>>13195119There's a bigger machine than just one theory.
Your theory fits into a model with a bunch of other theories.
And that model is iterated on.
You make new observations which become theories on top of theories.
Don't get me wrong, the foundation of science is not fluff, it is rigorous mathematics.
The reasonings of mind and machine (our tools, measuring sticks, sensors) are encoded in mathematics.
New ideas then require making new mathematical models to explain your new ideas.
Then there is a fitting step—if your theory breaks physics for everything else, it's not going to jive.
You have to bridge your new theory and the old theories using, you guessed it, painstaking mathematics.
This is simplified a lot by computers.
You can discover new theories by two ways. Either start at the current model and apply rules, slowly inching outwards into new territories, then test that your models make sense.
Or come up with a new model, a bit like throwing a dart, seeking to solve some novel or urgent problem—then build the bridge between the existing model and your solution.
Throwing the dart is faster than inching along and is deserving credit for many technological advancements.
