I think ML depended on math/statistics/computing much more than physics. It's much easier to see what Hopfield and Hinton did as inspired by mathematical models that were created to help study statistical mechanics. Which is fine and just shows how stochastic scientific discovery is.