← HN · Best

Hacker News·3 min read

AI is killing All About Berlin

Learn how AI training pipelines can unintentionally sabotage the very sources they consume.

AI models are siphoning traffic from small sites, and the author of All About Berlin says his readership has collapsed as a result. He notes that the site still sustains him, but the outlook is bleak and may end next year. The core claim is that AI is killing the ecosystem that fed it.

These models are trained on massive web crawls that include the exact articles, tutorials, and code snippets from sites like All About Berlin and Stack Overflow. The training process treats the text as data points, not as living content, so the models can reproduce similar answers without sending users to the original pages. This removes the click‑through that powers ad revenue and subscriptions.

Because the AI outputs are “inaccurate shadows” of the source material, users get a quick answer but lose the context, updates, and depth that only the original authors provide. The author observes that readers now prefer an AI‑generated summary of topics like opening a German bank account rather than navigating multiple links. The convenience of a single answer outweighs the incentive to visit the source.

If you are building a service that relies on web traffic, you need to anticipate that AI summarizers will divert users away from your site. Consider adding APIs, premium content, or licensing your data to AI providers instead of relying solely on ad clicks. Diversifying revenue streams becomes essential when the primary traffic source is being abstracted away.

TakeawayAI models can drain traffic from niche sites, so creators must adapt their monetization strategies.

Prodigy briefing — continue on the original for source material, discussion, and updates.

Read original ↗