I'm very pleased with this announcement even if price didn't rise like it did I'd still be extremely happy.It's very pleasing to see our Co BRN back with a Market Capital of just over $1 BILLION again.
who can google has a big advantageNot sure if I’ve missed reading the announcement being on my phone, does anyone know if we can watch this platform announcement somewhere (?on Brainchip’s website)?
\
Why share content from this grub of a business?
Sure hope so, blackouts on Sydney Northern Beaches right now from a little extra air conditioning now the sun is going down and solar not contributing I suspect… we need lower power solutions for everything! Let’s start with akida for AI!We are outside of market hours so here is something that could be upramping if taken too seriously.
ChatGPT needs to lower its electricity consumption about 60 billion percent. Slight exaggeration but what it uses boggles the imagination and is unsustainable. There have been some articles citing the need for a neuromorphic solution to some of the processing needs of ChatGPT.
Moving to 8 bit adds an important strategic opportunity to the AKIDA -v- GPU competition.
My opinion only DYOR
FF
AKIDA BALLISTA
I must be blinded by that
Might be first use case for AKIDA scent detector?I must be blinded by thatimage but I can't see a link to the platform viewing
And what smell is it sniffing off its hand?
Not sure if I’ve missed reading the announcement being on my phone, does anyone know if we can watch this platform announcement somewhere (?on Brainchip’s website)?
here it isIch muss von diesem Bild geblendet sein,aber ich kann keinen Link zur Plattformansicht sehen
Und welchen Geruch riecht es von seiner Hand?
Thanks IDD,A few thoughts on today's announcement:
The progression to 8-bit is interesting. I personally thought we would move to 8-bit, but not for at least a year or two. My take is that 8-bit isn't exclusively for edge processing anymore. If customers were looking at AI processing in places like data centres, I don't think 4-bit would have sufficient accuracy for those heavier workloads and they would also have a mains power supply. Up to 4-bit is great for most sensor processing at the edge, while 8-bit indicates to me that they are expanding their market reach. I think it's just a matter of time till we start migrating away from the edge, and this seems to be a step in that direction.
I also don't think they would have started designing 8-bit processors YET if 4-bit applications weren't successful with some projected cash flows in place. We know there are multiple customers that are likely already using the first Akida generation chips (Renesas, Mercedes, Valeo?, other EAPs?), so my guess is things are still going ok in this area. The second gen IP will likely be more expensive to develop and buy, so they wouldn't be going down the 8-bit path if they weren't confident of customer interest. As others have suggested they may be trying to work towards the 8-bit standardisation other companies like Intel and Qualcomm are planning for AI workloads (either way it's a great move strategically to have a solution that slots directly into their workflows), or alternatively it may have been explicitly requested by some existing customers for their applications. Regardless, this is good news for Brainchips future success.
I think this is not what he is asking for, but rather about the event that takes place just before NYSE open.
Yes, we spotted your deliberate mistake! (when you pointed it out)If we're having this much trouble, with the naming of the product lineup, imagine what trouble Kochie would be having..
View attachment 31282
He's still trying to work out how to pronounce nuromorphic!
(edit. Of course I know how to spell neuromorphic! Sheezz.. Do you think I'm heavily invested in something I can't even spell, let alone understand? That was put in for effect
)
Maybe you can still do 4 and 2 bit processing as efficiently as before, with the option of doing 8 bit if you need to?Thanks IDD,
The need for 8 bits is a bit (no pun intended) surprising.
As you point out, the cloud is one option, and is urgent from a climate perspective.
Also, as you said, compatibility with the big boys could be useful.
PvdM said in "4 bits are enough"
https://brainchip.com/4-bits-are-enough/ :
With the focus on edge AI solutions that are extremely energy-sensitive and thermally constrained and require efficient real-time response, this advantage of 4-bit weights and activations is compelling and shows a strong trend in the coming years.
After mentioning that BrainChip first used 4 bits in 2015 (Studio?), he referred to some 2020 IBM research which found that 4 bits could be as accurate as 16 bit deep learning, only faster and more efficient, so maybe they need the accuracy of 32 bit or 64 bit weights and measures.
We've been working with Valeo for a few years, so it would be late in the day for the requirement to originate with them. However, we have only just learned of the Luminar/Mercedes tie up, so it may be that their foveated LiDaR needs higher precision.
ISL and the USAF radar may be another option. All those fighter planes look the same on radar.
Of course there are many things which we do not know ...
I think the issue here is how to benchmark new ways of doing AI. How do you benchmark 1 shot learning? Or edge learning? E.t.c. I don't think we really know how to do that yet.Excellent announcement. Cant wait to see the bench marking and what the new version is capable of
View attachment 31317
EgI think the issue here is how to benchmark new ways of doing AI. How do you benchmark 1 shot learning? Or edge learning? E.t.c. I don't think we really know how to do that yet.
Yes, and 1 bit as well.Maybe you can still do 4 and 2 bit processing as efficiently as before, with the option of doing 8 bit if you need to?