I keep coming back to the scene in the DeepMind video documentary where Demis Hassabis realises that it’s possible just to brute-force compute all possible protein structures.

Now listening to a talk where Stephen Wolfram uses the same brute-force computation to enumerate all possible rules of a computational scheme.

Plus, recently listened to a podcast with a scientist who advised grad students to read old papers and consider how they made great conceptual leaps in the midst of computational machinery droughts; they only had their minds to figure stuff out.

The coarse take is that brute-force computation is acting as an easy-out for deep conceptual thinking. I don’t think that’s really the case but I do think that, with the rise of intelligence-on-demand, there’ll have to be practices that enshrine cognitive airgapping in order to facilitate deep, original, paradigmatic thought.