>Want more R&D? Pay "smart" people more, instead of forcing them to try and do the startup.
... having found that I can attack problems which are 10x harder in academia, as a lowly PhD student, than I was ever allowed to come close to touching in industry, because I didn't have both a PhD and my own company...
Yeah. I would never claim I've done particularly great or world-changing research so far (all my conference papers get rejected, after all), but nonetheless, I definitely think I make more contribution to society overall working on, say, applying Bayesian deep learning to study individual differences in neuroimaging than, say, doing web apps for sports teams or embedded firmware for a product that gets canceled (my previous two jobs).
And I definitely contribute more to society working on the derp larnin' for neuro-imaging than I did when I was unemployed and getting turned down for ML engineer jobs because I hadn't done enough ML before and the stars didn't align my way.
Arrange the money so that it pays to throw smart people at hard, useful problems.
... having found that I can attack problems which are 10x harder in academia, as a lowly PhD student, than I was ever allowed to come close to touching in industry, because I didn't have both a PhD and my own company...
Yeah. I would never claim I've done particularly great or world-changing research so far (all my conference papers get rejected, after all), but nonetheless, I definitely think I make more contribution to society overall working on, say, applying Bayesian deep learning to study individual differences in neuroimaging than, say, doing web apps for sports teams or embedded firmware for a product that gets canceled (my previous two jobs).
And I definitely contribute more to society working on the derp larnin' for neuro-imaging than I did when I was unemployed and getting turned down for ML engineer jobs because I hadn't done enough ML before and the stars didn't align my way.
Arrange the money so that it pays to throw smart people at hard, useful problems.