A Post-Google World?
This article by Drew Magary gets at the heart of the craziness behind Google’s recent decision to go all in on AI and essentially declare war on its own historical business model. I’m not really sure how that’s going to work out for them, but I also have some thoughts — thoughts that I imagine would be incredibly unpopular across parts of the internet and Bluesky: which is that the World Wide Web is not the only model for the information superhighway, and I think we possibly forget that. Google’s search/ad dominance led to the world of SEO, ragebait, and our current international nightmare…maybe its interaction with the web just wasn’t healthy at any point.
Back in the ’90s, things like William Gibson’s Neuromancer, David Brin’s Earth, and the Knowledge Navigator video from Apple — which is one of my secret origin stories — presented a different world, a world that was not taken. You can also see it in the documentary on General Magic, where their initial vision was very much ‘agentic’ before the Web came and killed them stone dead. You can catch glimpses of another world in The Net, which had indepth segments on anonymous FTP servers and USENET.
But whilst I agree with Drew that Google declaring war essentially on the rest of the web is a terrible idea, I’m also wondering: is it possible that the web itself is something that might need to go away? This is one of those things I keep toying with in the back of my head; I do not particularly want to live in a world where everybody lives in Sam Altman’s walled gardens.
The new frontier models are powerful and amazing, and everybody should have them. I remember arguing on the web a few years back that all I wanted was an army of trans cat girls with a lava lamp in their basement doing crazy things locally on Llama models. I still believe that (although I’ve moved on to Qwen models in 2026). I think democratizing these things to the extent that they run in everywhere is potentially as revolutionary a change as having a computer in every home.
But I’m starting to think: can we take that a bit further? Like: “No, son, we have Google at home.” What would it take ti give everybody had their own search engine at home? Could we put something akin to 2000-era Google on your PC / Mac / Laptop? What would that look like? Is it feasible? How does it update? Sure, it doesn’t solve the money problem, but it might help keep the open web alive.
Obviously a lot of this comes from the fact that I’m completely obsessed with search. If Google isn’t going to be the only search engine around any more, then maybe we have to look back to the situation that existed in the mid-to-late ’90s. Infoseek, AltaVista, Ask Jeeves, MetaCrawler, HotBot — all the different search engines that were the way we interacted with the web; each engine had its pluses and minuses and required a bit of knowledge (or research) to work out which one to use for what scenario.
What I envision is a three tier local system. Tier 0 is your emails and documents. Tier 1 is the main search engine, where something like 100m documents from the web can be searched locally. This engine is updated on a monthly basis based on a user’s explicit wishes and implicit signals from what they’re searching, drawing on the monthly Common Crawl dump to stay current. And then there’s Tier 2, which is the connection to the outside world. Paid API sources like Kagi, Bing, and anything else that you’d like to plug in. There would be a program on top of this1 that would do routing for search queries, working out when a queries need to go to the outside world, or if they can be answered locally and delegating appropriately (balancing API costs and whatnot, obviously).
But who would build such a thing. Well…I have the local search engine prototyped already…
-
okay, I’ll call it an agent if you force me… ↩︎
