thanks John, really appreciate you writing. the purpose of the project is it put work out there and then get feedback from the community to increase the accuracy! So this is super valuable. I will make sure to include this in this week's newsletter. Please don't hesitate to correct future work.
Yes Lawrence, the future of neuromorphic computing is exciting especially for companies that can deliver edge inference near the sensor with no reliance on the cloud. Throw in low wattage and second generation advancements like TENNs, on chip learning, and ViT. And ownership of the JAST tech. Loihi is only a reference chip at this stage.
I struggle somehow putting all this in context. The potential benefits of neuromorphic computing are obvious ... while at the same time the commercially available tech (actual neuropmorhic chips availalable on the web site from BrainChip, via Megachips and others) just receiving scant academic attention. Is their Akida solution a diamond in the rough, or is there something else that is prevening widespread adoption on the edge?
no clear use case yet for which the cost trade-off is worth the performance gains. my guess is we are waiting for an obvious and fast growing applicaiton area like event-driven cameras or something.
Super interesting + well-digested take on neuromorphic computing! I especially like the open questions posed. There are two professors at my university working in collaboration with a couple national labs on the hardware-software boundary of neuromorphic computing with applications geared towards particle physics. Would love to get you all in touch!
Very rich insights as always from Lawrence!
I am sorry, but your statement about what is available, is factually incorrect.
BrainChip does have a commercially available product, and it has been available for months.
Their "Akida" chip technology is evolving and the second generation chip is soon to be released.
You might want to review your investment strategy. Their tech is certainly not OLD school.
thanks John, really appreciate you writing. the purpose of the project is it put work out there and then get feedback from the community to increase the accuracy! So this is super valuable. I will make sure to include this in this week's newsletter. Please don't hesitate to correct future work.
Yes Lawrence, the future of neuromorphic computing is exciting especially for companies that can deliver edge inference near the sensor with no reliance on the cloud. Throw in low wattage and second generation advancements like TENNs, on chip learning, and ViT. And ownership of the JAST tech. Loihi is only a reference chip at this stage.
I struggle somehow putting all this in context. The potential benefits of neuromorphic computing are obvious ... while at the same time the commercially available tech (actual neuropmorhic chips availalable on the web site from BrainChip, via Megachips and others) just receiving scant academic attention. Is their Akida solution a diamond in the rough, or is there something else that is prevening widespread adoption on the edge?
no clear use case yet for which the cost trade-off is worth the performance gains. my guess is we are waiting for an obvious and fast growing applicaiton area like event-driven cameras or something.
Super interesting + well-digested take on neuromorphic computing! I especially like the open questions posed. There are two professors at my university working in collaboration with a couple national labs on the hardware-software boundary of neuromorphic computing with applications geared towards particle physics. Would love to get you all in touch!
Please do make the introduction, would be very appreciated. Ping me Lawrence@lunar.vc. Thanks!