They are complicated because they got to do something much harder than neural nets - they have to build themselves starting from a single cell. When a neural ne

They are complicated because they got to do something much harder than neural nets - they have to build themselves starting from a single cell. When a neural net will be able to secrete GPUs to enlarge itself like the brain, then it will do the same thing with brains. But then it would have to carry extra apparatus. A neural net doesn't need to operate for 80-90 years without servicing the hardware, brain builds redundancy from the start to compensate. And brains keep an energy budget of just a few watts, not kilowatts. Out of all that complex biology in the brain maybe just a small part of it is contributing to intelligent behavior. Most of it is for support or redundancy. So I don't care you can equal a neuron with a 3 layer MLP with 1000 neurons or more. A model that fits on a DVD can speak fluent language, write code, and solve tasks - that is the useful capacity we are looking for.