Procrastination is good
I think procrastination can be very helpful, depending what you’re procrastinating on.
Procrastinating on that main bit of functionality, that algorithm you need to test, the machine learning idea you need to verify? No good.
Procrastinating on the toil, the data collection, the styling, database setup? Very good.
There’s the caveat that the toil needs to be known to be doable, but just about all toil should be pushed back till later, once you’ve verified the idea.