FF, very timely to go back and relook at those comments made by Peter.
It was at the 2018 AGM in Syd.
Question was put to Peter at the Meeting - In an autonomous vehicle, would the Akida chip sit side by side with the GPU which is doing the computation or could Akida replace the GPU altogether? Peter said yes you could potentially replace the GPU?
Naturally many of the SH's attending the Meeting almost fell of their seats.
Louis had to qualify this in a Webinar following the AGM.
Peter's comments relate to far in the future. The spiking neural network will likely first find its way into transducers and sensor interfaces and then maybe as a card level solution or multiple chips, off loading certain tasks as a kind of co-processor to the GPU, with the end game seeing customers
get more familiar with the architecture and what they can accomplish with Akida eating more and more into the GPU, CPU landscape.
I remember the other "bombshell" comment by PVDM at that Meeting. For all the SH's attending I'm sure right then and there they all saw the light. Peter dug deep into the science concerning the fatal flaw in current autonomous vehicle technology. He wasn’t just talking about it in terms of Von Neumann bottlenecks processing massive datasets to continue the development of AV’s using Nvidia’s GPUs. It was like the moment of truth. Peter stated that with current AV's there are no second chances. Not even with a 1 second delay. We all understand this now. The unacceptable delay in applying the breaks in an emergency due to the heavy maths compute, connectivity issues or latency getting up and back from the cloud or whatever in current CNN based tech.
I know we’re hard on any critical analysis of the Company at the moment that can’t see beyond company finances and revenues. But we have to remember how difficult it is for analysts to come to grips with this neuromorphic stuff and see the true potential of the Co. and the disruptive nature of Akida. Lets face it, even an entrepreneurial visionary like Elon Musk is going his own way. For the moment he’s sticking with Nvidia’s GPUs and continuing to develop his Tesla Autopilot tech backed up at the data center with his own Tesla supercomputers.
Ah well, pity Elon didn’t get to hear PVDM in the presentation years ago at that AGM.
(20min delay)