BRN Discussion Ongoing

[Boab]

Ambiq Apollo510 Arm Cortex-M55 MCU delivers up to 30x better power efficiency for AI/ML workloads - CNX Software

Ambiq Apollo510 Arm Cortex-M55 microcontroller delivers 30 times better power efficiency than typical Cortex-M4 designs and 10 times the performance of

www.cnx-software.com

More details if someone with better technical understanding can help us.

 

[Stable Genius]

Hyundai autonomous car passed a driving test in New York!

4 min video featuring a robotaxi. Starts bringing a blind lady into the story to humanise it and help show value.




Enjoy

 

[Rach2512]

From Dec 2022, sorry if already posted.

Evaluation of neuromorphic AI with embedded Spiking Neural Networks – LEAT

This internship project aims at confirming this assumption by converting classical CNN to SNN from standard Machine Learning frameworks (Keras) and deploy the resulting neural models onto the Akida neuromorphic processor from BrainChip company [4]. The results obtained in terms of accuracy, latency and energy will be compared to other existing embedded solutions for Edge AI [2].

Project mission

The project mission will be organized in several periods:

Bibliographic study on spiking neural network trainingIntroduction to the existing Sw framework from BrainChipTraining of convolutional neural networks for embedded applications [1] and conversion from CNN to SNN from KerasDeployment of the SNN onto Akida processing platformExperimentations and measurementsPublication in an international conference.


Below is link to their publications.

Toutes Publications LEAT

 

[Tothemoon24]

AI workloads are rapidly flowing out to the network edge, accelerating the demand for even more compute power that established chipmakers and startups alike are rushing to meet.

Silicon vendors over the past month have rolled out new CPUs, accelerators, and platforms designed to address the power, analytical processes, and security demands in a compute environment that by its nature is constrained in many of those areas. That said, the overriding need is to bring all these capabilities as close as possible to where the massive amounts of data are being generated, so for chipmakers, the race is on.

The result will be a range of new silicon options for AI workloads at the edge as the year unfolds. The same day this week that Arm unveiled its Ethos-U85 neural processing unit(NPU) and Corstone-320 Internet of Things (IoT) reference design platform for edge AI applications at the Embedded World 2024 show, rival Intel showed off its Gaudi 3 acceleratorsand Xeon 6 CPUs for the AI edge at the company’s Intel Vision 2024 event, with CEO Pat Gelsinger calling AI “the next killer app for the edge.”

“As the edge becomes increasingly important, it’s going to become the dominant AI workload resource,” Gelsinger said. “Research indicates that by 2026, 50% of edge computingdeployments will involve machine learning and AI, compared to just 5% today. A killer use case.”

For its part, Qualcomm at Embedded World announced its RB3 Gen 2 Platform, a hardware and software offering aimed at embedded and IoT uses, including AI edge boxes. All this comes less than a month after NVIDIAtook the covers off its upcoming Blackwell family of GPUs, maintaining its pole position for AI workload accelerators.

Startups Are in the Mix​

It’s not only the established players but also startups looking to get a foothold in the edge AI space. Israeli startup Hailo this month raised $120 million in Series C funding — bringing to more than $340 million that total amount it’s brought in — and launched the Hailo-10 accelerators, designed to run generative AI applications locally without the need for cloud-based services. Another startup, SiMa.ai, which makes Systems on a Chip (SoCs) of edge AI systems, this month pulled in $70 million in funding — including from Dell Technologies’ investment arm — growing the amount it’s collected to $270 million.

It will be a boon for chip makers, with Omdia researchers expecting the market for AI processors at the edge to grow from $31 billion in 2022 to $60 billion in 2028.

 

[TECH]

View attachment 60826


AI workloads are rapidly flowing out to the network edge, accelerating the demand for even more compute power that established chipmakers and startups alike are rushing to meet.

Silicon vendors over the past month have rolled out new CPUs, accelerators, and platforms designed to address the power, analytical processes, and security demands in a compute environment that by its nature is constrained in many of those areas. That said, the overriding need is to bring all these capabilities as close as possible to where the massive amounts of data are being generated, so for chipmakers, the race is on.

The result will be a range of new silicon options for AI workloads at the edge as the year unfolds. The same day this week that Arm unveiled its Ethos-U85 neural processing unit(NPU) and Corstone-320 Internet of Things (IoT) reference design platform for edge AI applications at the Embedded World 2024 show, rival Intel showed off its Gaudi 3 acceleratorsand Xeon 6 CPUs for the AI edge at the company’s Intel Vision 2024 event, with CEO Pat Gelsinger calling AI “the next killer app for the edge.”

“As the edge becomes increasingly important, it’s going to become the dominant AI workload resource,” Gelsinger said. “Research indicates that by 2026, 50% of edge computingdeployments will involve machine learning and AI, compared to just 5% today. A killer use case.”

For its part, Qualcomm at Embedded World announced its RB3 Gen 2 Platform, a hardware and software offering aimed at embedded and IoT uses, including AI edge boxes. All this comes less than a month after NVIDIAtook the covers off its upcoming Blackwell family of GPUs, maintaining its pole position for AI workload accelerators.

Startups Are in the Mix​

It’s not only the established players but also startups looking to get a foothold in the edge AI space. Israeli startup Hailo this month raised $120 million in Series C funding — bringing to more than $340 million that total amount it’s brought in — and launched the Hailo-10 accelerators, designed to run generative AI applications locally without the need for cloud-based services. Another startup, SiMa.ai, which makes Systems on a Chip (SoCs) of edge AI systems, this month pulled in $70 million in funding — including from Dell Technologies’ investment arm — growing the amount it’s collected to $270 million.

It will be a boon for chip makers, with Omdia researchers expecting the market for AI processors at the edge to grow from $31 billion in 2022 to $60 billion in 2028.

Great post ! Once again Brainchip is quietly sitting in No. one pole position, patiently waiting for the light to turn green, surrounded by all our 60 odd customers/clients and partners, 99% secured with breath-taking patents worldwide...oh the timing and digital as well !!!

AKD II can't wait to show off

 

[mrgds]

Hyundai autonomous car passed a driving test in New York!

4 min video featuring a robotaxi. Starts bringing a blind lady into the story to humanise it and help show value.




Enjoy

Thnx @Stable Genius , definately a "moving video" ( pun intended )
But seriously, if the driving test was held in any other city outside of the cars "geo-fenced area " then i feel the examiners tick would turn to a cross.
This technology ie; lidar, radar, ultrasonic sensors are not scalable due to the prohibitive costs involved.

This is where i believe Tesla"s approach is far more favourable for scalability and wide spread adoption.
We humans learn to drive a motor vehicle with just 2 eyes and a biological brain, so why can"t a vehicle with 8 eyes ( cameras ) and a Ai neural net brain be equally efficent?

Autonomy of transport is inevitable imo, maybe 1-2yrs away,
Check this out;

tesla fsd 12.3.3 video ai driver update free - Yahoo Video Search results

The search engine that helps you find exactly what you're looking for. Find the most relevant information, video, images, and answers from all across the Web.

au.video.search.yahoo.com

Tesla also doing "multi modality " now with their super computer Grok.
Please let Akida be part of their solution .....................

 

[Food4 thought]

Great post ! Once again Brainchip is quietly sitting in No. one pole position, patiently waiting for the light to turn green, surrounded by all our 60 odd customers/clients and partners, 99% secured with breath-taking patents worldwide...oh the timing and digital as well !!!

AKD II can't wait to show off

Tech
How and where do you get the information that our beloved company Brainchip is Number One???
I am a massive fan and investor but I haven’t see any evidence that we are the number one

 

[Smoothsailing]

Go Brainchip

 

[TECH]

Tech
How and where do you get the information that our beloved company Brainchip is Number One???
I am a massive fan and investor but I haven’t see any evidence that we are the number one

Fair enough question.

First off, nothing on this site is fact, unless it comes directly from within the company, so that leaves us all with an opinion.

My question to you would be, please name another company that you believe deserves No. 1 on the grid ? and I'm referring
to a company that basically has 3 NSoC that all have "proof of concept', are commercially available now, specializing in Spiking
Neural Networks and have On-Chip Learning, currently protected by a solid patent.

Until such time as another entity raises their head and says...."look at me...look at me" I shall continue to believe that my personal
opinion, based on what I know to be true is a fair and reasonable inference.

Kind regards...Tech

 

[supersonic001]

#5g #iot | Microchip Technology Inc.

Join us for our new livestream series, "Designs for the Edge," airing weekly from May 8, 2024 to June 12, 2024. Don't miss out on key insights into #5G, #IoT and more. Learn more and subscribe now: https://mchp.us/3vPUysx.

www.linkedin.com

 

[Diogenese]

Fair enough question.

First off, nothing on this site is fact, unless it comes directly from within the company, so that leaves us all with an opinion.

My question to you would be, please name another company that you believe deserves No. 1 on the grid ? and I'm referring
to a company that basically has 3 NSoC that all have "proof of concept', are commercially available now, specializing in Spiking
Neural Networks and have On-Chip Learning, currently protected by a solid patent.

Until such time as another entity raises their head and says...."look at me...look at me" I shall continue to believe that my personal
opinion, based on what I know to be true is a fair and reasonable inference.

Kind regards...Tech

Hi Tech,

I think your reasoning holds water.

It should be part of a BRN investor's continuing DD to maintain an ongoing evaluation of the competition's comparative advantages, and, so far, the thousand eyes have not blinked.

Some of the pretenders have gained a commercial toehold due to the commercial availability of physical chips, but. at any level of implementation, a degree of technical expertise is required, whether it be;

installing a PCIe board in a PC, (hobbyist);

incorporating a SNN chip into a PCB, (system designer);

incorporating Akida IP into an IC SoC, (IC designer).

Even the edge box will require some degree of competence above plain PC user capability.

And the rapid development of AI technology means that what is developed today was obsolete yesterday.

 

[Bravo]

I find the below extract from an article dated 6 Nov 2023 very interesting. View attachment 56338

https://www.geekwire.com/2023/openai-ceo-sam-altman-on-possibility-of-chatgpt-device-wed-like-to-figure-out-something-amazing/

We all know OpenAI and Mercedes are collaborating.

And remember that the Mercedes EQXX featured AKIDA-powered neuromorphic AI voice control technology which Mercedes engineers said was 5 to 10 times more efficient than conventional voice control.

Sooo...I wonder....




View attachment 56337

Is OpenAI about to take on Alexa and Siri? ChatGPT maker files trademark for Voice Engine​

News
By Ryan Morrison
published 27 March 2024

Trademark covers voice assistants

(Image credit: Shutterstock)

OpenAI may have Apple, Amazon, and Google in its sights for its next big artificial intelligence push, taking on the voice assistant market with a new Voice Engine tool.

While ChatGPT does have a voice-friendly interface on mobile — and recently introduced a way to have it speak its responses on desktop — a new trademark application from OpenAI for the words Voice Engine relates specifically to building digital voice assistants.

It is now possible to swap out the default voice assistant on Android. Apple seems to be in talks with a range of AI companies over the future of artificial intelligence on the iPhone, so this could be a preemptive move from OpenAI building on a potential new market.


Apple is also rumored to be opening a dedicated AI App Store with the next major upgrade to iOS, which would create a new market for AI-powered assistants.

Sam Altman, OpenAI CEO said there are "many different things" being released this year. While it is expected this will include Sora, the AI video tool it could include a new AI voice system.

What do we know about Voice Engine​

(Image credit: Getty)
We don’t know much about Voice Engine or whether it will even be a product. OpenAI hasn’t commented publicly on it, so all we have is rumor and the trademark filing.
While Voice Engine could be a new model built specifically for speech applications, it is also likely this is part of an enterprise play for OpenAI. It could be building a high-quality speech system that would let companies build out more efficient call center bots.

Sign up to get the BEST of Tom’s Guide direct to your inbox.​

Upgrade your life with a daily dose of the biggest tech news, lifestyle hacks and our curated analysis. Be the first to know about cutting-edge gadgets and the hottest deals.
Contact me with news and offers from other Future brandsReceive email from us on behalf of our trusted partners or sponsorsBy submitting your information you agree to the Terms & Conditions and Privacy Policy and are aged 16 or over.

It sounds a lot like all the pieces you'd need for a fully functional, fully interactive AI voice assistant that can not only handle complex tasks but chat naturally and even take phone calls on your behalf.

The new trademark application was filed with the U.S. Patent and Trademark Office last week. While an application doesn't necessarily mean it will result in a product, this does line up with the wider market shifting more to voice and OpenAI's direction to targeted models.
The filing covers the creation of software used for building digital voice assistants, audio generation from text prompts, voice command processing, and voice service delivery.
The full application covers the development of voice service delivery, using AI for text or-voice and text-to-audio, natural language, and speech processing, generating audio and voice from a prompt (text, speech, visual, image), processing voice commands, speech recognition, and building digital voice assistants.
That sounds a lot like all the pieces you'd need for a fully functional, fully interactive AI voice assistant that can handle complex tasks, chat naturally, and even take phone calls on your behalf.

Where does GPT-5 fit in this?​

(Image credit: OpenAI)
OpenAI released GPT-4 a year ago. At the time this was a groundbreaking generative AI model that powers ChatGPT and Microsoft Copilot.
The company also started training GPT-5 late last year, resulting in speculation over its release date. Altman told podcaster Lex Fridman, "We will release an amazing new model this year," but wouldn't confirm whether this was GPT-5 or some precursor.
He also said there would be "many different things" released over the coming months. According to OpenAI CTO Mira Murati, this will include the AI video platform Sora.
There is some speculation on social media that Sora and this new Voice Engine are different modal interfaces for GPT-5.
It is very likely that GPT-5 will be a true multimodal model, able to understand video, images, speech, text, and code — as well as generate all those content types.

Voice Engine could be a new Assistant​

Given the trademark's description, it is also possible that Voice Engine could be a new voice assistant, merging Siri, Alexa, or Google Assistant's wider capabilities with ChatGPT's reasoning and natural language capabilities.
Google has already started upgrading Gemini to work in that way, Apple is rumored to be building a new version of Siri with large language model functionality, and Amazon is already testing Alexa Plus with similar underlying skills.
OpenAI may offer Voice Engine to power such systems in the future or as an alternative interface to ChatGPT that can run on smart speakers, phones, or even headphones.
Or it could just be OpenAI playing it cautious with trademarks. It had a bid to protect GPT rejected, so it now has filed trademark applications for GPT-5, 6, and even GPT-7. The latter includes music generation, converting text and data to code, and writing code from scratch.

 

[Bravo]

AI hardware company from Jony Ive, Sam Altman seeks $1 billion in funding​

A venture fund founded by Laurene Powell Jobs could finance the company.​

SAMUEL AXON - Tuesday at undefined

Jony Ive, the former Apple designer.
Gary Cohen
131
Former Apple design lead Jony Ive and current OpenAI CEO Sam Altman are seeking funding for a new company that will produce an "artificial intelligence-powered personal device," according to The Information's sources, who are said to be familiar with the plans.
The exact nature of the device is unknown, but it will not look anything like a smartphone, according to the sources. We first heard tell of this venture in the fall of 2023, but The Information's story reveals that talks are moving forward to get the company off the ground.


Ive and Altman hope to raise at least $1 billion for the new company. The complete list of potential funding sources they've spoken with is unknown, but The Information's sources say they are in talks with frequent OpenAI investor Thrive Capital as well as Emerson Collective, a venture capital firm founded by Laurene Powell Jobs.
Advertisement

FURTHER READING​

OpenAI CEO Altman wasn’t fired because of scary new tech, just internal politics
SoftBank CEO and super-investor Masayoshi Son is also said to have spoken with Altman and Ive about the venture. Financial Times previously reported that Son wanted Arm (another company he has backed) to be involved in the project.
Obviously, those are some of the well-established and famous names within today's tech industry. Personal connections may play a role; for example, Jobs is said to have a friendship with both Ive and Altman. That might be critical because the pedigree involved could scare off smaller investors since the big names could drive up the initial cost of investment.
Although we don't know anything about the device yet, it would likely put Ive in direct competition with his former employer, Apple. It has been reported elsewhere that Apple is working on bringing powerful new AI features to iOS 18 and later versions of the software for iPhones, iPads, and the company's other devices.
Altman already has his hands in several other AI ventures besides OpenAI. The Information reports that there is no indication yet that OpenAI would be directly involved in the new hardware company.

 

[7für7]

Maybe I’m wrong.. but I guess we will need a deeper breath for a nice rising after this what happens In Middle East. I don’t expect any Green from tomorrow! Not because of brainchip. Hope I’m wrong

 

[Food4 thought]

Fair enough question.

First off, nothing on this site is fact, unless it comes directly from within the company, so that leaves us all with an opinion.

My question to you would be, please name another company that you believe deserves No. 1 on the grid ? and I'm referring
to a company that basically has 3 NSoC that all have "proof of concept', are commercially available now, specializing in Spiking
Neural Networks and have On-Chip Learning, currently protected by a solid patent.

Until such time as another entity raises their head and says...."look at me...look at me" I shall continue to believe that my personal
opinion, based on what I know to be true is a fair and reasonable inference.

Kind regards...Tech

I personally am hoping that we are all thinking the same way and that we are right in the thinking. ( otherwise I would not have invested so much of my hard earned and time)
I am excited every day to wake up and hope that this day will be the day the penny drops and that something positive in the way of a ASX Announcement is made, it will happen some day soon for sure.
Have a safe positive journey

My heart goes out to the people who unfortunately passed away in the Bondi Westfield tragedy on Saturday afternoon.
Such a sad time for so many people.

 

[manny100]

Fair enough question.

First off, nothing on this site is fact, unless it comes directly from within the company, so that leaves us all with an opinion.

My question to you would be, please name another company that you believe deserves No. 1 on the grid ? and I'm referring
to a company that basically has 3 NSoC that all have "proof of concept', are commercially available now, specializing in Spiking
Neural Networks and have On-Chip Learning, currently protected by a solid patent.

Until such time as another entity raises their head and says...."look at me...look at me" I shall continue to believe that my personal
opinion, based on what I know to be true is a fair and reasonable inference.

Kind regards...Tech

Yep, and we are the only one that goes avoids the trip to tge cloud making for better latency, privacy/security plus no cloud fees.
According to Nandan (recently) as far as he is aware we are the only company that does 'on device learning'. This saves cost and time. The others do training and palm it off as learning.

 

[Diogenese]

Thnx @Stable Genius , definately a "moving video" ( pun intended )
But seriously, if the driving test was held in any other city outside of the cars "geo-fenced area " then i feel the examiners tick would turn to a cross.
This technology ie; lidar, radar, ultrasonic sensors are not scalable due to the prohibitive costs involved.

This is where i believe Tesla"s approach is far more favourable for scalability and wide spread adoption.
We humans learn to drive a motor vehicle with just 2 eyes and a biological brain, so why can"t a vehicle with 8 eyes ( cameras ) and a Ai neural net brain be equally efficent?

Autonomy of transport is inevitable imo, maybe 1-2yrs away,
Check this out;

tesla fsd 12.3.3 video ai driver update free - Yahoo Video Search results

The search engine that helps you find exactly what you're looking for. Find the most relevant information, video, images, and answers from all across the Web.

au.video.search.yahoo.com

Tesla also doing "multi modality " now with their super computer Grok.
Please let Akida be part of their solution .....................

Well for a start, the processor load in processing video is enormous, so 8 cameras burn a lot of power. The inherent sparsity of lidar reduces this load significantly. Apparently Tesla's newest FSD processor uses more than 70 W and has 70,000 GPUs and its NN accelerators use MACs.

 

[Fullmoonfever]

Think we all recall the TinyML pedestrian hackathon from last year.

I recall some written summaries / articles and other info around it but not the actual project presentations video.

This is it for anyone interested as just popped up on my YouTube feed....probs cause all my various google and YouTube Akida searches

It's over 2 hrs but Akida is used in the first two solutions and can skip to relevant points with the transcript if wanted.

 

[MDhere]

I'm back but never disappeared. Kind of like bravos run but my jog didn't seem to move the needle so will leave the runs to bravo.

Will catch up on what I've missed. I think we are due for some juicy announcements, just over one month to Agm.

Happy week fellow brners ❤

 

[TECH]

Good morning..
I believe I'm correct in saying that Sean recently stated that the time gap between new iterations of the Akida architecture would be between 12 to 18 months, well that was the suggested 'hope' I'm not sure whether Peter or Anil were onboard with that timeframe, except to say Anil told me a week ago that he was "very busy".

Lets all be 100% honest here, so far ALL the articles and photos show AKD 1000 being referenced in potential product/s or papers or patents...nothing referring to AKD ll as yet, yes Dr. Tony and the Perth Research Institute must keep driving the innovations forward to 'maintain' our alleged lead of 3-5 years..but confirmed market based revenue has to be the immediate goal.

As Akida II was announced to the market in March 2023 with general availability around September/October 2023 I'm expecting news on AKD III within the next 5-6 months..what's your view ?

Tech.