2024 BrainChip Discussion, page-6768

This is an interesting article from 2022.

I have read it before but it now pops up on Google Scholar. While it is general in nature it references Brainchip AKIDA and others in the list of companies in this space

My renewed interest is based on the fact that since this article Brainchip unlike the other listed companies in this paper has made a quantum leap over all its competitors with the release of AKIDA 2.0 with TENNs & ViT which makes it highly compelling in the robotic space the topic discussed in this paper:

“Neuromorphic computing systems have a lot of potential for the development of smart robots and embodied AI for robotics. First, they natively support multiple parallel and asynchronous processing streams at very low power, high throughput, and low latency, leaving enough time and energy to perform intelligent computing. Second, they can run neuronal networks with a wide range of topologies, including sparse networks and feedback loops on different spatial and tem- poral scales, supporting massively parallel distributed algorithms. Last, local always-on plasticity in neuromorphic hardware enables continual learning to shape and adapt neu- ronal networks using both training data and behavioural experience. Thus, neuromorphic hardware can enable perceptive and cognitive capabilities that are extremely costly in conventional computing hardware. Neuromorphic hardware continues to develop and can be further enriched with the alternative computing prin- ciples derived from neuroscientific insights, as briefly, albeit not exhaustively, discussed above. Next, we review some examples of applications of neuromorphic hardware in robotics that have already shown promise.”

https://www.researchgate.net/profile/Yulia-Sandamirskaya/publication/361650236_Neuromorphic_computing_hardware_and_neural_architectures_for_robotics/links/64d2863cc80b930269f95ca3/Neuromorphic-computing-hardware-and-neural-architectures-for-robotics.pdf

For Intel Loihi 2 to compete with AKD1000 & AKD1500 it will need to have overcome the problems Loihi 1 had with being too cold to function without being artificially warmed up and not being able to continue operating after 220 minutes because of over heating. This disadvantage of Loihi 1 most certainly would not be an attractive feature for industrial, military, or disaster recovery robotics when you consider the value of such robotics is in being able to function in extreme conditions which exclude human participation.

My opinion only DYOR

Fact Finder