> Even buying that premise, is there anything in the applied physics fields that is comparable in impact to the neural network?
Why does it have to be? It's a different field. You would not expect the Nobel in chemistry to be awarded to Linus Torvalds, impact or no impact.
And the connection to physics is beyond tenuous here. The toy neural networks they cite in the document, including the Boltzmann machine, have very little to do with the power of ANNs to learn complex patterns that made such a splash recently. That is basically what the Bitter Lesson is all about. The interesting stuff does not arise from clever theory, but from practical tinkering and loooooots of compute.
Why does it have to be? It's a different field. You would not expect the Nobel in chemistry to be awarded to Linus Torvalds, impact or no impact.
And the connection to physics is beyond tenuous here. The toy neural networks they cite in the document, including the Boltzmann machine, have very little to do with the power of ANNs to learn complex patterns that made such a splash recently. That is basically what the Bitter Lesson is all about. The interesting stuff does not arise from clever theory, but from practical tinkering and loooooots of compute.