Some surprising things about DuckDuckGo

Original Article ↗ Hacker News Discussion ↗

Meta: nature of the post and HN norms

  • Some called the article a “shill/fluff piece” because it’s written by the CEO, others replied that company blogs by founders are normal HN content if interesting.
  • It emerged the CEO didn’t submit it to HN, and moderators reminded people not to attack submitters and to follow site guidelines.

Censorship, Bing dependency, and torrent results

  • A major thread questioned the “we don’t censor results” claim.
  • Critics argue DDG effectively serves censored results because upstream providers (especially Bing) already censor, and from a user’s perspective that distinction is meaningless.
  • Specific tests around DMCA‑sensitive queries (pirated media, torrent domains) suggest some sites and hashes don’t appear, while Russian torrent sites do, fueling accusations of selective or inherited censorship.
  • DDG’s responses:
    • They don’t remove results themselves, monitor upstream removals, and can add back missing results.
    • They comply with DMCA. Some argue that is censorship by any practical definition.

Search quality, speed, and captchas

  • Several long‑time users say DDG results have declined recently, especially for obscure or “literal” queries, driving them back to Google, Bing, Brave, Yandex, or SearXNG.
  • Others report the opposite: DDG is solid for literal/technical searches, using !g only when needed.
  • Complaints about over‑aggressive autocorrect and query rewriting: users want empty or sparse result pages rather than “made‑up” queries.
  • Some users in Asia report DDG feeling noticeably slower than competitors.
  • Captcha prompts on DDG (and Google) are a pain point; people speculate they’re tied to VPNs, privacy tools, or fingerprinting defenses.

AI, duck.ai, and “no AI”

  • Some came to DDG explicitly to escape Google’s AI‑heavy UX; they like being able to disable DDG’s AI features or use noai.duckduckgo.com.
  • duck.ai is praised as a simple, privacy‑oriented way to try multiple LLMs, though the interface and model‑switching UX draw criticism.
  • Others think DDG will become irrelevant if Google/Bing keep AI results proprietary and DDG can’t differentiate in AI.
  • Competing AI search experiences from Brave and Kagi are mentioned, with mixed views on quality and the broader “AI‑everywhere” trend.

Bangs and power‑user tooling

  • Bangs remain a widely loved differentiator, especially !g, !w, and site‑specific shortcuts.
  • Multiple users say bang maintenance feels neglected: broken entries, ignored submission forms, no changelog or public issue tracker.
  • DDG staff cite overwhelming spam and limited team capacity; they say submissions aren’t ignored but are de‑prioritized and tooling needs improvement.
  • Several users replicate/extend bangs via:
    • Browser keyword search/bookmarks (especially in Firefox).
    • Self‑hosted frontends like SearXNG.
    • Custom “search routers” and launchers (e.g., Alfred workflows).
  • Kagi’s similar “bangs/snaps” system is referenced as inspiration, including an open‑source shortcode list (though too large for some client‑side uses).

Privacy, tracking, and business model skepticism

  • Some distrust DDG’s privacy claims, pointing to:
    • Lack (in their view) of deep, open third‑party code audits.
    • Click‑tracking URLs that require tools like Privacy Badger/AdGuard to strip.
    • Heavy employee count vs. unclear revenue.
  • DDG counters with:
    • A U.S. market share around 3%, implying substantial ad revenue even with lower monetization than Google.
    • A formal review by an advertising industry body that accepted their privacy claims.
    • Emphasis that ads are anonymous, optional, and not based on personal profiles; they sell ad slots by keyword, not by user identity.
  • Some users remain unconvinced, preferring paid options like Kagi or other engines (Qwant, Brave, Ecosia) to avoid ad ecosystems altogether.

Interfaces, APIs, and missing features

  • Lightweight interfaces (html.duckduckgo.com, lite.duckduckgo.com) get strong praise for speed and lack of clutter/JS, though one user notes curl requests are blocked via iframe on some endpoints and wants a proper paid API instead.
  • DDG says they can’t easily offer a general search API due to upstream licensing (e.g., Bing); others point to Brave and various search APIs (SERP, Exa, Tavily) filling that niche.
  • Users request:
    • Reverse image search (image‑as‑query, like Google Images’ camera icon).
    • Better dark mode options on the HTML interface.
    • Bookmark/password sync independent of the DDG browser, for people using mixed browser stacks.

Reputation, alternatives, and miscellany

  • Opinions diverge sharply: some use DDG 80%+ of the time and celebrate its longevity and privacy mission; others say Brave, Kagi, Yandex, or even Yahoo now outperform Google and DDG.
  • There’s appreciation for extras like duck.com, no‑AI endpoints, Email Protection (tracker stripping), and support for Perl and open‑source orgs.
  • Minor topics:
    • The long, nursery‑rhyme name is still seen as a branding handicap despite the duck.com shortcut.
    • Frustration with search engines (not just DDG) rewriting queries (“did you mean…”) is widespread.
    • A few comments branch into DDG’s tech stack (Perl), hiring experiences, and remote‑work/timezone logistics mentioned on DDG’s “How We Work” page.