I wouldn't ask copilot. It doesn't use the most up to...

I wouldn't ask copilot. It doesn't use the most up to date nor properly referenced information even though it can search the web. I'd lean on Perplexity so that you can see where it derived its information.

As for the current thread - the comparison between Nvidia and BrainChip is akin to apples and oranges.

They're not gunning for generative capabilities. They're going for autonomous functionality. That is AI that doesn't need to connect to a central server. Much like how ChatGPT is mostly hosted on Azure(currently) and could split between Azure and Apple's servers.

The whole 'edge' mantra has been pretty frustrating in the dev world as well as it holds several different meanings and so I'd dig in a little more to get some clarity.

The way I perceive their idea of 'edge' currently is not through hosting a server that has their IP, but rather the companies with licenses to said IP to be able to build servers that use that tech.

This is why I don't think comparing Nvidia to BrainChip is a decent comparison and would rather look into avenues outside of the generative realm. They're currently marketing that platform because it's the current hype cycle.

I can see this having a better comparison specific to AI in the security(things like cameras for example) and biotech(blood analysis, hearing devices, lab grown meat).

One particular point where Nvidia might collide with BrainChip would be in manufacturing and production. This might be where BrainChip would hold the advantage because the electricity cost of running several bots in a warehouse is already pricey, let alone pinning the cost of running AI on Nvidia's servers at an enterprise scale would definitely make your eyes water.

I get the point about the current SP and I'm eagerly awaiting one of the larger companies to even utter the word 'neuromorphic' because that would be the catalyst to the current SP soaring to stupid levels.
But at the moment I don't think anything drastic will happen until this hype cycle for MOAR COMPUTE MOAR STRONG AI hits the asymptote and dips to get other companies to market neuromorphic computing more eagerly.

That being said, I definitely would like to see the numbers go up.