[tylervigen]

I think this is a funny joke about how HTTP is technically an API. However, with more and more website fighting back against scraping (in part due to LLMs), I suspect explicit APIs are going to become increasingly important and useful.

[mckn1ght]

This made me wonder why websites would be against scraping, at least as long as the content has DoS protection behind an edge cache. Aside from the copyright/etc issues.

Then it made me wonder, since AWS/cloudflare already have ML products and have data stored in edge caches already, if they don’t/couldn’t train directly from the caches?