO Slop, and AI Summaries

  • Broad frustration with Google’s organic results: more ads, SEO‑spam articles, and CAPTCHAs; useful content pushed down.
  • Some use AI summaries explicitly as a shield against SEO garbage, letting the model read low‑quality pages for them.
  • Others insist AI summaries can’t be trusted without manual verification and miss the old “snippets + links” flow.

Ads, Business Model, and Enshittification

  • Many frame Google as an ad company first; search and now AI are just funnels for ad revenue and data collection.
  • Concerns that AI answers will become stealth ad placements and further cut traffic and income to independent sites.
  • Recurrent theme of “enshittification”: deliberate degradation (more ads, worse UX) once dominance is secured.

Broader AI Sentiment and Control

  • Mixed feelings: heavy personal use of LLMs coexists with distrust of how they’re force‑integrated, trained on user data, and used to shape what information is surfaced.
  • Anger at opt‑out‑only data collection and aggressive, confusing AI rollouts across products.

Interpreting the “28% Increase”

  • Multiple commenters question the statistic: relative growth on a tiny base (DDG <1% share) implies only a very small fraction of Google users actually moved.
  • Some still see it as a meaningful signal of dissatisfaction and a “crack” in Google’s dominance, even if economically insignificant today.