Brainchip's Edge AI revolves around the convolutional spiking neural network and if you look at all its patents (
https://patents.justia.com/assignee/brainchip-inc), it covers how you receive, classify, store and learn from a stimulus in this spiking neural network at the edge.
Knowledge on using spiking neural network to mimic functions of the brain has been around for over 20 years and it's only recent that manipulation of this network can be scaled down at the edge like with Brainchip's IP. Now, how Brainchip's Edge AI works and its uniqueness appears clear through its patents but Brainchip competitors' "Edge AI" technology appears a bit vague to me and very hard to find any information at all.
Obviously they cannot go through the same pathway as Brainchip so the question is whether these alternative pathways to "Edge AI" are efficient, robust and allow "true" one shot learning capabilities like Brainchip's IP?....the devil's in the detail!
So for me, I mainly gauge Brainchip's IP based on its acceptance by NASA and US defence force. You just cannot risk taking something up into space or used in defence systems if it's not efficient, flexible and robust! Of course the other partnerships just seals Brainchip's IP as the real deal for Edge AI...which other "Edge AI" technology currently has a wider acceptance than Brainchip's?
Just making a point here so don't please get too emotional!
(20min delay)