BRN Discussion Ongoing

[SharesForBrekky]

Not sure if this upcoming conference involving Brainchip has been mentioned, but there's NICE 2024 from 22-26 April in San Diego:

NICE 2024 - Agenda

flagship.kip.uni-heidelberg.de

11:05‑11:30
(25+5 min)
Invited talk: TENN: A highly efficient transformer replacement for edge and event processing.
M Anthony Lewis, Yan Ru Pei and Olivier Coenen (BrainChip)

 

[Kachoo]

Not sure if this upcoming conference involving Brainchip has been mentioned, but there's NICE 2024 from 22-26 April in San Diego:

NICE 2024 - Agenda

flagship.kip.uni-heidelberg.de

11:05‑11:30
(25+5 min)
Invited talk: TENN: A highly efficient transformer replacement for edge and event processing.
M Anthony Lewis, Yan Ru Pei and Olivier Coenen (BrainChip)

Oliver was the inventor or the guy that is listed on the TENNS patent!

WO2023250093A1 - Method and system for implementing temporal convolution in spatiotemporal neural networks - Google Patents

Disclosed is a neural network system generally relates to the field of neural networks (NNs). In particular, the present disclosure relates to event-based convolutional neural networks (NNs) that are trained to process spatial and temporal data using kernels represented by polynomial expansion...

patents.google.com

WO2023250092A1 - Method and system for processing event-based data in event-based spatiotemporal neural networks - Google Patents

Disclosed is a method for processing event-based input data using a neural network. The neural network comprises a plurality of neurons and one or more connections associated with each of the plurality of neurons. Further, each of the plurality of neurons is configured to receive a corresponding...

patents.google.com

 

[Fullmoonfever]

Oliver was the inventor or the guy that is listed on the TENNS patent!

WO2023250093A1 - Method and system for implementing temporal convolution in spatiotemporal neural networks - Google Patents

Disclosed is a neural network system generally relates to the field of neural networks (NNs). In particular, the present disclosure relates to event-based convolutional neural networks (NNs) that are trained to process spatial and temporal data using kernels represented by polynomial expansion...

patents.google.com

WO2023250092A1 - Method and system for processing event-based data in event-based spatiotemporal neural networks - Google Patents

Disclosed is a method for processing event-based input data using a neural network. The neural network comprises a plurality of neurons and one or more connections associated with each of the plurality of neurons. Further, each of the plurality of neurons is configured to receive a corresponding...

patents.google.com

This is something on Rudy I posted previously and Mamba which has links to transformers and a version of it that Rudy appears to have worked on or developed?

BRN Discussion Ongoing

Careful...

thestockexchange.com.au

 

[jtardif999]

Hi Hoohoo,

The kits that are sold out are not physically sold out.

At these prices 5000 and 10000 included so many hours of engineering support.

The availability of these people is in short supply so the items are sold out.

I guess the employees are busier on other projects and they are unable to meet the customer's request hence sold out.

Hope that helps but if you want some chips possibly contact the company but you may be on your own.

Cheers

‘I am the walrus, I am the walrus, I am the walrus, hoo hoo kachoo’ sorry but I couldn’t help myself guys!

 

[Pmel]

Episode 31: BrainChip Talks SpaceTech and Beyond with European Space Agency | Michael Novak

big things in the pipe.......👍

www.linkedin.com

He works for samsung.

Dyor

 

[Esq.111]

Morning Pmel ,

Hopefully another Duck, was lost but now found .......



Regards,
Esq.

 

[Iseki]

The valley of death he was referring to did not relate to BRN or Akida. He was referring to the DoD SBIR process, or to be more precise, the follow up to the SBIR, going from sample and test to practical adoption and implementation.

The VoD relates to BRN. We announced we had a relationship with NASA, and since then we are seeking commercialization. We manufactured the Akida1000 so potential customers could sample and test our solution, which M-B did. We receivd feedback from other companies and designed the upgraded Akida2. This is why availability of a manufactured Akida2 chip is important. Yes, it'd be brilliant if someone else did that at their expense, someone wanting a million of them, but the clock is ticking. We are traversing the VoD.

 

[jtardif999]

The VoD relates to BRN. We announced we had a relationship with NASA, and since then we are seeking commercialization. We manufactured the Akida1000 so potential customers could sample and test our solution, which M-B did. We receivd feedback from other companies and designed the upgraded Akida2. This is why availability of a manufactured Akida2 chip is important. Yes, it'd be brilliant if someone else did that at their expense, someone wanting a million of them, but the clock is ticking. We are traversing the VoD.

Accenture Global Solutions, that patent of a week ago flags their intent. They will produce a chip with Akida2 IP embedded sooner than later imo.

 

[rgupta]

The VoD relates to BRN. We announced we had a relationship with NASA, and since then we are seeking commercialization. We manufactured the Akida1000 so potential customers could sample and test our solution, which M-B did. We receivd feedback from other companies and designed the upgraded Akida2. This is why availability of a manufactured Akida2 chip is important. Yes, it'd be brilliant if someone else did that at their expense, someone wanting a million of them, but the clock is ticking. We are traversing the VoD.

There are more than one reason you are wrong
1. He was talking about policies of DOD which leads into failure of companies getting success at the initial stages but later on died coz no one was the taker.
2. He clearly mentioned brainchip is not dependent upon DOD only but have its own commercial success plan.
3. We have already generated so much of attention and publicity that dieing from here will be very difficult. We are already into space thanks to ant 61 brain. If you think about size of ant61 and the competitors you will realise the same was not possible without akida.
4. We have more than 60 partners and the only way to get into valley of death is when the technology is not real. Otherwise we need not look back.
5. We have engagements with a lot of big companies like arm, renasas, microchip, Intel etc ( we now know) and they won't engage with a company for wasting time.
Dyor

 

[stuart888]

Agentic Ai is special. Distinguished Andrew Ng explains it well.

My take Workflow to Reasoning, Double Checking, then presenting the Results to the Requestor.

 

[Smoothsailing]

The VoD relates to BRN. We announced we had a relationship with NASA, and since then we are seeking commercialization. We manufactured the Akida1000 so potential customers could sample and test our solution, which M-B did. We receivd feedback from other companies and designed the upgraded Akida2. This is why availability of a manufactured Akida2 chip is important. Yes, it'd be brilliant if someone else did that at their expense, someone wanting a million of them, but the clock is ticking. We are traversing the VoD.

Accenture Global Solutions, that patent of a week ago flags their intent. They will produce a chip with Akida2 IP embedded sooner than later imo.

Yes indeed that’s good thinking huge company like that will be all over it

 

[stuart888]

Condensed. Power of Agents!

The Agentic Workflow is a waterfall. The video link at the bottom starts later, more of summary mode.

 

[stuart888]

He says the Next AI Wow Moment, like the ChatGpt Moment is Planning!!!

 

[stuart888]

Deeper:

One Agent For Many Worlds, Cross-Species Cell Embeddings, and more | Andrew Ng | 124 comments

Last week, I described four design patterns for AI agentic workflows that I believe will drive significant progress: Reflection, Tool use, Planning and Multi-agent collaboration. Instead of having an LLM generate its final output directly, an agentic workflow prompts the LLM multiple times...

www.linkedin.com

Last week, I described four design patterns for AI agentic workflows that I believe will drive significant progress: Reflection, Tool use, Planning and Multi-agent collaboration. Instead of having an LLM generate its final output directly, an agentic workflow prompts the LLM multiple times, giving it opportunities to build step by step to higher-quality output. Here, I'd like to discuss Reflection.

It's relatively quick to implement, and I've seen it lead to surprising performance gains. You may have had the experience of prompting ChatGPT/Claude/Gemini, receiving unsatisfactory output, delivering critical feedback to help the LLM improve its response, and then getting a better response.

What if you automate the step of delivering critical feedback, so the model automatically criticizes its own output and improves its response? This is the crux of Reflection.

ake the task of asking an LLM to write code. We can prompt it to generate...

 

[stuart888]

 

[Diogenese]

View attachment 60091

In some ways, programming is simpler than NLP because of the limited vocabulary. However the syntax is a rectal ache.

 

[Iseki]

Accenture Global Solutions, that patent of a week ago flags their intent. They will produce a chip with Akida2 IP embedded sooner than later imo.

Really? Then why don't they just come out and say it? You might be right, and maybe all those extra ML programmers are going to be working on thier use cases. Alternatively Accenture might take out a patent as a defense against their competitors. Cost of applying for a patent $10K, cost for producing some Akida2 chips $10M, so yeah, the odds are stacked against us.

 

[Smoothsailing]

Really? Then why don't they just come out and say it? You might be right, and maybe all those extra ML programmers are going to be working on thier use cases. Alternatively Accenture might take out a patent as a defense against their competitors. Cost of applying for a patent $10K, cost for producing some Akida2 chips $10M, so yeah, the odds are stacked against us.

public holiday off ridicules thinking

 

[rgupta]

The VoD relates to BRN. We announced we had a relationship with NASA, and since then we are seeking commercialization. We manufactured the Akida1000 so potential customers could sample and test our solution, which M-B did. We receivd feedback from other companies and designed the upgraded Akida2. This is why availability of a manufactured Akida2 chip is important. Yes, it'd be brilliant if someone else did that at their expense, someone wanting a million of them, but the clock is ticking. We are traversing the VoD.

Hi Isleki
I relisten to the conversation and would like to amend my views.
1. He said 'they were told jetson is here' which means DOD chose old technology over new
2. But then he also mentioned that akida they tried out was not custom made chip and they need for product only for visual identification purposes. And he further states if someone develop a chip as per DOD requirements that can change the whole landscape.
3. So yes you may be partially right as far as DOD is concerned and it proves brainchip will find it difficult to get adjusted without a god father.
But on the same hand that God father can take technology from brainchip and make a chip for DOD.
4. As a business no one is unknown to those type of issues and brainchip must of aware of the same as well.
So does we have someone that godfather in our portfolio ? Yes Intel is that godfather who will manufacture a customised chip for DOD and use our ip. ( My belief , everyone should DYOR)
5. So yes in the end he used the term valley of death for DoD policies and with jetson is on board confirms that includes akida as well.
Dyor

 

[TECH]

This podcast goes for an hour so might try and listen later. Happy Easter!

‎From the Crows' Nest: Advancing Cognitive Warfare: Unveiling Neuromorphic Frontiers on Apple Podcasts | Information Systems Laboratories

Check out the Association of Old Crows (AOC) podcast with ISL’s CEO Dr Joe Guerci. Covers a lot of ground! https://lnkd.in/gVjkSGpY

www.linkedin.com

View attachment 60057

Thanks TC...

That was an excellent Podcast, and some very interesting words expressed by Dr. Joe with regards Brainchip the company out of Australia !

Peter and I shared the odd laugh over the years about Brainchip being an Australian company with it's HQ being Perth, but I realize that we are really (now) a global company, with our HQ in California, spiritually it's always going to be Perth Western Australia.

With the amount of articles coming out now on Neuromorphic Computing, it's certainly just reconfirms what many of us have known for the last 9 years, we (Brainchip) have been way ahead of the mob, ALL thanks to our brilliant visionary, Peter, who without all his years and years of research and never give up attitude this dream may never have been possible, the Brainchip team and all shareholders are poised to reap the benefits of dreaming big and truly believing that one is only ever restricted by ones own lack of imagination.

Products and IP License agreements are coming, when ? sometime is my best guess Tech.