TEACHING AI HOW SCIENCE ACTUALLY WORKS

Summary from the TLDR newsletter

Current AI for science consists of specialized tools built on curated data sets, but much of science is hands-on experimentation and unrecorded tacit knowledge. AI scientists will need multimodal datasets that capture that tacit knowledge for them to become truly capable. The IFP has proposed an effort to generate and use the multimodal data necessary to unlock the full potential of AI for science. The program involves creating a lab that would simultaneously conduct research while recording everything and then using that data to make the whole process more productive.