Competitive landscape, page-625

Sounds like it could eventually (how long do you expect it to take them to develop a commercially ready chip btw?) provide a nice way to implement conventional Neural Networks down the line but upon a cursory glance I can't tell if this is asynchronous like akida or essentially just a more efficient GPU. As in, there are energy savings from CIM but no powre savings from only processing events (ie they still process the "zeroes").

The brainchip biology inspired approach is far more exciting, these guys are playing catch up to a first gen akida while PVDM and co are studying more complex structures in the brain (cortical columns and god know what else) and developing ways to implement them in silicone. While these guys are building their chip, BRN engineers are also developing new learning rules (having already licensed some already from fantastic researchers like Simon Thorpe), building larger patent moats and creating larger technology leads.

Interesting bit of news, thanks for sharing but tbh unless you have some other insight to offer I'm not even sure if this worth looking into further at this point. Just my opinion though.