The point of search is finding, and Google increasingly fails at its core job. I increasingly use AI for questions I used to go to Google for, threatening their very foundation of organizing the world’s information.

For example, I was reading this lovely textbook about how, precisely, we’re fucked, and I had a question. Normally I would type this query into Google, but then I’d get shit like this, which is just more work and doesn’t actually answer my question:

In addition to being more work, these links don’t even answer my question, which is really why does burning 37 kg of gas produce more mass than you started with, how is this even possible?

Meanwhile typing the same query into ChatGPT answers my question immediately:

ChatGPT quickly tells me the answer (carbon bonds with oxygen atoms in the air), walks me through the math, and then corrects my estimate (it actually produced 114 kg of CO2). Hence, it finds what I’m looking for, which is what I used to go to Google for. Whereas Google helps me with research, the AI is actually helping me with search. This is obviously a problem for Google.

Is Google working on this? Surely. They have working AI models and could provide the same output. But there’s a big problem here. ChatGPT is such a devilish competitor because it’s not a direct competitor. ChatGPT’s business model is more like Amazon AWS, they provide a backend service that anyone can use through their API. Like, for example, Microsoft Bing. What ChatGPT creates for Google is not one competitor but millions, because the API is open to anyone.

Technology is weird, the behemoths usually eat everybody and win, but what I am personally experiencing is a break of the hold Google had on everybody, whereas it was just the place to find answers. Google used to feel magically good at this, whereas now it feels drearily shit. With Google, I have to wade through links and hope someone had the same problem/question that I had, whereas AI actually does that part for me and tries to answer my question. Whereas Google was a tool, AI is a solution, and that is of course a problem for Google.

Another problem for Google, from a business perspective, they will have to lose money to compete here. Their current ad/search business is wildly profitable because it’s computationally very cheap (near zero) to serve a page of text links. Meanwhile running an AI query burns a tree to draw a picture of one. It’s computationally very expensive, which is fine for scrappy startups and fourth-place search engines, but not for a mature company like Google which is just about rent-seeking on text links now.

Even if Google has the technology to incorporate AI into search (which it does), the cost would be astronomical at that scale, and serving their generally useless ads won’t come close to covering it. They can’t deploy the technology even though they have it because it’ll blow up their operating expenses and they won’t be able to keep buying back stock and firing people like the greedy monopolists they are. So they’re stuck, like the hypothetical monkey with his hand in the candy jar.

They had an amazing run with search, but now their core product looks like Yahoo, ie a lame collection of links, while AI is actually answering people's questions. A lot of people’s questions? No. But as AI becomes a backend service behind every app and website, the moat Google built around search becomes increasingly irrelevant. Castles will start popping up everywhere. Sophisticated users will just pay to get better search results (like I do) but even the unwashed masses will be given free AI by some venture capitalist trying to herd them into his digital domain.

Google can do the same thing, but only by making its core business less profitable, which no shareholder wants to do. It’s basically Hayden Christensen’s Innovators Dilemma, wherein big incumbents can’t move to the next domain without becoming less profitable, which no (‘public’) business is designed to do. Hence while Google can technically adapt to AI overnight, financially it’s much more difficult. And so other, hungrier, companies eat their lunch, in this case search.

Is this a prediction? No. Who would’ve predicted Microsoft threatening Google in search (through its investment in OpenAI)? It’s really just an observation that I, as a user, no longer trust Google to answer questions for me. Honestly, it had being shit at its job for a long time, but now it’s become obvious that I can get answers somewhere else.

I don’t know how this plays out for everyone else, but monopolies are built on the belief that there is no alternative, and I don’t believe that anymore. I honestly use Google less than AI for daily queries, and while that certainly doesn’t threaten its market position now, it definitely threatens its monopoly position, because monopoly depends on belief, and I just don’t believe it anymore. Google is falling behind, and their current search engine and business model has no answer for that. Maybe they should ask ChatGPT lol.