[joshmarlow]

I've often wondered about this. My suspicion is that there is a limit to the complexity of mental models that humans can fluently manipulate and I think we're starting to bump into it in some cases.

I think we will eventually need a paradigm shift from science being built around human grokable models (e=mc^2) to external human manipulatable models (ie, large scale machine derived models that we can't actually grok but can use for analysis and engineering). I think we're already starting to see this - there are already mathematical proofs that are so large and complex (in the GB range) that they had to be found by automation and only other automation can verify them.

[raxxorrax]

We have tons of numerical simulations in engineering. Light modulation alone, just 3-5 lenses with different qualities can occupy a modern processor for a few hours.