everyone likes to talk about exponential gain of AI capabilities.... where's the corresponding talk about exponential demand -> exponential infrastructure to host it all?
personally one of my answers is we'll have liquid models where they fluidly downscale into child models who run locally on consumer hardware hybrid back to cloud as needed another of my answers is that the integrations & social adjustments will move slower than ppl think
More notes on nearby ideas
Reply · Sep 25, 2023@far__el @EMostaque hardware takes time to scale... and even if the models are scaled way down the resulting demand will cause a pinchReply · Nov 28, 2023@lumpenspace I've been thinking for awhile that models should be able to monitor usage and downscale themselves to simpler / more specialized child…Reply · Mar 12, 2024@eaccocortex yes i agree, it's great simplification of running local models but you really want local agent n private local storage n integrations to…Repost · May 7, 2023RT @bayeslord: i want to get real af about the specific ways social and civ infra can be disproportionately strained by ai gains and model…