There are some who will tell you that. I’m less sure that’s true. In fact, I am starting to think that if you are a technical person who is considering a startup, Google’s fascination with adding a slow and unreliable AI chat to its results is an opening for you to put a brick on the gas pedal and absolutely run Google over.
Ask Jeeves launched in 1997, and the idea was that you’d type a natural-language query into the box, and the valet would come back with an answer. (Jeeves is named for P.G. Wodehouse’s famous character, a near-omniscient man with a “feudal spirit.” Disclosure: my cat is also named for this character.) It was popular until Google Search entered the scene — an engine with a better ability to crawl the web, among its other strengths.
This is coming after a period of obvious degradation of Google Search, caused in part by Google’s own success
One way early Google Search distinguished itself from Ask Jeeves was that it was easier to use for anyone who knew anything about. Boolean operators (“and,” “or,” and “not”) are powerful tools for narrowing search results and getting the results you want. Ask Jeeves guessed at a Boolean search for its users when they entered natural language terms — as Google does now — but guessing at a Boolean term for a user is much less powerful than a user simply entering Boolean search terms themselves. For power users, Google was obviously superior.
And now Google, for reasons best known to itself, has chosen to hamstring its search operation: by adding. This is coming after a period of obvious degradation of Google Search, caused in part by Google’s own success: whole search engine optimization teams have been built to make sure websites show on the first page of search since most people never click through to the second. And there’s been a rise of SEO-bait garbage that surfaces first.
SEO bait isn’t really new. If you look at HTML for, which is basically a time capsule from 1997, you’ll see a string of words at the end, rendered in the same color as the background of the website. Those words are search terms and reflect the SEO of the time: “extraterrestrial,” for instance, is repeated 14 times, to give it priority among crawlers. But trying to find that website illustrates the problem with Google Search now — because it prioritizes recency and because SEO has changed since then, the original cult website isn’t in the top 10 results, or even the top 50, for “Heaven’s Gate.”
I amto observe that over the years. Some of the problem is, of course, that the web is bigger than it used to be. Some of it is new SEO tactics. Some of it is the increasing volume of — Google really shat in its own nest with that one. Some of it is inexplicable: what deal with the devil did Pinterest make to get its largely useless results ranked so high?
“I looked for some products in Bard, and it offered me a place to go buy them, a URL, and it doesn’t exist at all.”
The switch to an AI chatbot-like search response, though, tells me that Google doesn’t understand why it beat Ask Jeeves. It was a better experience for power users. Most of us don’t want a questionably accurate summarized answer derived from unknown websites — we want to be able to judge the source of our answer by looking at the original copy. Downranking links — that is, the websites that might be potentially made by real actual people — means removing that. And crucially, AI has slowed Google’s responses. Are you patient enough to wait for a maybe-wrong answer from God knows where? I’m not.
I mean, even Sundar Pichai has noticed the problem: “I looked for some products in Bard, and it offered me a place to go buy them, a URL, and it doesn’t exist at all,”. Cool product, bro!
For the first time in my life, I think Google Search is vulnerable. Take Bluesky, which is — to a user — almost exactly like Twitter, except that it’s not full of, , and Elon Musk. One possibility to beat Google: a startup that looks and feels like old Google Search but returns better results, perhaps by choosing different dimensions to index on. Maybe someone clever can use AI under the hood to filter out the garbage.
But maybe trying to figure out how to crawl the web in its degraded, AI-glurge-ridden, SEO-bombed state is too technically tricky. There has been a view —— that AI will devalue human labor. But what if the rise of AI-written nonsense makes humans more valuable? If you want a human result, not some machine-generated SEO bait, you may want a human involved in the process of search. You might move toward having a curator.
In fact, that’s another early-internet search model that could be resurrected: the Yahoo! Directory, which started in 1994. At the time, it was one of the most useful ways to get around the web: a human-indexed directory meant to be browsed by users. It was crushed by Google Search’s supremacy and closed in 2014. There’s even an existing signal here that’s valuable: the popularity ofI don’t know how long that will work in the LLM era, but it does suggest a number of users really want results from other people.
I don’t know what the actual solution is. All I know is that I smell weakness, and if you’ve ever been the kind of person who longed to take down a giant using only a slingshot, this is your moment. Because the main problem with turning Google Search into Ask Jeeves is that we already know what happened to Ask Jeeves.