|
|
|
|
|
by hedora
7 days ago
|
|
I'm increasingly tempted to just archive.is 100% of anubis page loads or something like that, then instruct my LLM to try archive.is if it sees anubis. If there was a plugin for this, it'd probably get close to 100% cache hit rate for agents pretty quickly. The plugin could even have an MCP endpoint that popped up windows for me to solve captchas or whatever in my primary browsing session, then uploaded the result to the shared cache. Archive.is blocks agents, I think, but you get the idea. Also, I really wonder what the economics are for your example. The JS boilerplate for the anubis landing page is usually longer than the piece of information I need. (That's ignoring the actual anubis JS, which is probably on a privacy-destroying shared CDN link). |
|