Scale of the threshold. The bar is 95% alive and cancer-free at 18 months after diagnosis, in four of nine countries. US five-year relative survival across all cancers is about 69% (SEER, 2014 to 2020 diagnoses), and the average hides the sites that decide the question: pancreas 13%, lung about 27%, liver about 22%, glioblastoma under 10%, and those are five-year survival figures, not cancer-free status. Getting every site to 95% remission at 18 months is not an improvement in treatment, it is the disappearance of the disease as a cause of death, and cancer is the second cause of death in the United States at roughly 600,000 deaths a year. The largest single-site jump of the last twenty years, metastatic melanoma on checkpoint inhibitors, took five-year survival from under 10% to around 50%: one site, one mechanism, fifteen years. So 2037 is less a drug-timeline question than a question about whether the resolver would accept a redefinition, and 33% is a bet on the resolver. For scale from the species that gets the same cancers at similar rates in the same houses: osteosarcoma, lymphoma and melanoma trials run in pet dogs because the natural disease is closer to ours than any mouse model; registry of those trials (mine): https://w0lph.github.io/k9/trials/

https://x.com/DrSamuelBHume/status/2102463284536328379

https://www.bbc.com/news/articles/cj38ymeyg4mo