[danbruc]

My ad hoc response would be that this is a question of maturity. In an unexplored area you have to experiment to find the principles that work. Once you have found them, you should use them, everyone rediscovering them on their own is probably a waste of time. On the other hand discovering something on your own can aid a deeper understanding, so maybe it is not too bad after all. And it provides an opportunity to reevaluate, maybe the environment has changed in such a way that the principles also have to change.

[makeitdouble]

> On the other hand discovering something on your own can aid a deeper understanding, so maybe it is not too bad after all

That is the take we globally landed on I think.

For instance students will be asked to prove known theorems, solve classic problems, literally reinvent the wheel. And sometimes realize it's all wrong and there's better approaches or more accurate theories.

[cushpush]

"Please derive in this 90-minute exam what took 100 years to comprehend"

[wolverine876]

I was going to write the same. Also, it implies (rightly) that changing high-impact theories should be left to experts who have fully explored the domain and the mechanisms.

There are times to disregard inconvenient aspects of theory:

Quick-and-dirty development: If you are building an passenger jet, yes stick to theory. But if you are hacking together a quick tool that won't be used past the end of the week and which will have little impact, then just get it done!

Early stage development: You're in an experimental stage anyway. Better to experiment more than to perfect the product. You can address theory later when your concept has matured.

For low-impact development, I think it's often better to encourage people to experiment, explore, and learn - to have agency - than to worry about theories. That's how people learn and engage best.