What Are We Going to Do With All This Intelligence?

Tim Lewis hosts the stargazing at Cal-a-Vie's observatory, where Steven Muskal met him one night after hours of talk under the dome. He grew up in Michigan, the youngest of five, with the freedom to be home by dark. He got his first telescope at 11, a wobbly two-and-a-half-inch scope he rebuilt in the basement, then traded the stars for a guitar at 14 and came back to astronomy around 2000. Underneath all of it is an enthusiasm he cannot quell: the best part of understanding something is getting to tell somebody else. In this conversation on The Renaissance Circle, Tim and Steve talk about: - Why people are not insignificant next to the universe: in numbers of atoms and cells, we are more a universe than the universe is - Consciousness as a leading candidate for what lies beneath spacetime, and 30 years of approaching the mystical with the same rigor as everything else - The ego as an expert pursuer and a terrible keeper of things - Science as an all-hands-on-deck engine for a civilization, and AI as a leveler that lets anyone ask questions - Why some mathematicians are uneasy with AI-solved proofs, and the printing press's early detractors - Steve's rule for his own work: he does not want to write another line of code, only to say what should be built, evaluate it, and prove it - The AI investment boom, the dot-com parallel, and Max Tegmark's Life 3.0 - Incentives: if you don't like the way something is happening, look at what the incentive system rewards - Musicians predicting the beat by milliseconds, and the few seconds when a band locks in - Conscious evolution and the widening sphere of care - Giving time as well as money, and why giving is a gift to yourself - Wealth, islands, and why the ego is never satiated - Universal income, abundance, and who benefits from AI's productivity - Going back to old science: Steve's 1998 human intestinal absorption model, built on 86 compounds, rerun on about 780, and the discovery that the original test set was a favorable one. It still reproduces, and it was not as good as reported. - Moderna's personalized cancer vaccines - Consortium voting for toxicity prediction, releasing models as free websites before the paper, and why pharma should share basic toxicity and ADMET data (OpenADMET) - What a pharma incentive structure built around human wellbeing could look like, wearables, and GLP-1 drugs Companion article on Substack: drstevenmuskal.com