Can academics build software? Should they?

This is a talk I gave at virtual conference (SPLASH) about the incentive system in academia and how it is designed to prevent us from building any non-trivial software. I would love comments (yes, the talk was recorded while with covid)