Peer to peer science is proposed as a future of science that is messy, computationally active 24/7, and examines myriad contexts rather than clean contexts. It involves computational consensus rather than peer review and journals. A decentralized sensor mapping protocol is suggested to make the best science available to all by scoring predictions based on complexity, context, and decentralized sampling. Blockchain could provide a way to account for science through proof of work and allow anyone to propose scientific computations, with a potential new post-monetary currency.