Experimental browser for the Atmosphere
Loading post...
{ "uri": "at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.like/3looojnlttk26", "cid": "bafyreihtwww4ltotd7l27bdtmqkrhqc7byiz4cunbf57egll4olzh7vt2m", "value": { "$type": "app.bsky.feed.like", "subject": { "cid": "bafyreigmvnpzig5lhvxvkn4mx6bk2wx4lk3budere3idxca7stpitriqui", "uri": "at://did:plc:6ond5sxlegjxpe3ismrczk3r/app.bsky.feed.post/3looobwigpc2j" }, "createdAt": "2025-05-08T19:47:39.474Z" } }
Yeah, there's lots of interest in better attention algorithms but part of me thinks we'll just switch to RAG. I want an LLM to use all human knowledge as context. Subquadratic scaling wont help much there.
May 8, 2025, 7:43 PM