BRN Discussion Ongoing

[Evermont]

The time is fast approaching where the risk/reward currently attached to BrainChip will reverse. Stock shorted will adjust accordingly.

 

[Vladsblood]

I wonder what the yearly fools will say about this lil gem of an announcement? Don’t think anyone will listen to those idiots anyways.
Nice part for me is it puts OUR lead further out in front….$$$ to us Chippers. Our luck is just beginning hold on tight!! Vlad

 

[McHale]

I'm starting to think, if I concentrate hard enough, I can simply will things into existence. That's what happens when you spend too long on the WBT thread talking to yourself.

Maybe the Jungian 100 monkey concept has become the 1000 monkeys.

 

[Euks]

The formal platform launch press release sounds interesting…

This part is so so intriguing.

I wonder if they (Peter and Sean) will be literally standing up in front of a small media team with a few partners outlying the significance of this achievement. It’s scheduled for 8am in Los Angles time….

How cool would that be….

 

[Vladsblood]

Spell check will be correct with our chip doing the work hehehe. Should’ve been Twatly fools! Lol

 

[Cardpro]

Hi Green, long time no see!

Glad to see you

 

[Cardpro]

I wonder what the yearly fools will say about this lil gem of an announcement? Don’t think anyone will listen to those idiots anyways.
Nice part for me is it puts OUR lead further out in front….$$$ to us Chippers. Our luck is just beginning hold on tight!! Vlad

Google should remove them from their search results... most of their articles look like its auto generated garbage...

 

[supersonic001]

My previously mentioned thoughts on some of the partnerships is that they may be joint ventures where both parties agree to co-develop a product (possibly Socionext, Prophesee, ...). This is not a licence agreement. There is no licence fee involved and there are no royalties payable. The parties share in the revenue from sales in proportion to their contribution to the JV.

A short time ago someone posted an email from Tony to the effect that partnerships are different from customer licence arrangements.

and after all, Partnerships are part of Rob T's remit as VP of Ecosystems & Partnerships !!

 

[JDelekto]

Hi FJ,

As we don't have the full details or patent specification, the following is just speculation.

We haven't seen anything to suggest AKD1500 can handle 8 bits. On the other hand, do we have the resources to run two developments in parallel? The release date of Q3 2023 does tie in with AKD1500.

The 8 bits will increase accuracy at the expense of time and power, so the applications will be in fields where accuracy is paramount.

The time penalty will not be such as to take it outside the concept of "real time", but will probably reduce the FPS equivalent rating. I'm not sure how this would affect our compatibility with Prophesee, but I assume we will still have the 1, 2, and 4 bit options.

I suppose the 8 bits will also apply to the model libraries.

This emphasizes the importance of N-of-M coding to provide performance advantages over the competition.

Edit: As @Boab points out, "scene segmentation" is new. This may dovetail with Luminar's "foveated LiDaR".

I'm also hoping they stick with the 1, 2, and 4-bit options. I assume they will handle 8-bit models as it looked like other members of the AI standards committee were leaning in that direction. Better to conform to the standards now than later.

I am hoping that it doesn't affect the speed too much as they claim to be scalable, where I see it having an effect on benchmarks is in the size of the training data. However, providing modes for the other "bit-ness" would allow for certain applications to still be able to take advantage of less memory for certain IoT devices.

While NVISO provided its own testing benchmarks for the Akida AKD1000, Jetson Nano, and Tegra GPU, I would really like to see some type of independent testing lab providing its own measurements. Hopefully, a full testing suite will emerge and we can compare Akida with the likes of SySense, Qualcomm, GraiMatter, and other competitors, in various AI workloads ranging from image and audio to various sensor processing tasks.

It seems that there are potential customers (such as Prophesee potentially with a commercial engagement) that seem to be very impressed by what Akida has to offer. I'd really like to be able to quantify their excitement.

 

[Fastback6666]

Without going into full detail that ANN has some very powerful wording from start to finish.
Its like Sean has gone into full beast mode.

Yes the announcement is so strong - so many technical advancements and elements in there that are seriously exciting for the techo's.

There is a small problem of most of the market not being able to understand the "additional" technical breakthroughs here...but the market does know AI is where it is at now and Brainchip are leading in the AI IP/chip arena...and brainchip has just unloaded their technical advancement dump truck on it.

What we all are keen to see are follow up $$$ related deals are announced in relation - now that is where we get back to $2 and beyond.

Though today I will us the words of Taylor Swift - the shorters are going to shake, shake, shake, shake shake and haters going to hate, hate, hate, hate, hate.

 

[robsmark]

I’ll voice my concerns when I’m unhappy with this company, but I’ll also give praise when deserved.

This announcement was great from a development standpoint, but equally as important in my eyes is that it was written with some conviction. Much better.

Now go sell it.

 

[Quatrojos]

I’ll voice my concerns when I’m unhappy with this company, but I’ll also give praise when deserved.

This announcement was great from a development standpoint, but equally as important in my eyes is that it was written with some conviction. Much better.

Now go sell it.

Yes, a formal product platform launch should do the trick. Time for Hehir to air his product to the general public...

 

[RobjHunt]

Whatta wonderful morning

Pantene Peeps.

 

[TheDrooben]

Great "management worthy" announcement. Our technological lead extended even further

Shorters getting ready to feel the burn........

Larry

 

[Draed]

For me, this states... we have produced the hardware, it was verified, but there was some modifications required. Hence brainchip went back to the bench and made the required changes... why? Because the akida 1000, although science fiction, didn't have the real world applications needed for massive gross sales... so a bit of feedback from tech monsters in the industry, a bit more tweeking and bam. Here you go, a chip that will work across multiple industries with multiple use cases, with significantly less customisation.

I ask, why would these tech companies help us design a better selling chip, rather that producing their own? Because we have the patents on the base level technology... much easier for them to have us modify it and sell it for us, than risk patent litigation. For me that's a huge deal. Our success is their success... and they are working with us because they know the sales are around the corner. Its just a matter of time now.

I felt like the tech was validated a couple years ago, but this announcement says, the sales are validated....

My opinion, DYOR.

 

[dippY22]

Ok simple. Do I have milk in my cornflakes or 30 year old scotch in my cornflakes?

start another poll and see what people say....

 

[DingoBorat]

Looks like today might be the day @DingoBorat or do I hold of a little bit longer with the overall SP still falling due to LDA cashing in.

View attachment 31233

Are you okay Rocket?..

I can imagine a sinking feeling of dread..

 

[Rise from the ashes]

Already did on another platform.
"Who's the biggest toss" you'd be surprised who's leading
@dippY22

 

[Diogenese]

I'm also hoping they stick with the 1, 2, and 4-bit options. I assume they will handle 8-bit models as it looked like other members of the AI standards committee were leaning in that direction. Better to conform to the standards now than later.

I am hoping that it doesn't affect the speed too much as they claim to be scalable, where I see it having an effect on benchmarks is in the size of the training data. However, providing modes for the other "bit-ness" would allow for certain applications to still be able to take advantage of less memory for certain IoT devices.

While NVISO provided its own testing benchmarks for the Akida AKD1000, Jetson Nano, and Tegra GPU, I would really like to see some type of independent testing lab providing its own measurements. Hopefully, a full testing suite will emerge and we can compare Akida with the likes of SySense, Qualcomm, GraiMatter, and other competitors, in various AI workloads ranging from image and audio to various sensor processing tasks.

It seems that there are potential customers (such as Prophesee potentially with a commercial engagement) that seem to be very impressed by what Akida has to offer. I'd really like to be able to quantify their excitement.

Well I suppose one advantage of having different offerings in the IP licencing business is that the customers can choose what best fits their needs.
That said, I would think that if a customer wanted the Vision Transformer (ViT), they would need to take the 8 bit silicon even if they only wanted to use 4 bits. However, they could use a 4 bit model library rather than an 8 bit library - a bit like keeping the Roller in the garage and just driving the Jag.

 

[robsmark]

For me, this states... we have produced the hardware, it was verified, but there was some modifications required. Hence brainchip went back to the bench and made the required changes... why? Because the akida 1000, although science fiction, didn't have the real world applications needed for massive gross sales... so a bit of feedback from tech monsters in the industry, a bit more tweeking and bam. Here you go, a chip that will work across multiple industries with multiple use cases, with significantly less customisation.

I ask, why would these tech companies help us design a better selling chip, rather that producing their own? Because we have the patents on the base level technology... much easier for them to have us modify it and sell it for us, than risk patent litigation. For me that's a huge deal. Our success is their success... and they are working with us because they know the sales are around the corner. Its just a matter of time now.

I felt like the tech was validated a couple years ago, but this announcement says, the sales are validated....

My opinion, DYOR.

That being the case, I wonder how it changes the testing process for these potential buyers. One would think if they were waiting on aspect improvements, testing could be accelerated and licences could be signed.
I would expect to see those on the sideline putting their money where their mouths are in the next six month if this is the case.