Also, there's quite a lot of value in figuring out what doesn't work. A colleague of mine says that his entire PhD could have been completed in about 4 months o

Also, there's quite a lot of value in figuring out what doesn't work. A colleague of mine says that his entire PhD could have been completed in about 4 months of work if only he knew what to do to begin with. Perhaps some ai system can try a bunch of different pathways and explain why they went wrong. Perhaps that's educational for human scientists. I dunno.