@lumpenspace I've been thinking for awhile that models should be able to monitor usage and downscale themselves to simpler / more specialized child models for tasks that don't require full AGI capabilities... also manage compute for themselves to run tasks locally vs cloud
More notes on nearby ideas
Reply · Feb 12, 2024@deepfates "AGI" for the masses will be tactically downscaled from real AGI, and tbh what the avg user wants to do w it could be done on consumer…Reply · Jul 29, 2023@LastNPCAlex @chopstickfury01 haha idk, perhaps... ideally I'd want the AI to run on my own cloud instance w backups...Reply · Nov 15, 2023personally one of my answers is we'll have liquid models where they fluidly downscale into child models who run locally on consumer hardware hybrid…Reply · Nov 17, 2023@foomagemindset @BasedBeffJezos sama wanted compute to scale & continue training more powerful models, any way to achieve that was means to a 🚀 end