1) Human neurons are not just input-weight-output; they are more complex: more intricate inputs, cumulative processes, "bleeding off" if not enough is accumulated, etc. There are already experiments with architectures that resemble human ones more closely.
2) The human brain has significantly more neurons - it's a matter of time and computational power.
3) The process of continuous learning and usage, unlike AI where we first train and then use.
4) The sleep process - a systemic process of restructuring connections.
5) External memory sources - similar architectures are already being tested.
6) The presence of external receptors inaccessible to computers (for example, smells).
7) The hallucination system - illogical restructuring of structure or signal failures leading to unexpected results - mood, drugs, etc.
8) The ability to manipulate the environment and receive new inputs - soon possible with the connection of robot manipulators and universal AI. By the way, what will we do when AI can independently extract the most valuable resource - energy?
9) Hundreds of thousands of years of evolution + the evolutionary mechanism as a whole.
10) The ability for rapid communication among themselves - say, a car that informs another car about what is happening around the corner.
What else?
P.S. Overall, the further we go, the more it evokes not admiration for how cool neural networks are, but fear of how, in fact, simply humans are structured. In this sense, I see no reason why artificial intelligence shouldn't surpass humans in the future.