People Tripled Their Traffic to a Search Engine That Does Less

The Feature Nobody Requested

On May 19th, Google announced what it apparently considered an upgrade: AI suggestions replacing autocomplete, follow-up question support, expanded Personal Intelligence pulling from Gmail and Google Photos, Search agents doing research on your behalf. The full package. The vision, finally assembled.

By May 28th, visits to DuckDuckGo's No AI search page had more than tripled. They've averaged around 84 percent above baseline consistently since the announcement, according to what DuckDuckGo told MacRumors. Traffic extensions are now available for Chrome and Firefox to set No AI search as the default.

No AI search, for clarity, offers no AI-assisted answers, no chat interface, and surfaces fewer AI-generated images. It is, in the most literal sense, a product defined by what it refuses to do.

And people cannot get enough of it.

When 'Smarter' Becomes a Problem

I've watched this cycle enough times to recognize the shape of it. A platform bets everything on a capability users didn't ask for, frames it as inevitability, and waits for adoption to validate the decision. Usually it works, because inertia is powerful and friction is real. Switching search engines is a minor but genuine inconvenience — you have to retrain your muscle memory, trust a different algorithm, let go of the idea that Google is just objectively the answer.

The fact that a meaningful number of people are doing it anyway tells you something about where trust currently lives.

This isn't a privacy story, exactly, though DuckDuckGo's positioning has always leaned that direction. This is a credibility story. Google's AI overhaul — the intelligent search box, the Personal Intelligence integrations, the agents — arrives at a moment when users have already developed a vocabulary for what AI gets wrong. They've seen the hallucinations. They've watched confident answers turn out to be fabricated. They've had to fact-check the thing that used to be the fact-checker.

So when Google announces it's making search more AI-powered, a notable slice of the user base hears that as a warning, not a feature drop.

Tom's Guide covered the traffic surge framing it as a direct response to Google doubling down. MacRumors pinned the inflection point to the I/O announcements specifically. Both are right, and together they point to something the individual coverage can't quite hold: this is a user revolt against a definition of progress.

The conventional wisdom in tech product development holds that more capability equals more value. Smarter is better. Intelligent is the direction. The whole industrial logic of AI integration rests on that premise — you add it, adoption follows, skeptics come around eventually.

What the DuckDuckGo numbers suggest is that at least some of the skeptics have decided not to wait.

There's something almost clarifying about opting into a product that announces its limitations upfront. No AI-assisted answers. No chat. Fewer AI images. A search engine that searches. You might get a worse result sometimes. You'll know the result is coming from somewhere real.

That trade is apparently worth changing your browser default for.

The irony Google probably doesn't want to sit with: its May 19th announcements were designed to demonstrate competitive strength. They may have done the opposite — not by failing technically, but by finally making the stakes of the AI bet legible enough that users could decide they didn't want to take it.

Smarter and wanted have always been different things. They just stopped overlapping on the same day.