BRN Discussion Ongoing

[Labsy]

In my opinion this is fantastic news for brainchip. It is evidence that we are in the sweet spot and an untouchabke entity on its own. It makes us very attractive to Nvidia and competitors but just out of reach ATM. When akida 2.0 is launched stand back and watch the billion dollar budgets froth at the mouth...

Nvidia Acquired AI Startup That Shrinks Machine-Learning Models

Nvidia in February quietly acquired OmniML, a two-year-old artificial intelligence startup whose software helped shrink machine-learning models so they could run on devices rather than in the cloud, according to a spokesperson and LinkedIn profiles of former OmniML employees who now work at ...

www.theinformation.com

Even Nvidia acknowledges cloudless function is the holy grail....

 

[Cgc516]

As a shareholder, 1000 eyes are doing brilliant job, however , we are still waiting for the updates from company not ASX announcement.

 

[rgupta]

🔮 E05: Neuromorphic Computing & The Future of Edge AI

By 2030, neuromorphic chip designs win 20% of the edge AI hardware market

stateofthefuture.substack.com

On a few occasions I felt author is wrong
e.g there is no commercially available neurophonic chip and even brn akida 1000 is a laboratory chip.
Much more emphasis was laid on memory especially ReRam than the the neurophonic architect.

 

[HopalongPetrovski]

On a few occasions I felt author is wrong
e.g there is no commercially available neurophonic chip and even brn akida 1000 is a laboratory chip.
Much more emphasis was laid on memory especially ReRam than the the neurophonic architect.

A dude called John took him to task in the comments below the article , and he says he will investigate further.

I must admit though, I enjoyed his explanations and also found the points made greatly favouring and mirroring what has been revealed about our approach thus far.

As an example, He says......

"Neuromorphic designs are similar. They will not replace non-Neumann designs but will "win market share" in tasks where parallelism, real-time processing, and pattern recognition are required. Neuromorphic designs are particulary suited to edge computing and AI, two huge growth areas."

(***should read Von-Neumann designs)


And further as his point 4 on open questions in the industry?.......

"How will the software ecosystem play out? A start-up sort of needs to develop their own software to run on the hardware to demonstrate capabilities. This is especially true for AI, there is no point offering a new piece of hardware and hoping developers will put in the time to figure out how to write software for it. Especially when they can just go to Nvidia, deploy and start making fun and profit. My bet is that any neuromorphic startup needs to launch with an algorithm that is competitive with transformers."

We have been attending to these issues for a number of years now with our converters and dedicated software specialists and our high spec Gen 2 version soon to be available is transformer capable, so we seem to be hitting many targets.

However he does also seem to be indicating that we may have to endure through till 2030 before we are really hitting our straps and beginning to be mainstream.

 

[HopalongPetrovski]

A dude called John took him to task in the comments below the article , and he says he will investigate further.

I must admit though, I enjoyed his explanations and also found the points made greatly favouring and mirroring what has been revealed about our approach thus far.

As an example, He says......

"Neuromorphic designs are similar. They will not replace non-Neumann designs but will "win market share" in tasks where parallelism, real-time processing, and pattern recognition are required. Neuromorphic designs are particulary suited to edge computing and AI, two huge growth areas."

(***should read Von-Neumann designs)


And further as his point 4 on open questions in the industry?.......

"How will the software ecosystem play out? A start-up sort of needs to develop their own software to run on the hardware to demonstrate capabilities. This is especially true for AI, there is no point offering a new piece of hardware and hoping developers will put in the time to figure out how to write software for it. Especially when they can just go to Nvidia, deploy and start making fun and profit. My bet is that any neuromorphic startup needs to launch with an algorithm that is competitive with transformers."

We have been attending to these issues for a number of years now with our converters and dedicated software specialists and our high spec Gen 2 version soon to be available is transformer capable, so we seem to be hitting many targets.

However he does also seem to be indicating that we may have to endure through till 2030 before we are really hitting our straps and beginning to be mainstream.

"However he does also seem to be indicating that we may have to endure through till 2030 before we are really hitting our straps and beginning to be mainstream."

Just in answer to myself cause I'm a bit schitzo today.......
I think there is going to be an element of luck involved and we should be making money well before then, but just how much is going to depend on just who is involved and when.
Big players with longer investment timeframes seem to be gradually accumulating which bodes well in the longer term, but I hope and think we should start seeing "some" traction over the next 12 months.
Just have to factor it in and hope for pleasant surprises to be revealed and that the macro world doesn't turn too nasty in the meantime.

 

[Diogenese]

"However he does also seem to be indicating that we may have to endure through till 2030 before we are really hitting our straps and beginning to be mainstream."

Just in answer to myself cause I'm a bit schitzo today.......
I think there is going to be an element of luck involved and we should be making money well before then, but just how much is going to depend on just who is involved and when.
Big players with longer investment timeframes seem to be gradually accumulating which bodes well in the longer term, but I hope and think we should start seeing "some" traction over the next 12 months.
Just have to factor it in and hope for pleasant surprises to be revealed and that the macro world doesn't turn too nasty in the meantime.

Nothing til the 30s - he knows of Akida, but not about Akida.

Which applications need Akida:

One of the most urgent commercial applications is automotive.

DMS is legislated in Europe for all new models NOW.

The switch to EVs is generating urgency to reduce power consumption in pursuit of the extra km.

There is a life and death struggle between car makers to perfect ADAS/AD. Are we in SCALA 3?

In-vehicle infotainment - "Hey Mercedes!" always-on key word spotting.

SDR - the demand for bigger and better communication pipes (MorseMicro?) requires a means of squeezing more out of bandwidth because they stopped making bandwidth 14 billion years ago.

NASA - TTM and Mars ...

CubeSats - we are already there ...

Cloud - the ever increasing power consumption will melt the icecaps. They are even building server farms in the Arctic????

Defence - sadly in the 21st century, we still urgently need the latest and fastest and smartest. This is being driven by the global arms race, US Chips Act ...

 

[Dhm]

A dude called John took him to task in the comments below the article , and he says he will investigate further.

I must admit though, I enjoyed his explanations and also found the points made greatly favouring and mirroring what has been revealed about our approach thus far.

As an example, He says......

"Neuromorphic designs are similar. They will not replace non-Neumann designs but will "win market share" in tasks where parallelism, real-time processing, and pattern recognition are required. Neuromorphic designs are particulary suited to edge computing and AI, two huge growth areas."

(***should read Von-Neumann designs)


And further as his point 4 on open questions in the industry?.......

"How will the software ecosystem play out? A start-up sort of needs to develop their own software to run on the hardware to demonstrate capabilities. This is especially true for AI, there is no point offering a new piece of hardware and hoping developers will put in the time to figure out how to write software for it. Especially when they can just go to Nvidia, deploy and start making fun and profit. My bet is that any neuromorphic startup needs to launch with an algorithm that is competitive with transformers."

We have been attending to these issues for a number of years now with our converters and dedicated software specialists and our high spec Gen 2 version soon to be available is transformer capable, so we seem to be hitting many targets.

However he does also seem to be indicating that we may have to endure through till 2030 before we are really hitting our straps and beginning to be mainstream.

The author Lawrence was very polite with John and I added my 2c as well. Politeness is an often missed virtue.

 

[HopalongPetrovski]

The author Lawrence was very polite with John and I added my 2c as well. Politeness is an often missed virtue.

Politeness has evolved as a useful social lubricant.
And nice to see Lawrence was not only open to new information, but was positively welcoming of it to expand his knowledge base.

We have all been somewhat brutalised by the likes of the fools, the Fin and the conflict model clowns over at the crapper and perhaps even by the merciless, persistent and orchestrated shorter's attack's on our share price and so can tend to be a touch reactive to supposed criticism.
Not everyone is out to get us.

 

[IloveLamp]

On a few occasions I felt author is wrong
e.g there is no commercially available neurophonic chip and even brn akida 1000 is a laboratory chip.
Much more emphasis was laid on memory especially ReRam than the the neurophonic architect.

 

[Quercuskid]

Politeness has evolved as a useful social lubricant.
And nice to see Lawrence was not only open to new information, but was positively welcoming of it to expand his knowledge base.

We have all been somewhat brutalised by the likes of the fools, the Fin and the conflict model clowns over at the crapper and perhaps even by the merciless, persistent and orchestrated shorter's attack's on our share price and so can tend to be a touch reactive to supposed criticism.
Not everyone is out to get us.

So politeness is the lubricant to social intercourse, hmmmmm

 

[HopalongPetrovski]

So politeness is the lubricant to social intercourse, hmmmmm

Yes, now please, bend over and spread 'em!

 

[Frangipani]

A dude called John took him to task in the comments below the article , and he says he will investigate further.

I must admit though, I enjoyed his explanations and also found the points made greatly favouring and mirroring what has been revealed about our approach thus far.

As an example, He says......

"Neuromorphic designs are similar. They will not replace non-Neumann designs but will "win market share" in tasks where parallelism, real-time processing, and pattern recognition are required. Neuromorphic designs are particulary suited to edge computing and AI, two huge growth areas."

(***should read Von-Neumann designs)


And further as his point 4 on open questions in the industry?.......

"How will the software ecosystem play out? A start-up sort of needs to develop their own software to run on the hardware to demonstrate capabilities. This is especially true for AI, there is no point offering a new piece of hardware and hoping developers will put in the time to figure out how to write software for it. Especially when they can just go to Nvidia, deploy and start making fun and profit. My bet is that any neuromorphic startup needs to launch with an algorithm that is competitive with transformers."

We have been attending to these issues for a number of years now with our converters and dedicated software specialists and our high spec Gen 2 version soon to be available is transformer capable, so we seem to be hitting many targets.

However he does also seem to be indicating that we may have to endure through till 2030 before we are really hitting our straps and beginning to be mainstream.

Not only did Lawrence (the author) reply to John (a reader correcting him on what he had written about Brainchip) directly in the comment section below, he also refers to this very comment in his latest blog post (and even links to the Brainchip website), in which he also mentions that he will post on neuromorphic computing again next week (let’s see whether he will get that analog vs. digital right), and is thrilled that a tech expert, whom he seems to revere, provided a link to the deep tech tracker curated by him in his latest newsletter.

🔮 E06: Large Language Models and the Assembly Line

By 2030 LLMs will have increased labour productivity by 20% in OECD countries versus 2022 benchmark levels.

stateofthefuture.substack.com

E06: The Model T Moment: LLMs and the Assembly Line​

Finishing the job the Internet started and solving the productivity paradox​

LAWRENCE LUNDY-BRYAN
30.06.2023
1
Share

Started a little blog just to get some traffic. How do we all feel about a hip hop differentiated deep tech newsletter? Ben of a16 did the hard thing about hard things. I had more feedback on Eminem and Nate Dogg than on neuromorphic computing. Sad.

E05: The Future of Edge AI: "Brain-inspired" Neuromorphic Computing​

LAWRENCE LUNDY-BRYAN
·
23. JUNI
Read full story
But the law of audience capture begs me to do more.

There are three rap quotes in this newsletter. Bring me the quotes and some fair criticism in the comments and you get a prize I guess? You want money? You want street cred? I tell you what, I’ll go out and buy the CD and ship it to you. Old school. Who said newsletters were boring? Tell you friends.
Why bother reading on?

• How LLMs finish the job the Internet started in the digital supply chain
• Why LLMs are the only technology that scored 5+ in our assessments
• Why LLMs will rescue crypto (Yes, it does need rescuing!)

Onward Into Battle, Lawrence

---

Mailbag​

​

On neuromorphic, Thanks to
Azeem Azhar
for the shout out. “Lawrence Lundy and Lunar Ventures built a research tool to understand the deep tech landscape, and they’re making it freely available. “

Exponential View by Azeem Azhar
ByteDance’s big leap; grande langue; sodium batteries; predictably random ++ #428
Hi, I’m Azeem Azhar. As a global expert on exponential technologies, I advise governments, some of the world’s largest firms, and investors on how to make sense of our exponential future. Every Sunday, I share my view on developments that I think you should know about in this newsletter…
Read more
5 days ago · 37 likes · Azeem Azhar

I made it ma.

Some feedback from John Dent:

“I am sorry, but your statement about what is available, is factually incorrect. BrainChip does have a commercially available product, and it has been available for months. Their "Akida" chip technology is evolving and the second generation chip is soon to be released. You might want to review your investment strategy. Their tech is certainly not OLD school.”

Fair play. More here. The press release says: “Introduces Vision Transformers and Spatial-Temporal Convolution for radically fast, hyper-efficient and secure Edge AIoT products, untethered from the cloud”

This hints at some early thinking about value proposition: edge, low-latency, secure, and local. They also say “Akida’s neuromorphic processing platform is event-based, fully digital, portable, and proven in silicon.” This pitch validates the assessment that digital will be more popular over analog short-term.

I spent some more time this week on neuromorphic and analog computing. I’ll explain/try to explain the difference next week. What they both have in common is a massive change on the demand-side….

(…)

——————————————————————————————————-

It is worth noting that Lawrence Lundy-Bryan is a research partner with deep tech venture fund Lunar Ventures, based in Berlin, Germany (“We are technical investors commited to the European DeepTech ecosystem”) as well as an angel investor in Rain AI as member of a family fund (something he also mentions in the original blog post):

Lawrence Lundy-Bryan FRSA​

Partner, Research @Lunar Ventures. Angel Investor. stateofthefuture.substack.com​

Lunar Ventures​

Uckfield, England, United Kingdom​

4K followers 500+ connections​

​

​

​

​

Join to follow​

About​

I lead research at Lunar Ventures (https://lunar.vc), a deep tech venture fund. I’ve advised the WEF, OECD, and the UK Government on emerging technology. I’ve consulted with ARM, Shell, Amadeus and Verizon. I spend my days exploring technologies and trends and seeing if I can make sensible predictions. I've made two big predictions:

1/ In 2015, I thought blockchains were computing platforms, not just a ledger for cryptocurrencies. This insight led to my Convergence thesis. The prediction could still go either way, but I’m worried I was short-term right (speculation) but long-term wrong (value generation).

2/ In 2021, I wrote that privacy-enhancing technologies (PETs) enabled collaborative computing. If I'm right, privacy-enhancing tools are one of the most misunderstood and undervalued technologies today. I predict they'll be fundamental pieces of the web. It's too early to tell if I'm right.

I'm finishing up a 2-year horizon-scanning project called the State of the Future, launching at the end of February. We’ve explored 100+ technologies and assessed, compared and combined them—everything from xenobots to VTOLs, solar sails to fusion, and neuromorphic computing to whole-brain emulation.

(…)

Partner, Research​

Lunar Ventures​

Sep 2020 - Present 2 years 11 months
I'm a research partner with Lunar Ventures, a European early-stage deep tech venture fund. I try to figure out what the future might look like and invest in startups building it. I am interested in anything deep tech related that might be important in the future, and I'm actively exploring:

- Privacy-enhancing technology like homomorphic encryption and multi-party computation
- Cryptographically secure data infrastructure with things like data unions and data markets
- Edge hardware with things like analog computing, photonics, and neuromorphic computing

We've invested in 18 companies:

1. Mutable: Moving cloud computing to the edge for ubiquitous computing
2. Mobius: AI for superhuman vision in every application
3. Kyso: Scaling data-driven decision making with data science collaboration
4. Molecule: Decentralized financial ecosystem for funding new therapies
5. Neurolabs: Shortcut the way to more accurate ML using synthetic vision
6. Zama: Making private computation for ML a reality with end-to-end encryption
7. deepset: Open source platform for a new generation of language-aware applications
8. NannyML: Monitoring mission critical systems for ML degradation
9. Wasp: Language for developing full-stack web apps with zero boilerplate code
10. Codavel: Latency-resistant content delivery for mobile connections
11. Mystic.ai: Bringing ML to deployment with a high performance serverless cloud
12. Stack: Reimagining the web browser for the post browsing era
13. Vaxine: Global database with speed and integrity
14. Stealth: Blood-based screening test and platform for personalised drug development
15. Stealth: Rethinking how people interact with tools, computers and ideas
16. Samudai: Productivity tools for DAOs
17. Stealth: AI-supercharged automation
18. Stealth: Framework for building multiplayer gamesShow less

​

Angel Investor​

Lundy-Bryan & Sons Family Office​

Jan 2020 - Present3 years 7 months
1. Charm Therapeutics: Drug discovery through protein folding
2. Bonnet: Making EV charging accessible to all
3. Rain Neuromorphics: Brain-inspired hardware for artificial intelligence
4. Uhubs: Intelligent and personalised training for sales teams
5. Pool: Infrastructure for data unions to create world-class experiences
6. Elemendar: Automating cyber threat intelligence
7. Nevermined: Enterprise-grade data sharing
8. Telesqobe: Grammarly for video
9. Roleshare: Smart matching site for shared roles
10. Bluejay: Local stablecoins for local economies
11. Catapult: Delightful onboarding experiences for DAOs
12. Upside Money: Frictionless cashback

 

[Mt09]

I’m feeling a Monday partnership announcement.

 

[Stable Genius]

Hi all,

This is not new but reassuring (sort of, in a necessary kind of way) to see where we could end up:



Enjoy

 

[Sirod69]

really interesting that Thomas Hülsing from Airbus Defence and Space GmbH is so committed to Brainchip

Thomas Hülsing System-of-Systems Engineering
1 Std.
BrainChip vs Brain Chip — What is the difference?

More information about BrainChip Akida can be found here:
https://brainchip.com/
https://lnkd.in/eyDw6ZZZ

Neuralink vs Akida — What is the difference?
https://lnkd.in/eRx2srt4



Thomas Hülsing System-of-Systems Engineering
51 Min.
Here is an article on neuromorphic computing that explains how it works.
BrainChip Akida realizes this in an IP (Intellectual Property) that can be used by smart device vendors.
https://www.redeweb.com/en/articulos/dispositivos-neuromorficos-en-tinyml/

 

[Boab]

I’m feeling a Monday partnership announcement.

Yes @Mt09 I'm feeling it too. Sean did say his Sunday afternoon/evening hookup with all the company sales and marketing team will hopefully have something very positive to announce.
Can't remember if he said it at the AGM or the interview with Noah the next day?

 

[MrRomper]

New Brainchip Patent (US) Published 29/6

https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/20230206066

 

[Fullmoonfever]

Nice to see Raytheon Technologies playing with neuromorphic in their advanced sensing programs incl RF.

https://www.rtx.com/who-we-are/we-are-rtx/transformative-technologies

Transformative Technologies​

Advanced sensing

• Our RF systems use higher-power microelectronics, increased processing power and software-defined apertures to achieve next-generation capabilities for radar, electronic warfare, communications and multifunction radio frequency applications.
• We’re advancing electro-optical/infrared and other systems such as space-based multispectral sensors and electro-optical distributed aperture systems variants.
• We are providing increased capability against advanced threats and countermeasures by enhancing high-bandwidth digital waveform generation, AI-enabled intelligent signal processing and advanced neuromorphic processing.
• Our acoustic systems enable advanced mine-hunting and undersea networking capabilities through high sensitivity, directionality, multiple access and multi-mod active and passive capabilities for sonar, communications and navigation.
• Our missile seekers counter a wide range of advanced threats through advanced processing and algorithms, all while achieving low size, weight, power and cost.

 

[Diogenese]

New Brainchip Patent (US) Published 29/6

https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/20230206066

Nice pick up Mr Romper,

The full documentation is not up on Espacenet yet (published 20230629).

Inventors:
MCLELLAND DOUGLAS [FR]; CARLSON KRISTOFOR D [US]; JOHNSON KEITH WILLIAM [AU]; JOSHI MILIND [AU]

No PvdM as inventor.

Milind Joshi is our patent attorney in Perth.



US2023206066A1 SPIKING NEURAL NETWORK

Disclosed herein are system, method, and computer program embodiments for an improved spiking neural network (SNN) configured to learn and perform unsupervised, semi-supervised, and supervised extraction of features from an input dataset. An embodiment operates by receiving a modification request to modify a base neural network, having N layers and a plurality of spiking neurons, trained using a primary training dataset. The base neural network is modified to include supplementary spiking neurons in the Nth or N + 1th layer of the base neural network. The embodiment includes receiving a secondary training dataset and determining membrane potential values of one or more supplementary spiking neurons in the Nth or Nth + 1 layer which learn features based on secondary training data set to select a supplementary/winning spiking neuron. The embodiment performs a learning function for the modified neural network based on the winning spiking neuron.

Disclosed herein are system, method, and computer program embodiments for an improved spiking neural network (SNN) configured to learn and perform unsupervised, semi-supervised, and supervised extraction of features from an input dataset. An embodiment operates by receiving a modification request to modify a base neural network, having N layers and a plurality of spiking neurons, trained using a primary training dataset. The base neural network is modified to include supplementary spiking neurons in the Nth or N + 1th layer of the base neural network. The embodiment includes receiving a secondary training dataset and determining membrane potential values of one or more supplementary spiking neurons in the Nth or Nth + 1 layer which learn features based on secondary training data set to select a supplementary/winning spiking neuron. The embodiment performs a learning function for the modified neural network based on the winning spiking neuron.

Looks like the supplementary spiking neurons are 1002, 1004.

One thing it is designed to do is to adjust the multi-bit weights, so I guess that's why they need the ALU.

 

[IloveLamp]