[moomin]

The thing about superdeterminism is that it's only interesting if you want to argue philosophy. If you're dealing with hidden variables (or even measurement errors) the only practical tool in your box for handling them is probability distributions.

So either way, you've got a probability distribution. And at this point people just apply Occam's Razor and get on with their lives. You can theorize an infinite number of systems that work exactly like the real world. The question is whether they're useful.

[kgwgk]

> The thing about superdeterminism is that it's only interesting if you want to argue philosophy.

Like the Many Worlds Interpretation!

[trevyn]

Could an untestable underlying theory inspire new models and implications that are testable?

[dharmab]

In a general case, yes. This happened recently with gravitational waves. The construction of multiple gravitational wave detectors across the planet allowed us to test a previously untestable hypothesis about the number of dimensions that gravity can act in.

https://www.youtube.com/watch?v=3HYw6vPR9qU

[SideburnsOfDoom]

> allowed us to test a previously untestable hypothesis

So it is testable now? Is everyone here using the word "untestable" in the same way, i.e. "untestable today with the current state of the art, but it might be tomorrow" vs. "untestable ever, as a principle, even with perfect tech"

[namanyayg]

Untestable is a layman's word, everyone here should be saying "falsifiable".

Many Worlds isn't falsifiable and hence untestable; but untestable can also mean not possible to test with current tech. Falsifiable is more accurate here.

[klodolph]

"Falsifiable" is a fantastic term but the philosophy of science did not stop with Popper, and it's not the only term we have available to discuss competing theories.

[namanyayg]

What other terms would you use? "Falsifiable" is more accurate that "untestable"?