... and then you need to deal with Ella ...
... and then you need to deal with Ella ...
If the Helium vector extensions sole purpose is to extend the ARM instruction set improving CPU performance in more complex use cases then as you say the Helium technology is not an accelerator per say, but rather it is directly CPU enhancing. Therefore imo Akida IP could still be a coprocessor to any Cortex M IP containing Helium.To shine a light on ARM Helium again:
It's a vector extension which extends the instruction-set architecture. That leads to faster and more complex processes. It enables a run without additional accelerator, delivering better results than earlier generations and keep the costs down.
This is not the competitor. Of course you can reach even better results with the use of additional accelerators, so it's more a question of AND not OR.
Also the full commercialisation of Helium now gives a good idea about industry timelines, as the extension is in the pipeline since 2019.
In first place each company in the world will favour their inhouse solutions, all are in it for maximum profit, just a natural thing. Secondly stands the aquisition of new technology to the own pipeline. The history of the big tech players is full of it.
The least attractive solution is the use of IP from a start-up.
This is the world Brainchip is living in. For a breakthrough it needs a full commercialised product with Akida inside, which shows real advances to the competition. This could be the initialization of a snowball effect with a big wave starting.
So the remaining question is, who will be the first using not much loved IP as a standard in a commercial product, not as an option.
If the Helium vector extensions sole purpose is to extend the ARM instruction set improving CPU performance in more complex use cases then as you say the Helium technology is not an accelerator per say, but rather it is directly CPU enhancing ..and not an inferior alternative to Akida IP - therefore potentially Akida IP could now/in future be the AI component in Cortex M chip designs.
Here is the toy robot SynSense unveiled in April - unlike the robot by Kosmos, it does not need an app on a smartphone or tablet to work.Don’t forget SynSense CEO stated they were being used in a toy because being wrong occasionally is acceptable for toys unlike motor vehicles read in BMW probably a fail.
I have not read that SynSense have revealed what the toy is so you need to eliminate SynSense before you can have any confidence this is Brainchip.
My opinion only DYOR
FF
AKIDA BALLISTA
What do you think?
could it be Akida IP?
There is a new toy for kids aged between 10 -16 jears old in germany on market :
https ://www.kosmos.de/de/miika-k.i._1620899_4002051620899
especially the reading :
"To use the app, all you need is a smartphone or tablet. In order to make Miika AI a safe toy for children between 10 and 14 years old, the app works purely locally without the Internet, it does not save any image or audio files and does not send any data to a server. A real innovation that guarantees data protection!"
....makes me think of that possibility.
I like it as a present for christmas maybe.
@Fact Finder
Thank you for your contribution.
VVDN make Edge Boxes for Nvidia and Qualcomm. The Nvidia box uses Jetson. I believe it is the TX2.
As far as a competition with Nvidia’s Edge Box goes built on the Jetson TX2 the winner has already been announced by Tata Consulting Services and Brainchip:
1. TCS has published that “We design and train a CNN and convert it into SNN using the CNN2SNN conversion toolkit of Brainchip Akida neuromorphic platform. We achieve 95.46% accuracy while power consumption and latency are at least 35x and 3.4x more efficient respectively in stage-1 (and 230x & 7x in stage-2) compared to the equivalent CNN running on Jetson TX2.”
2. Brainchip has stated it will be a smaller slimlined Edge Box than the existing offerings from VVDN. So wins on size unless you believe size does matter.
3. AKD1000 has been priced at $AU25.00 on multiple occasions by Anil Mankar, Peter van der Made and Mr. Dinardo. Simon Thorpe formerly on the Brainchip Scientific Advisory Board has a number of papers where he suggested a price of about $10.00. Suffice to say it would be under $30.00.
4. Nvidia Jetson TX2 is significantly more expensive:
View attachment 50331
5. AKD1000 can be scaled to 64 chips so even if you added a couple of extra to up the processing power it will fall well short of $AU 100.00.
6. Assuming that the metal box and accessories required to build the Boxes will be comparable and the additional costs will be approximately the same.
On any measure therefore an AKIDA VVDN EDGE Box has already won the contest with Nvidia.
My opinion only DYOR
Fact Finder
AKIDA BALLISTA
Here is the toy robot SynSense unveiled in April - unlike the robot by Kosmos, it does not need an app on a smartphone or tablet to work.
![]()
Neuromorphic Robot Redefining Interaction | SynSense
SynSense presents the world’s first neuromorphic programmable robot with dynamic vision to enable strong human-machine interaction.www.synsense.ai
SynSense presents the world’s first neuromorphic programmable robot with dynamic vision to enable strong human-machine interaction
2023-04-13
By SynSense
[printfriendly]
This month, SynSense, the world-leading commercial supplier of ultra-low-power sensory processing hardware and application solutions, unveiled the world’s first neuromorphic programmable robot at the 22nd China Shantou (Chenghai) International Toy Fair, alongside QunYu, the leading Chinese company for programmable intelligent toys. The robot, integrating SynSense’s dynamic vision SoC “Speck™”,is capable of human body recognition, visual perception, and mimicry, expanding the forms of human-robot interaction.
Dr. Yannan Xing, senior algorithm application engineer at SynSense, explained that the robot’s sensing ability is achieved through the Speck™ chip embedded in the robot: “With our chip, the robot has the ability to ‘see’ and ‘learn’, allowing it to ‘mimic’ human movements. By waving your arms, the robot can learn your movements and wave its arms in response. When you cross your arms in front of your chest, the robot senses this and also crosses its arms.”
![]()
The neuromorphic vision solution is lightweight, low-power, low-latency, privacy-protecting, and low-cost, making it ideal for toy robots. Battery-powered and without the need for external devices such as smartphones or cameras, the robot can recognize eight different human body postures. The feature has flexible scalability and can recognize additional postures and perform additional visual functions through software updates.
Integrated vision processing and fluid human-robot interaction typically requires lots of energy and computational power. Speck™ overcomes these and enables a new class of toy that is both low-power and intelligent.
Speck™’s low cost enables large-scale deployment of vision-based human interaction applications to consumer devices. Speck™’s privacy-first design, and independence from cloud transmission or processing, meet the increasing demand for privacy and security in consumer electronics products.
View attachment 50388
Ruifeng Chen, CEO of QunYu said, “Neuromorphic intelligence is a cutting-edge AI technology, and we are very proud to launch the world’s first neuromorphic toy and provide a new form of human-robot interaction.” SynSense’s Senior Business Director Hongmin Chen said, “SynSense and QunYu have signed a strategic cooperation agreement and will jointly launch neuromorphic interactive blocks.”
I had a look at this little fella. In Germany, the robot is marketed as Miika K.I. (KI= Künstliche Intelligenz/Artificial Intelligence), internationally (by Thames & Kosmos) as KAI (= Kosmos Artificial Intelligence).
I watched part of the company’s step-by-step assembly video in the hope of finding out more about its electronic innards, but couldn’t make out what was written on the chip, as it was too blurry. But luckily I then found this picture in one of the German Amazon reviews, which shows that the robot uses an ESP32 MCU by Espressif, a Chinese company. It is hard to decipher the exact model, but it looks like it is part of the WROOM series.
ESP32 Wi-Fi & Bluetooth Modules I Espressif
ESP32 series of modules are powerful Wi-Fi+Bluetooth/Bluetooth LE modules that target a wide variety of AIoT applications, ranging from low-power sensor networks to the most demanding tasks.www.espressif.com
View attachment 50384
While Edge Impulse and Espressif are a match…
![]()
Adding Vision to ESP32 with Edge Impulse | Espressif Systems
You only need a few lines of additional code to run powerful ML algorithms on ESP32, especially if you also use the Arduino framework and Edge Impulse.www.espressif.com
…. and Plumerai, for example, can run their people detection on an ESP32-S3 MCU:
https://www.edge-ai-vision.com/2023...ection-ai-now-runs-on-espressif-esp32-s3-mcu/
… the Kosmosneeds an app run on a smartphone or tablet to work, unlike the neuromorphic SynSense robot toy.
Also, Miika K.I. is actually not that new - on January 31, it was awarded the ToyAward 2023 in the category ‘Teenagers & Adults’ during the Nürnberger Spielwarenmesse. So that alone would make Akida being involved highly unlikely, as SynSense would then not have been able to claim to have launched the world’s first neuromorphic programmable robot in April 2023.
![]()
Das sind die „ToyAward“-Gewinner 2023
Die Entscheidung ist gefallen: Feierlich eingebunden in die PressPreview wurden heute Mittag die Sieger der begehrten Awards in erstmals sechs Kategorien ...dasspielzeug.de
VVDN-EDBX-JXN Edge AI Box Powered by NVIDIA® Jetson Xavier NX™ Module
VVDN-EDBX-JXN Edge AI Box includes a power-efficient, compact NVIDIA® Jetson Xavier NX™ Module and a custom carrier board developed by VVDN for edge applicationswww.vvdntech.com
![]()
NVIDIA Jetson Xavier NX for Embedded & Edge Systems
The World’s Smallest AI Supercomputer & Supports all AI Frameworks.www.nvidia.com
GPU | 256-core NVIDIA Pascal™ GPU architecture with 256 NVIDIA CUDA cores |
CPU | Dual-Core NVIDIA Denver 2 64-Bit CPU Quad-Core ARM® Cortex®-A57 MPCore |
Memory | 8GB 128-bit LPDDR4 Memory 1866 MHx - 59.7 GB/s |
Storage | 32GB eMMC 5.1 |
Power | 7.5W / 15W |
Gern geschehen!Thank you for your time and effort in researching my query... it would have been too nice.
In the video is explained that all AI is made on the smartphone offline and there is also no voice recognition possible.
But I don't think it's a bad alternative to exploring AI as a gift for a child of the right age and interest.
Great research!
Good Morning Boab ,
Yes , With any luck one of the bot programers overseeing the manipulation of BRN got hit by a buss on the way into work today.
We can only hope.
Three day chart ... volume pattern looking promising for another good surge today ... any time now.
.......Wen Gulfstream........
Regards,
Esq.
Thanksgiving holiday in USA so with no lead for traders to follow and exploit and no price sensitive news popping likely just to continue with current trend but with withering momentum.Trading volume is extremely low? @Esq.111 ???
Well called, lets hope it keeps that direction for the rest of the day.Good Morning Boab ,
Yes , With any luck one of the bot programers overseeing the manipulation of BRN got hit by a buss on the way into work today.
We can only hope.
Three day chart ... volume pattern looking promising for another good surge today ... any time now.
.......Wen Gulfstream...............G700 - Gulfstream Aerospace
Regards,
Esq.
Sheesh Esq. orchids to B yesterday followed by compliments today. Both considerations understandable of course. Is there something you want to tell us perhaps?Morning Bravo & Fellow Chippers ,
Nice find on the above , interesting ...on a side note... Yesterday there were NO SHORTS TAKEN OUT ON BRN, according to the ASX site.
Regards,
Esq.