AI killed the city guide (and that's fine)

You can now generate a perfectly plausible city guide in eleven seconds. "Hidden gems in Lisbon", "Tokyo like a local", "48 hours in Mexico City" — a language model will produce any of them, complete with confident descriptions of ambiance and a warm closing paragraph. The generic city guide, as a genre, is dead. It was killed by infinite supply.

We build with AI ourselves, so this isn't a lament about the technology. It's an observation about what happens to the value of content when its marginal cost hits zero. And what happened is simple: the web is filling with travel guides describing restaurants the author has never smelled, written by nothing, for everyone, indistinguishable from the SEO sludge that came before — just faster.

Here's the thing though: the generic city guide was already worthless before AI. The listicle written by a content farm employee who'd never left their desk was not meaningfully better than what a model generates today. AI didn't destroy something valuable. It exposed how little value was there — and made the sludge so abundant that pretending is no longer viable.

So what survives? The one thing a model cannot generate: having been there. A model can write "this izakaya has a cozy atmosphere." It cannot know that the counter seats are the only ones worth having, that the owner switches the menu when the good fish comes in, that you should say you want whatever he recommends, or that the place two doors down with the identical facade is a tourist trap. That texture — bought with time, money, and repeated visits — is now the entire ballgame.

This is why we think verification through payment matters more every month. When a creator sells a map under their real name, to buyers who can judge and review it, "I actually go to these places" stops being a vibe and becomes an accountable claim. Charging money is, weirdly, a proof of humanity: nobody stakes their reputation on fifty pins they've never visited, because refunds and reviews exist.

The likely future of travel content splits in two. An infinite, free, AI-generated layer — genuinely useful for logistics, visas, transit, the commodity questions. And a small, paid, human layer for the only question that matters: where should I actually go, according to someone I trust who has actually been?

AI killed the city guide. Good riddance. It never deserved to live. What it can't kill is taste — and taste just became the scarcest resource in travel.