This.
I can crunch BOINC 24/7 for weeks without problems, but I can't train my TensorFlow model faster than my computer allows it.
If it can be done in a way that I can "accumulate" resources in the form of GRC to then exchange it for real resources, it will drag LOTS of attention from the machine learning community. There is nothing like this out there.
RE: Brainstorming new BOINC projects!