BRN Discussion Ongoing

[Dhm]

I know the Prophesee >> Sony >> Brainchip thing has been proven. I don't know about anyone else but to me, this rates almost as highly as the Mercedes Benz reveal. Here we have a Fortune 500 company revealed as a powerful client of Brainchip. Excuse me, I'm just going to get a ladder to SCREAM IT FROM MY ROOTOP!!!!!!

I couldn't find any relevant footage so I just got my wife to film me.

 

[robsmark]

View attachment 18434

Appreciate it Mia, thanks.

 

[Fact Finder]

Thoughts anyone?

View attachment 18442

Link here

Power-Efficient Edge AI Processing Targets Next-Gen HMI

Sponsored by Texas Instruments: Featuring low-power modes and low active power consumption, the AM62x processor family brings AI-enabled human machine interface to industrial ...

www.electronicdesign.com

Power-Efficient Edge AI Processing Targets Next-Gen HMI

Sponsored by Texas Instruments: Featuring low-power modes and low active power consumption, the AM62x processor family brings AI-enabled human machine interface to industrial ...

www.electronicdesign.com

Leaving aside all the features it does not have the ability to do that AKD1000 does they claim to be low powered and able to run off two pencil lite batteries for 1,000 hours which is about 44 days.

And they think this is low power .

This is what Peter van der Made said about a full size AKD1000 on Stocks Down Under last year confirmed by CEO Sean Hehir since:

“That means that you can run very complex processes off a small battery. Like for an example, order declassification, it can run off a small one and a half volts AA battery for five months.”

One battery 5 months -v- two batteries 44 days and they think this is low power.

They achieve this by basically turning it off whereas AKD1000 because it is event sensitive is running at full throttle 24/7 but just like Walter Brennan with his shotgun rocking on his porch you varmints better not move one single hair on your head or AKIDA1000 will have you.

Pretenders everywhere.

There is only one Tony Lockett and there is only one AKD1000.

My opinion only DYOR
FF

AKIDA BALLISTA

Dedicated to all the Ai pretenders in the semiconductor world:

 

[Diogenese]

There was only one Tony Lockett and now we have one further clue to your identity.

Warwick Capper?

 

[equanimous]

Thoughts anyone?

View attachment 18442

Link here

Power-Efficient Edge AI Processing Targets Next-Gen HMI

Sponsored by Texas Instruments: Featuring low-power modes and low active power consumption, the AM62x processor family brings AI-enabled human machine interface to industrial ...

www.electronicdesign.com

Power-Efficient Edge AI Processing Targets Next-Gen HMI

Sponsored by Texas Instruments: Featuring low-power modes and low active power consumption, the AM62x processor family brings AI-enabled human machine interface to industrial ...

www.electronicdesign.com

My thoughts on Akida and your smart phone will do wonders to save power and efficiency. Your phone is always on low red..

 

[equanimous]

Warwick Capper?

Best mullet in town

 

[Deleted member 118]

My thoughts on Akida and your smart phone will do wonders to save power and efficiency. Your phone is always on low red..

Lol

 

[JDelekto]

Just because AKIDA is being used by Mercedes Benz it does not mean it is being adopted by Qualcomm.

Facilitating or providing a place for Mercedes Benz to plug in their independently acquired or developed technology with Brainchip does not create a partnership between Qualcomm and Brainchip.

I agree with this, but there is something to consider. It does not necessarily mean Qualcomm is replacing BrainChip's role in Mercedes' plans.

If you note in the blurb, their MBUX system currently uses NVIDIA's technology. If anything, I think they are looking to replace the NVIDIA processors with higher power requirements with the Snapdragon processors for video and compute processing.

I imagine the ARM-based Snapdragon processors would be in line with the same family as those used to power Meta Quest's VR products.

One thing to note is that the Snapdragon XR2 processor (used in the Meta Quest 2 HMD) touts AI capabilities, such as voice activation and context awareness (to detect certain background noises). I think it still lacks the learning capabilities Akida would make a much more personalized experience.

I also think Akida may still have a role in the component of the MBUX system used for personalization and security. However, I am optimistically speculating here.

If anything, I would like to see Qualcomm's next XR generation incorporate BrainChip's IP because I think it would bolster its AI capabilities and align with the goal of low-power consumption the Qualcomm processors strive to achieve.

 

[Terroni2105]

Looking forward to this presentation from Anil. BRN are amongst some big names here

Linley Fall Processor Conference (powered by TechInsights) - Day One | TechInsights

Presentations will address processors and IP cores for AI applications, embedded, data-center, automotive, and server designs. In-person attendees will be able to hear presentations and interact with the speakers during Q&A, lunch, and the networking reception.

www.techinsights.com

I notice they have the old company design / logo

edit. I have sent an email off to Integrous Communciations, hope they are the relevant person to rectify it

 

[robsmark]

Does anyone have a link to that table which shows a comparison of all the big boys chips and Akida? Pretty sure it was in an AGM but can’t find it..

Edit: Found it - 2021 AGM.

 

[Baisyet]

Leaving aside all the features it does not have the ability to do that AKD1000 does they claim to be low powered and able to run off two pencil lite batteries for 1,000 hours which is about 44 days.

And they think this is low power .

This is what Peter van der Made said about a full size AKD1000 on Stocks Down Under last year confirmed by CEO Sean Hehir since:

“That means that you can run very complex processes off a small battery. Like for an example, order declassification, it can run off a small one and a half volts AA battery for five months.”

One battery 5 months -v- two batteries 44 days and they think this is low power.

They achieve this by basically turning it off whereas AKD1000 because it is event sensitive is running at full throttle 24/7 but just like Walter Brennan with his shotgun rocking on his porch you varmints better not move one single hair on your head or AKIDA1000 will have you.

Pretenders everywhere.

There is only one Tony Lockett and there is only one AKD1000.

My opinion only DYOR
FF

AKIDA BALLISTA

Dedicated to all the Ai pretenders in the semiconductor world:

Hi FF you mean Akida 1.0

 

[Deleted member 118]

If this ain’t proving Moschip and Brainchip are working together then nothing will. Sod those NDAs

https://media.sumo.com/99344e074741688f6d6dab85ced4997ce3335106ec10c54b899442f3f2fd0f5a

 

[Baisyet]

They were very excited about Brainchip and since last time they presentated together we have not heard anything yet. Interesting

 

[Deleted member 118]

https://www.ipsos.com/sites/default/files/ct/publication/documents/2021-12/Ipsos_Strategy3_Whitepaper_2020_-_China%27s_Tech-driven_ICV_for_the_Next_Decade.pdf

 

[Baisyet]

If this ain’t proving Moschip and Brainchip ain’t working together then nothing will.


View attachment 18447

View attachment 18449

https://media.sumo.com/99344e074741688f6d6dab85ced4997ce3335106ec10c54b899442f3f2fd0f5a

They were very excited about Brainchip and since last time they presentated together we have not heard anything yet. Interesting

Some big name attending and speaking but it says confirmation awaiting..

IESA Vision Summit 2022

iesavisionsummit.com

 

[TECH]

Hi Tech,

To state the obvious, Akida is no ordinary product.

I'm still hoping for the hockey stick.

There are external forces which tend to expedite the adoption of Akida, the main one being ADAS/AD.

Governments are mandating the end of ICEs, so there is pressure to adopt EVs.

Europe is introducing legislation requiring driver monitoring.

The sensors for ADAS/AD generate huge amounts of data which require real time processing.

Then there's defence spending ...

Several countries are setting up AI research organizations, but she'll be right.

The counter arguments include the fact that BrainChip has needed to get the product to the attention of the OEMs who need to encourage their customers to adopt what to many may appear to be unproven, if not downright impossible tech. After all "what is a neuromorphic computer anyway?"

The dam is about to spill.

Hey Dio,

18 months was just an example of the case in hand, but you are correct, many early adaptors have been testing and tossing ideas around for 2 plus years, in some cases longer, but "it can take up to 12 months" to secure a sale, then all the other factors come into play, concept design to marketable product.

Each individual client will obviously have different lead in times, but 18 months is realistic going by the journey we've been on to date, take Renesas for example, 21 months and counting.

I agree with what you are implying, I'm just trying to keep the natives at bay, but we are due for an "explosion in sales"

ADAS/AD were always going to be the longest path to revenue, time wise, from what Lou used to tell us.

News is coming for sure.

Thanks for your outstanding contributions to this forum by the way, your input is excellent, sometimes I wonder if you are actually the man himself.... wink wink

Cheers for now.... Tech

 

[Deleted member 118]

Anything here worth getting excited about?

https://docs.xilinx.com/api/khub/documents/Xz0szg2HiN1YFYfaJVXcrQ/content?Ft-Calling-App=ft%2Fturnkey-portal&Ft-Calling-App-Version=4.0.14&filename=wp518-ai-edge-intro.pdf&download=true

 

[Deleted member 118]

Tata Elxsi - Democratizing Artificial Intelligence

Tata Elxsi provide Machine Learning & Advanced Analytics, Big Data Engineering, AI Tools & Frameworks, Cognitive Video Services, Conversational AI, TESA - Smart Annotation

www.tataelxsi.com

 

[Deleted member 118]

https://documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/collateral/white-paper/white-paper-how-5G-AI-and-IoT-are-shaping-future-of-storage.pdf

 

[Diogenese]

Hey Dio,

18 months was just an example of the case in hand, but you are correct, many early adaptors have been testing and tossing ideas around for 2 plus years, in some cases longer, but "it can take up to 12 months" to secure a sale, then all the other factors come into play, concept design to marketable product.

Each individual client will obviously have different lead in times, but 18 months is realistic going by the journey we've been on to date, take Renesas for example, 21 months and counting.

I agree with what you are implying, I'm just trying to keep the natives at bay, but we are due for an "explosion in sales"

ADAS/AD were always going to be the longest path to revenue, time wise, from what Lou used to tell us.

News is coming for sure.

Thanks for your outstanding contributions to this forum by the way, your input is excellent, sometimes I wonder if you are actually the man himself.... wink wink

Cheers for now.... Tech

Yes. Renesas is a funny one because they have their own DRP-AI which they prefer for the heavy duty jobs. I think that there is a bit of "not invented here" syndrome at play.

I've said before that I think that whatever application they have for 2 nodes of Akida isn't going to be a make-or-break product when you consider their whole catalogue, so there's no rush as far as they are concerned, despite the fact that they have all the wherewithal to get to production faster than most. Still, they have paid the up front licence fee ...

I think that other companies with a smaller catalogue and needing the services of designers and manufacturers the likes of Socionext, TSMC or MegaChips would be more anxious to get to market asap once they pay the upfront fees, but they are dependent on the timetables of the chip manufacturers.