Post

I wonder if atmospheric websites (websites were most or all of their content is natively stored on the AT protocol) would alleviate this because there are so many places to get the content that aren’t just the website

If you run a blog, website, other publicly readable web thing: do you have any kind of noticeable LLM scraper bots problem (i.e. ongoing huge number of bot requests that you have to fight with)? If you do, what is it that you run and that they're requesting (and how are you defending)?

it's better than that, by design atproto alleviates the need to crawl entirely by providing incremental updates as a primitive (relays) this is actually what got me on here, I think it may be the only way to save the web

going by vibes mostly bc i am not really technically competent at ATProto, i have nevertheless fully committed to the idea that ATProto is to the 2020s what HTTP was to the 1990s

It's more like a modern RSS IMHO? It's RSS but instead of syndicating an XML document you syndicate a collection of JSON documents. Relays full the role of PuSH/WebSub. Bluesky is Google reader.

which sounds like hell for a non-wonky tool to make usefully accessible to create feeds,which in turn makes me sad that the attie announcement was met like the LORD OF THE FLIES kids killing Simon or the town chasing down Edward Scissorhands due to kneejerk ANY PLAIN LANGUAGE PROMPT AI IS BAD dogma

I say it's not ready for normal users because I think users are still going to want to examine the thing the LLM built and javascript is unfortunately not something the average user knows.

0 reposts1 like

I mean, it works for them in that they can say "give me a cat pictures feed, cutest cats first" and get roughly what they want, and they can treat the feeds like a set of saved searches. But I'd like it to be more legible.