BRN Discussion Ongoing

[ndefries]

i still dont understand how we dont have an announceable ageement with Mercedes. Did they really work through Renesas. Surely Mercerdes can't be this far ahead with relying on our tech that they are advertising it globally without coming to commercial terms. If it is the latter is there some model where the market can get some update on this?

 

[mcm]

Just finished listening but the big takeaways for me on the positive side are:

1. The fact that Mercedes is front and centre no questions asked it is happening approach. If you only listen to a couple of minutes listen to the beginning about Mercedes.

2. The fact that when showing later in the presentation the slide of the Early Adopters and Licensees Ken Scarince the Chief Financial Officer states regarding Early Adopters - 'We have got many more but we can't disclose them as we are viewed as a KEY DIFFERENTIATOR.'

3. The fact that when speaking about SiFive being announced Ken Scarince states that they have more similar announcements to come.

4. The fact that MegaChips is receiving two different forms of the AKIDA technology and they are actively out in the market selling AKIDA and Brainchip is very excited about what MegaChips is doing with its customers.

5. The fact that Ken Scarince confirmed that Renesas is bringing out an MCU with AKIDA technology on board. - no ifs, buts or maybes.

6. The fact that Ken Scarince stated their present focus is on adding to the sales and marketing teams to completely revamp that side of the business.

The final takeaway for me which is neutral at the moment was Ken Scarince received a direct question regarding AKD2000 and stated it was still in Perth at the Research Centre and he did not expect it would come out this year but stated he was hesitant to say much about it.

He did say that the Brainchip Engineers in the US were working on changes to meet existing customer requirements. He also said the Engineers are very, very busy and likely to be so for the foreseeable future.

Question/s for the AGM.

1. Are they holding back AKD2000 because of the overwhelming demand for the existing product AKD1000 requiring the complete attention of Anil Mankar and his engineering team in the US.

2. Is the AKD2000 IP available for licence or trial by customers in the same fashion that the IP for AKD1000 was available two years before the chip was released?

3. If so is it being trialled by automotive manufacturers or automotive OEM's?

My opinion only DYOR
FF

AKIDA BALLLISTA

Thanks for the excellent summary. Sure beats me why anyone would be selling BRN at the moment given some of the things Ken Scarince had to say. Akida viewed as the KEY DIFFERENTIATOR? Now there's a nice description of our chip.

 

[zeeb0t]

Hi Zeebot, how come Im tagged as an observer, I joined not long after the switch from HC?

Your title will change from observer as you post. Which should occur soon since you have

 

[TasTroy77]

Just finished listening but the big takeaways for me on the positive side are:

1. The fact that Mercedes is front and centre no questions asked it is happening approach. If you only listen to a couple of minutes listen to the beginning about Mercedes.

2. The fact that when showing later in the presentation the slide of the Early Adopters and Licensees Ken Scarince the Chief Financial Officer states regarding Early Adopters - 'We have got many more but we can't disclose them as we are viewed as a KEY DIFFERENTIATOR.'

3. The fact that when speaking about SiFive being announced Ken Scarince states that they have more similar announcements to come.

4. The fact that MegaChips is receiving two different forms of the AKIDA technology and they are actively out in the market selling AKIDA and Brainchip is very excited about what MegaChips is doing with its customers.

5. The fact that Ken Scarince confirmed that Renesas is bringing out an MCU with AKIDA technology on board. - no ifs, buts or maybes.

6. The fact that Ken Scarince stated their present focus is on adding to the sales and marketing teams to completely revamp that side of the business.

The final takeaway for me which is neutral at the moment was Ken Scarince received a direct question regarding AKD2000 and stated it was still in Perth at the Research Centre and he did not expect it would come out this year but stated he was hesitant to say much about it.

He did say that the Brainchip Engineers in the US were working on changes to meet existing customer requirements. He also said the Engineers are very, very busy and likely to be so for the foreseeable future.

Question/s for the AGM.

1. Are they holding back AKD2000 because of the overwhelming demand for the existing product AKD1000 requiring the complete attention of Anil Mankar and his engineering team in the US.

2. Is the AKD2000 IP available for licence or trial by customers in the same fashion that the IP for AKD1000 was available two years before the chip was released?

3. If so is it being trialled by automotive manufacturers or automotive OEM's?

My opinion only DYOR
FF

AKIDA BALLLISTA

FF brilliant summary
Ken also mentioned research with a mining and Transportation company towards the end.

 

[JK200SX]

Not if you are in Vic.
View attachment 3932
Its great to be a shareholder.

I don't actually love AKIDA, because that would be weird for me, but I do have AKIDA

 

[Fullmoonfever]

FF brilliant summary
Ken also mentioned research with a mining and Transportation company towards the end.

Wonder if that mining / transport research ties in with the below I posted on HC couple months back?

Anyone up for AEB.

Saw there was a handful of AEB posts back in 2020.

Searching for industrial application info and came across this guy in Perth?

Been on this project last year and half according to profile.

Could be independent but given he started a while ago, could have been given early access or engaged by BRN or someone else local.

Be handy for either road vehicles or even say mine vehicles like autonomous haulage at Rio.

https://www.fmgl.com.au/in-the-news...etes-world-leading-autonomous-haulage-project

https://au.linkedin.com/in/peter-zhou-2a745a1b0

https://s5system.com

 

[Learning]

I don't actually love AKIDA, because that would be weird for me, but I do have AKIDA

View attachment 3939

Since we're sharing, my love is not just for Akida,

Because the heart get what the heart wants.
Its great to be a shareholder

 

[Diogenese]

This presentation was on last night. Hope this link works.
https://vic-technology-apr22.hubb.me/fe/schedule-builder/sessions/902346


View attachment 3931

Hi White horse,

great find.

Among other things of interest - re MegaChips: We are giving them a couple of different versions of Akida - so maybe this is the origin of Akida 500, the pared down version of Akida for applications which do not need the full Akida 1000 capabilities?

Ken also suggested we look at MegaChips web page:

https://megachips.com/artificial-intelligence-machine-learning/

Brainchip features a unique neuromorphic processor which takes advantage of the sparse data streams associated with most edge applications to provide a very low power solution while maintaining high performance. The small footprint of the Brainchip IP makes it an ideal solution for low power, high volume applications.
Megachips’ engineering team is fully trained on Brainchip software and hardware and is prepared to work directly with our customers to evaluate the performance of the Brainchip Solution in your product.
Once selected MegaChips will work closely with our customers throughout the product development cycle, including licensing the SDK hardware and software, provide NN model consulting, and ensure the final design can be efficiently implemented into silicon.

Key Features​

• Very low power – microwatts to milliwatts
• On chip learning – no need to retrain in the cloud
• Complete development system support
• Silicon-proven IP
• Software Development Kit available

Key Applications​

• Keyword spotting
• Always on detection
• Gesture recognition

• Network packet inspection
• Person/object classification
• Object detection

Then they go on about quadric.

Quadric offers a unique “Vortex Core Architecture” which is unique in its ability to accelerate all steps in the processing of data from a variety of sensors. Typical AI engines only accelerate the Neural Network (NN) processing, while signal conditioning, non-maximal suppression, and other processing steps are normally done by the host processor.
This unique solution provides optimal performance in many demanding applications and offers the flexibility to adapt to new NN processing algorithms as they emerge.

Key Features​

• Scalable architecture to address a wide range of power/performance requirements
• Ability to accelerate all stages of the AI pipeline
• Flexible architecture can adapt to evolving AI models
• Silicon-proven IP
• Software Development Kit available

Key Applications​

• AI inferencing
• Data pre/post processing
• Image processing
• Path Planning

 

[Dhm]

For those who are only concerned with money the slide regarding the addressable market by 2025 has increased the 60 billion US dollars to 71 billion US dollars and the $60 billion has become 61 billion US dollars in 2024.

My opinion only DYOR
FF

AKIDA BALLISTA

I'm not sure if it has been previously mentioned, but it seems we have a Brainchip acknowledged competitor named GrAI Matter. Around the 16 minute mark in the video they are also in the Commercial sector aspect of the slide, above Intel and IBM.

https://www.graimatterlabs.ai

 

[Justchilln]

Hi White horse,

great find.

Among other things of interest - re MegaChips: We are giving them a couple of different versions of Akida - so maybe this is the origin of Akida 500, the pared down version of Akida for applications which do not need the full Akida 1000 capabilities?

Ken also suggested we look at MegaChips web page:

https://megachips.com/artificial-intelligence-machine-learning/

View attachment 3946

Brainchip features a unique neuromorphic processor which takes advantage of the sparse data streams associated with most edge applications to provide a very low power solution while maintaining high performance. The small footprint of the Brainchip IP makes it an ideal solution for low power, high volume applications.
Megachips’ engineering team is fully trained on Brainchip software and hardware and is prepared to work directly with our customers to evaluate the performance of the Brainchip Solution in your product.
Once selected MegaChips will work closely with our customers throughout the product development cycle, including licensing the SDK hardware and software, provide NN model consulting, and ensure the final design can be efficiently implemented into silicon.

Key Features​

• Very low power – microwatts to milliwatts
• On chip learning – no need to retrain in the cloud
• Complete development system support
• Silicon-proven IP
• Software Development Kit available

Key Applications​

• Keyword spotting
• Always on detection
• Gesture recognition
• Network packet inspection
• Person/object classification
• Object detection

Then they go on about quadric.

Quadric offers a unique “Vortex Core Architecture” which is unique in its ability to accelerate all steps in the processing of data from a variety of sensors. Typical AI engines only accelerate the Neural Network (NN) processing, while signal conditioning, non-maximal suppression, and other processing steps are normally done by the host processor.
This unique solution provides optimal performance in many demanding applications and offers the flexibility to adapt to new NN processing algorithms as they emerge.

Key Features​

• Scalable architecture to address a wide range of power/performance requirements
• Ability to accelerate all stages of the AI pipeline
• Flexible architecture can adapt to evolving AI models
• Silicon-proven IP
• Software Development Kit available

Key Applications​

• AI inferencing
• Data pre/post processing
• Image processing
• Path Planning

Sounds like we will be getting more license fees on top of royalties from megachips customers when they use akida…

 

[Diogenese]

I'm not sure if it has been previously mentioned, but it seems we have a Brainchip acknowledged competitor named GrAI Matter. Around the 16 minute mark in the video they are also in the Commercial sector aspect of the slide, above Intel and IBM.

https://www.graimatterlabs.ai

Hi Dhm,

We have been tracking GraiMatter for a few years (counting the time at the other place). I pointed to GraiMatter and Hailo as potential competitors

Hi,
as Uiux posted some research on JAST in the seperate thread, I did some research about Bernabe Linares-Barranco, a researcher on neuromorphic computing, and came upon this recently published paper: "2022 roadmap on neuromorphic computing and engineering" https://iopscience.iop.org/article/10.1088/2634-4386/ac4a83
I didn´t read the 170-pages completely - just looked for some keywords. I also like the abstract, which basically underlines the statement by a Ford-Manager (?) that von Neumann-architecture is dead.

First of all the volume of the neuromorphic chip market in above linked article seems to be in line with the prediction in rayzors article:
"If the fundamental technical issues are solved in the next few years, the neuromorphic computing market is projected to rise from $0.2 billion in 2025 to $22 billion in 2035 [7] as neuromorphic computers with ultra-low power consumption and high speed advance and drive demands for neuromorphic devices." (page 4, line 59)

Another passage I found was the following about lidar:
"Given the real-time nature of ADAS computation, and the necessity to process correlated streams of visual and 3D point-cloud data (from lidar systems), there is some expectation that event-based neuromorphic computation may be more suitable than current GPU-type computational hardware. At least one neuromorphic event-based hardware startup is focused on real-time vision processing for
this purpose [7]."(p. 130 ll 27) - I was a little suprised about the source [7]: an article from and about NeuroFlow by GrAI Matter Labs. The authors didn´t have Akida in mind. Has anyone had a look at NeuroFlow? Even if they do not seem to have some Akida-advantages and i cannot see that their chip is in production, they seem to have a shared market with Akida - Lidar, cameras, wearables...

Lastly I was wondering why there is no mention of Brainchip and Akida - TrueNorth and Loihi get mentioned for example here:
"Notably, working neuromorphic olfaction algorithms have been deployed on diverse edge-compatible hardware platforms including Intel Loihi, IBM TrueNorth, field- programmable gate arrays, and custom neuromorphic devices [7, 12, 15, 18, 19]." (page 134 line 7) - I guess that´s because their chips have been distributed to research-institutions and Sean is going to change that for Akida in 2022?!

Greetings from Germany!
Click to expand...

Hi Pepsin,

We have looked at GraiMatter a couple of times in the last couple of years.

In another place, I mentioned that I thought GraiMatter and Hailo were worth keeping an eye on:

31/08/21 20:58 Post #: 55728384

If I were worried about competitors, I'd keep an eye on Hailo and GrAi Matter both of which have silicon NNs, GrAi Matter claiming a SNN.

GrAiMatter: WO2020025680A1 DATA PROCESSING MODULE, DATA PROCESSING SYSTEM AND DATA PROCESSING METHOD
https://worldwide.espacenet.com/patent/search/family/063254640/publication/WO2020025680A1?q=wo2020025680


The GraiMatter NPU:-

HotCrapper

Just a dried up old turd.

hotcrapper.com.au

While GrAiMatter claim to be avoiding the von Neumann bottleneck, I don't think the apple has fallen that far from the tree.

#########################################################################

GraiMatter uses time multiplexing:

WO2020025680A1 DATA PROCESSING MODULE, DATA PROCESSING SYSTEM AND DATA PROCESSING METHOD

The paper you cited points out a problem with time multiplexing:

ShieldSquare Captcha​

iopscience.iop.org
page 88

On-chip learning requires precious memory and routing resources [21], which hinders scalability. On digital technologies, this problem can be sidestepped by time-multiplexing a dedicated local plasticity processor [6, 12]. The time multiplexing approach however suffers from the same caveats as a Von Neumann computer due to the separation between the SNN and the associated plasticity processor. Other promising alternatives for true local plasticity are emerging devices (Sec. 1) and related architectures (Sec. 2.1), which allow storage and computation for plasticity to occur at the same place.

Opinions and research expressed are my own as an unlicensed individual. External links are not endorsed. Do your own research or consult a licensed financial advisor.

Spiking Neural Networks: Research Projects Or Commercial Products?
Opinions differ widely, but in this space that isn’t unusual.
MAY 18TH, 2020 - BY: BRYON MOYER

Spiking Neural Networks: Research Projects or Commercial Products?

Opinions differ widely, but in this space that isn't unusual.

semiengineering.com

 

[Diogenese]

Hi Dhm,

We have been tracking GraiMatter for a few years (counting the time at the other place). I pointed to GraiMatter and Hailo as potential competitors


Hi Pepsin,

We have looked at GraiMatter a couple of times in the last couple of years.

In another place, I mentioned that I thought GraiMatter and Hailo were worth keeping an eye on:

31/08/21 20:58 Post #: 55728384

If I were worried about competitors, I'd keep an eye on Hailo and GrAi Matter both of which have silicon NNs, GrAi Matter claiming a SNN.

GrAiMatter: WO2020025680A1 DATA PROCESSING MODULE, DATA PROCESSING SYSTEM AND DATA PROCESSING METHOD
https://worldwide.espacenet.com/patent/search/family/063254640/publication/WO2020025680A1?q=wo2020025680


The GraiMatter NPU:-

HotCrapper

Just a dried up old turd.

hotcrapper.com.au

While GrAiMatter claim to be avoiding the von Neumann bottleneck, I don't think the apple has fallen that far from the tree.

#########################################################################

GraiMatter uses time multiplexing:

WO2020025680A1 DATA PROCESSING MODULE, DATA PROCESSING SYSTEM AND DATA PROCESSING METHOD

The paper you cited points out a problem with time multiplexing:

ShieldSquare Captcha​

iopscience.iop.org
page 88

On-chip learning requires precious memory and routing resources [21], which hinders scalability. On digital technologies, this problem can be sidestepped by time-multiplexing a dedicated local plasticity processor [6, 12]. The time multiplexing approach however suffers from the same caveats as a Von Neumann computer due to the separation between the SNN and the associated plasticity processor. Other promising alternatives for true local plasticity are emerging devices (Sec. 1) and related architectures (Sec. 2.1), which allow storage and computation for plasticity to occur at the same place.

Opinions and research expressed are my own as an unlicensed individual. External links are not endorsed. Do your own research or consult a licensed financial advisor.

Spiking Neural Networks: Research Projects Or Commercial Products?
Opinions differ widely, but in this space that isn’t unusual.
MAY 18TH, 2020 - BY: BRYON MOYER

Spiking Neural Networks: Research Projects or Commercial Products?

Opinions differ widely, but in this space that isn't unusual.

semiengineering.com

Added to that, in the video @TECH posted today Simon Thorpe disparaged rate coding, which GraiMatter proposes to use.

.

 

[alwaysgreen]

What are everyone's thoughts on this article?

Toyota adopts Tesla's camera-only approach to self-driving development

Toyota subsidiary Woven Planet will utilize a camera-only approach for its self-driving project development, joining Tesla as the only other company to adopt a vision-based strategy to advance itself in the race to autonomous driving. Woven planet said it can use low-cost cameras to collect data...

www.teslarati.com

Toyota adopting Tesla's camera only approach for self driving vehicles rather than Lidar.

 

[Slade]

What are everyone's thoughts on this article?

Toyota adopts Tesla's camera-only approach to self-driving development

Toyota subsidiary Woven Planet will utilize a camera-only approach for its self-driving project development, joining Tesla as the only other company to adopt a vision-based strategy to advance itself in the race to autonomous driving. Woven planet said it can use low-cost cameras to collect data...

www.teslarati.com

Toyota adopting Tesla's camera only approach for self driving vehicles rather than Lidar.

Akida works with both Lidar and Cameras. I’m very comfortable with Toyota’s choice.

 

[Fact Finder]

Wonder if that mining / transport research ties in with the below I posted on HC couple months back?

Anyone up for AEB.

Saw there was a handful of AEB posts back in 2020.

Searching for industrial application info and came across this guy in Perth?

Been on this project last year and half according to profile.

Could be independent but given he started a while ago, could have been given early access or engaged by BRN or someone else local.

Be handy for either road vehicles or even say mine vehicles like autonomous haulage at Rio.

https://www.fmgl.com.au/in-the-news...etes-world-leading-autonomous-haulage-project

https://au.linkedin.com/in/peter-zhou-2a745a1b0

https://s5system.com

View attachment 3940
View attachment 3941
View attachment 3942
View attachment 3943
View attachment 3944

WHY IS NOT THE CROWD ROARING. COME ON WAKE UP 1,000 EYES FMF HAS DONE IT AGAIN.

READ THE WORDS - "SINCE JUNE 2020" "IMPLEMENTING AKIDA FOR AUTONOMOUS EMERGENCY BRAKING"

OUT OF 1,000 EYES ONLY 16 EYES HAVE NOTICED.

STOP SPENDING THE MONEY YOU ARE GOING TO MAKE ON NUMBER PLATES AND TAKE NOTICE OF WHAT FMF HAS POSTED.

My opinion only please DYOR
FF

AKIDA BALLISTA

 

[MDhere]

very interesting .95 large scoop buy of shares on close !

 

[IloveLamp]

WHY IS NOT THE CROWD ROARING. COME ON WAKE UP 1,000 EYES FMF HAS DONE IT AGAIN.

READ THE WORDS - "SINCE JUNE 2020" "IMPLEMENTING AKIDA FOR AUTONOMOUS EMERGENCY BRAKING"

OUT OF 1,000 EYES ONLY 16 EYES HAVE NOTICED.

STOP SPENDING THE MONEY YOU ARE GOING TO MAKE ON NUMBER PLATES AND TAKE NOTICE OF WHAT FMF HAS POSTED.

My opinion only please DYOR
FF

AKIDA BALLISTA

 

[Gemmax]

Yes , some big buys went through!!
One @ $348,492.30

 

[Gemmax]

Approx $1.75M in traded at 4:10pm

 

[Csharmo]

Desperate times calls for desperate measures....
Shows how desperate the whole world his to hold something as special as Brainchip.

'Tip of the iceberg': Taiwan's spy catchers hunt Chinese poachers of chip talent

Taiwan's spy catchers have launched probes into around 100 Chinese companies suspected of illegally poaching semiconductor engineers and other tech talent, a senior official at the island's Investigation Bureau told Reuters.

www.reuters.com