"If Hanoch doesn't sell the company this could become a gigantic...

"If Hanoch doesn't sell the company this could become a gigantic stream of revenue for WBT."

Perhaps. I am not sure yet how an emulated spiking neural network in reram would compare to a dedicated hardware design for an SNN. The amount of memory to form the needed logic will dictate the cost effectiveness.

This is a pretty long road yet. I still think there is much higher payback in getting first silicon out, getting selectors developed and scaling as hard and as fast as we can. It will also help this use case immensely.

And if the SNN thing pans out, the place where it will play is in really big nets where the cheapness of the logic affords a cost advantage over specialist SNN hardware.

Again, WBT's absolute goal needs to be to get cost economies and scale to have that advantage. Let the Leti AI guys do their thing and we remain focused on the main game with the super chip heads.