I was on a Google Cloud Next panel in Vegas called “From Code to Cure” for @goearli, together with some of the best people in AI applications for Medicine: Isomorphic’s CTO @SergeiIakhnin (drug design with AI), Freenome’s (liquid biopsy) CSO C. C. Jimmy Lin, MD, PhD, MHS, and AI organism foundation model company GenBio’s CTO, Le Song, moderated excellently by Google’s Sr. Dir of Applied AI, @scottpenberthy. AI is finally helping to turn biology into an engineering discipline. Today, a single drug development costs over $2B and takes 12-15 years to develop. Why? One key reason is the flawed, slow and costly pre-clinical experiment process in mice and other animals, and the long, risky, and costly clinical trials in humans. This future is still 15-20 years out, but imagine this: One day, AI will help us to turn 90% of clinical trials into purely digital projects based on fully understood human bodies, tissues, cells and nuclei. In that world, fully *digitized human model bodies* get a *fully digitized new drug* *digitally injected*. That trial in, say 10,000 “digital human bodies” will take 1 day to run. Then the result will be read out, the drug iterated on, and then the next digital trial cycle begins, until the drug works. Only when all the digital trials work flawlessly, the drug gets injected into real human beings for the first time, just to confirm it all works as expected. Done. We would get drugs tested and approved in 1 year, at 1/10th the cost. Imagine what innovation in medicine that will spur - a Cambrian age of medicine is coming at that point. With AI developing exponentially, perhaps we get lucky and this will be here in ten years, not 15-20.