ClockBase Agent scanned 43,602 intervention comparisons and found 500+ that significantly reduce biological age. the top hits, rapamycin, metformin, quercetin, nicotinamide riboside, overlap almost perfectly with what the longevity community spent decades converging on. but it also flagged ouabain and fenofibrate, which nobody was really talking about. they tested ouabain on old mice and it prevented age related frailty
@danshipper@reidhoffman the hard part with this advice is that new model capabilities drop every few weeks and invalidate the workflow you just set up. it's hard to build habits on infrastructure that keeps shifting
@EricTopol@NEJM_AI@pranavrajpurkar specialist models win benchmarks but generalist models win deployment. hospitals can't run 47 different models for 47 different imaging tasks
@levelsio every workflow and work itself might end up wrapped in a world model built by AI. we're not far from work becoming something you navigate in 3D rather than click through in 2D
@danshipper the trench coat is the hard part though. sitting with teams, understanding their actual workflows, deploying forward engineers to figure out what to build. the agent is interchangeable, the context isn't