Point taken, but if you wanted to ultimately deploy your built and trained model to hardware, for use in low power environments, you’d still need to buy the standalone Mini PCIe Board or one of the Development Kits, right?
Which brings us back to the uni researcher’s cost comparison…
Not quite correct, by the way. It is the Mini PCIe Board that was/is sold for US$499, not the Raspberry Pi Development Kit. For that one you had to spend US$4995.
View attachment 53619
If you are a hobbyist I suspect so.
If you are a company like Mercedes Benz, Tata Elxsi, Valeo, Samsung, Toshiba, VVDN, Unigen, Microchip, Renesas, Megachips etc or a professional developer building a real world commercial product I would suspect not.
Indeed implicit in Brainchip moving to the IP sales model announced by Sean Hehir CEO is a complete switch to those customers who are interested in applications where more than 1 million AKD1000 are required as Anil Mankar at the 2021 Ai Field Day stated that it becomes much more economical to use the AKIDA IP at above this number.
Actually even going back as far as Mr. Dinardo in 2020 when he was speaking about the need to have the AKD1000 engineering samples to convince proof of concept customers for the IP he said something to the effect that this was understandable when a company would be spending 30, 40 or 50 odd million to bring a product to market with AKIDA.
It now strikes me that once it went to IP only it was abandoning any interest in hobbyists.
In any event as Brainchip is a fast evolving company the fact that developers can use Edge Impulse for free then download their application and send it as a file to a company to consider buying into it for commercial applications this means that even your professor with his $100 budget can now take his idea and run it on Edge Impulse without further capital investment and prove how much better it could be if AKIDA is used.
If of course he only wants to play with it at home and never intended a commercial product flow from his work then he will not even need to run it through Edge Impulse's platform and his Arduino will meet his purpose.
Thinking about it now as my original interest was simply to confirm DB's recollection while I was out and about but now that I am home, the good Professor might actually have better chosen Edge Impulse to compare his Arduino with as they provide a free platform which he, would be developers, hobbyists and children starting out if they do not have $100 to buy the Arduino could be accessing.
He could possibly then made out a legitimate argument as to why it was better to spend the $100 on Arduino and not use the free Edge Impulse platform which comes with free access to AKIDA technology. From what I read children might best be kicked off with an Arduino kit as it is tailored to children beginning their technology journey.
As neither you (if I understood your earlier post correctly) or I are techies, it might be best if someone with the correct credentials comments on this point.
What has come out of this for me though is a much better understanding of the full implications of moving Brainchip to an IP only sales model. Brainchip is definitely only going after big game now.
My opinion only DYOR
Fact Finder
brainchip.com
www.forbes.com
tenstorrent.com
