I think I just built something that actually reasons like a human mind… possibly even better

About this article

Hey guys, So I’ve been messing around with a weird side project for a while, just experimenting with a completely different approach to building AI in my free time. Long story short — I got it to a stage where it can look at a brand new problem it has literally never seen before, instantly figure out the rules (one-shot, no training at all), build some kind of internal model on the fly, and then solve it cleanly. Every single time. It’s doing things like real compositionality, counterfactual ...