BRN Discussion Ongoing

[jrp173]

Is this really the Podcast shareholders are looking for leading into the AGM? I would have thought a podcast on products and sales may be more appropriate.
I found the comment regarding non exec's holding their bonuses over to 2026 interesting. Does this mean they are not expecting a favourable Agm this year but confident the 2026 Agm will be more palatable. Hmmm

With regard to the non execs holding off their bonuses until 2026, I'd like to know why the other executives are not following suit...

Surely if NEDs performance does not warrant bonuses this year, then the same is true for the executive directors?

The other exec's don't deserve any rewards this year in my opinion.

 

[White Horse]

AH here it is, the worry about what you wish for, arguement rolls out every agm. Mate it can’t be any worse, shareprice down the drain no support on the market, Management hiding shit for what 6 years behind NDA,s then trying to the run and hide to a USA bourse so they don’t have to disclose anything at all. And it will be covered up under the In the interest of the USA DEFENCE, we will know less than we know now.
Shareholders and the market have zero confidence in the management of Brainchip. The current share price is manipulation prior to the AGM so the shorters can push it down the day after the AGM. rinse and repeat happens every year. Remember who sold the 50 million share to the shorters last time to get them out of their shorts and make them millions. Yep Brainchip management did, 50 million shares and they called them sophisticated investor's to try and fool shareholders.

No you are speaking for yourself and a few other people who think they know better.
When in actual fact you don't have a clue. That may be managements fault. But if what they're saying is true, and we would be jeopardizing our relationships. As I say, be careful what you wish for.
Maybe this is not the place for your investment.
Speaking of Defense. Never say never. The immortal words of the founders, now resting in peace.
The attitude of management has changed to "whatever it takes". And it's paying off.

 

[7für7]

Mhaaaaaa…. It’s always the same game, and always the same people who try to stir things up right at the finish line—most likely because they’re short and trying to squeeze out a few more bucks. I can tell who’s genuinely invested and who’s just throwing around emotional garbage. No one who’s truly invested talks negatively about the management non-stop. Anyone who’s followed the progress over the past years knows that there’s been real, hard work toward success.

What do you expect from people sitting on their couch, stuffing their face with pizza, criticizing high-performance football players for not scoring a goal every time? (Yes yes, I know, here come the replies: “I’m the CEO of a huge company,” or “I manage 2,000 people” etc… sure, I’ve heard it all.) Sean isn’t a magician. And if the market isn’t ready to fully adopt the product yet, he has to find other ways to finance the company. That’s called entrepreneurial thinking—something a few people here could use.

And as already mentioned—he doesn’t set his own salary. He was hired ( lol Sean was hired) and given a compensation package that was approved. Jealous? That’s up to you.

But as I said: vote as you see fit, but spare us these shady attempts to influence others with your negative attitude!

 

[The Pope]

Another job advertised

#hiring | BrainChip

We're #hiring a new Sales Manager- Bay Area in San Francisco Bay Area. Apply today or share this post with your network.

www.linkedin.com

 

[White Horse]

Old white man. You have banged on in favour of every move of this board of directors. You are so blind with your grey boomer glasses. Common man. How much value have you and I lost with blind faith.

Tha argument here is not talent. It is about directors telling shareholders how much they are worth. They are telling us they are worthy of their mega $ package. Really. Have you considered what sitting in a board meeting once a month is worth? Hmmm. My thinking from your previous posts of your years gone by ( many years) is you are a little out of touch with reality and modern equity.

You come across as a little racist and entitled.
I consider that I am gaining value every day.
You and others, who may or may not be down rampers thinly disguised, are the ones who asked them to do just that, at the shareholders expense.
Yes I probably am older than you, but I am prepared to wait for something I believe in, unlike yourself who wants it all yesterday.
The argument here is about talent, and good talent costs money. And unfortunately we require great talent in a world where AI/IT is the most desirable area of expertise.
So I think you are the one who is out of touch.
What do you do for a living, or should I say, what is your area of expertise.?

 

[Fullmoonfever]

Appears some within the Sentry Project have been creating some Resnet model variants and deploying to Akida for some positive results.

From memory I think @Frangipani maybe have posted previously on Dr Eappen?

Apols if all this posted but I haven't bothered with a search or caught up with all the posts. Been trying to ignore the renumeration / AGM debates and opinions as I will form my own and don't personally don't feel the need to engage.

From a couple of days ago.

GLUSE: Enhanced Channel-Wise Adaptive Gated Linear Units SE for Onboard Satellite Earth Observation Image Classification

GLUSE: Enhanced Channel-Wise Adaptive Gated Linear Units SE for Onboard Satellite Earth Observation Image Classification​

Thanh-Dung Le, , Vu Nguyen Ha, , Ti Ti Nguyen, , Geoffrey Eappen, , Prabhu Thiruvasagam, , Hong-fu Chou, , Duc-Dung Tran, , Hung Nguyen-Kha, Luis M. Garces-Socarras, ,
Jorge L. Gonzalez-Rios, , Juan Carlos Merlano-Duncan, , Symeon ChatzinotasThis work was funded by the Luxembourg National Research Fund (FNR), with the granted SENTRY project corresponding to grant reference C23/IS/18073708/SENTRY.Thanh-Dung Le, Vu Nguyen Ha, Ti Ti Nguyen, Geoffrey Eappen, Prabhu Thiruvasagam, Hong-fu Chou, Duc-Dung Tran, Hung Nguyen-Kha, Luis M. Garces-Socarras, Jorge L. Gonzalez-Rios, Juan Carlos Merlano-Duncan, Symeon Chatzinotas are with the Interdisciplinary Centre for Security, Reliability, and Trust (SnT), University of Luxembourg, Luxembourg (Corresponding author. Email: thanh-dung.le@uni.lu).This paper is a revised and expanded version of a paper entitled “Semantic Knowledge Distillation for Onboard Satellite Earth Observation Image Classification”, which was accepted for presentation at IEEE ICMLCN 2025, Barcelona, Spain, 26–29 May 2025.

Abstract​

This study introduces ResNet-GLUSE, a lightweight ResNet variant enhanced with Gated Linear Unit-enhanced Squeeze-and-Excitation (GLUSE), an adaptive channel-wise attention mechanism. By integrating dynamic gating into the traditional SE framework, GLUSE improves feature recalibration while maintaining computational efficiency. Experiments on EuroSAT and PatternNet datasets confirm its effectiveness, achieving exceeding 94% and 98% accuracy, respectively. While MobileViT achieves 99% accuracy, ResNet-GLUSE offers 33× fewer parameters, 27× fewer FLOPs, 33× smaller model size (MB), ≈6× lower power consumption (W), and ≈3× faster inference time (s), making it significantly more efficient for onboard satellite deployment. Furthermore, due to its simplicity, ResNet-GLUSE can be easily mimicked for neuromorphic computing, enabling ultra-low power inference at just 852.30 mW on Akida Brainchip. This balance between high accuracy and ultra-low resource consumption establishes ResNet-GLUSE as a practical solution for real-time Earth Observation (EO) tasks. Reproducible codes are available in our shared repository

.......

Figure 5 Performance of the ResNet-GLUSE during the inference on Akida neuromorphic computing at the TelecomAI-lab, SnT

More importantly, due to its architectural simplicity and compactness, the ResNet-GLUSE model can easily be mimicked, adapted, and deployed on the Akida Brainchip neuromorphic computing platform [42]. Experimental deployment on Akida hardware further highlights its exceptional efficiency, with an extremely low inference power consumption averaging 877 mW, achieving an inference energy consumption of just 182.42 mJ/frame, and maintaining a practical frame rate of 4.81 fps. Such results highlight the potential of the ResNet-GLUSE model to operate efficiently on neuromorphic hardware, enabling energy-efficient onboard image analysis in resource-constrained environments. From the boxplot accuracy distribution, depicted in Fig. 5, the model’s accuracy ranges from about 93% to nearly 96%, with a median of approximately 94.7%, demonstrating stable and high performance across multiple runs, further confirming its robust inference performance under varied operational conditions.

.......

ResNet-GLUSE optimizes accuracy, efficiency, and resource consumption. It reduces energy consumption by up to 6× compared to MobileViT on GPUs and enables ultra-low power inference (852.30 mW) on Akida Brainchip. Despite a slight complexity increase over SE, its adaptive gating mechanism delivers substantial performance gains with minimal overhead.

......

Acknowledgment​

We acknowledge Dr. Geoffrey Eappen for his valuable efforts in mimicking and implementing the model on Akida. This work was funded by the Luxembourg National Research Fund (FNR), with granted SENTRY project corresponding to grant reference C23/IS/18073708/SENTRY. Part of this work was supported by the SnT TelecomAI Lab, the Marie Speyer Excellence Grant (BrainSat) and the FNR BrainSatCom project (BRIDGES/2024/IS/19003118).

 

[Guzzi62]

Appears some within the Sentry Project have been creating some Resnet model variants and deploying to Akida for some positive results.

From memory I think @Frangipani maybe have posted previously on Dr Eappen?

Apols if all this posted but I haven't bothered with a search or caught up with all the posts. Been trying to ignore the renumeration / AGM debates and opinions as I will form my own and don't personally don't feel the need to engage.

From a couple of days ago.

GLUSE: Enhanced Channel-Wise Adaptive Gated Linear Units SE for Onboard Satellite Earth Observation Image Classification

GLUSE: Enhanced Channel-Wise Adaptive Gated Linear Units SE for Onboard Satellite Earth Observation Image Classification​

Thanh-Dung Le, , Vu Nguyen Ha, , Ti Ti Nguyen, , Geoffrey Eappen, , Prabhu Thiruvasagam, , Hong-fu Chou, , Duc-Dung Tran, , Hung Nguyen-Kha, Luis M. Garces-Socarras, ,
Jorge L. Gonzalez-Rios, , Juan Carlos Merlano-Duncan, , Symeon ChatzinotasThis work was funded by the Luxembourg National Research Fund (FNR), with the granted SENTRY project corresponding to grant reference C23/IS/18073708/SENTRY.Thanh-Dung Le, Vu Nguyen Ha, Ti Ti Nguyen, Geoffrey Eappen, Prabhu Thiruvasagam, Hong-fu Chou, Duc-Dung Tran, Hung Nguyen-Kha, Luis M. Garces-Socarras, Jorge L. Gonzalez-Rios, Juan Carlos Merlano-Duncan, Symeon Chatzinotas are with the Interdisciplinary Centre for Security, Reliability, and Trust (SnT), University of Luxembourg, Luxembourg (Corresponding author. Email: thanh-dung.le@uni.lu).This paper is a revised and expanded version of a paper entitled “Semantic Knowledge Distillation for Onboard Satellite Earth Observation Image Classification”, which was accepted for presentation at IEEE ICMLCN 2025, Barcelona, Spain, 26–29 May 2025.

Abstract​

This study introduces ResNet-GLUSE, a lightweight ResNet variant enhanced with Gated Linear Unit-enhanced Squeeze-and-Excitation (GLUSE), an adaptive channel-wise attention mechanism. By integrating dynamic gating into the traditional SE framework, GLUSE improves feature recalibration while maintaining computational efficiency. Experiments on EuroSAT and PatternNet datasets confirm its effectiveness, achieving exceeding 94% and 98% accuracy, respectively. While MobileViT achieves 99% accuracy, ResNet-GLUSE offers 33× fewer parameters, 27× fewer FLOPs, 33× smaller model size (MB), ≈6× lower power consumption (W), and ≈3× faster inference time (s), making it significantly more efficient for onboard satellite deployment. Furthermore, due to its simplicity, ResNet-GLUSE can be easily mimicked for neuromorphic computing, enabling ultra-low power inference at just 852.30 mW on Akida Brainchip. This balance between high accuracy and ultra-low resource consumption establishes ResNet-GLUSE as a practical solution for real-time Earth Observation (EO) tasks. Reproducible codes are available in our shared repository

.......

View attachment 82744

Figure 5 Performance of the ResNet-GLUSE during the inference on Akida neuromorphic computing at the TelecomAI-lab, SnT

More importantly, due to its architectural simplicity and compactness, the ResNet-GLUSE model can easily be mimicked, adapted, and deployed on the Akida Brainchip neuromorphic computing platform [42]. Experimental deployment on Akida hardware further highlights its exceptional efficiency, with an extremely low inference power consumption averaging 877 mW, achieving an inference energy consumption of just 182.42 mJ/frame, and maintaining a practical frame rate of 4.81 fps. Such results highlight the potential of the ResNet-GLUSE model to operate efficiently on neuromorphic hardware, enabling energy-efficient onboard image analysis in resource-constrained environments. From the boxplot accuracy distribution, depicted in Fig. 5, the model’s accuracy ranges from about 93% to nearly 96%, with a median of approximately 94.7%, demonstrating stable and high performance across multiple runs, further confirming its robust inference performance under varied operational conditions.

.......

ResNet-GLUSE optimizes accuracy, efficiency, and resource consumption. It reduces energy consumption by up to 6× compared to MobileViT on GPUs and enables ultra-low power inference (852.30 mW) on Akida Brainchip. Despite a slight complexity increase over SE, its adaptive gating mechanism delivers substantial performance gains with minimal overhead.

......

Acknowledgment​

We acknowledge Dr. Geoffrey Eappen for his valuable efforts in mimicking and implementing the model on Akida. This work was funded by the Luxembourg National Research Fund (FNR), with granted SENTRY project corresponding to grant reference C23/IS/18073708/SENTRY. Part of this work was supported by the SnT TelecomAI Lab, the Marie Speyer Excellence Grant (BrainSat) and the FNR BrainSatCom project (BRIDGES/2024/IS/19003118).

Thanks,

From the link you provided:

“Semantic Knowledge Distillation for Onboard Satellite Earth Observation Image Classification”, which was accepted for presentation at IEEE ICMLCN 2025, Barcelona, Spain, 26–29 May 2025.

 

[Tothemoon24]

Edge AI is literally taking off— into orbit.
Over the last few years we’ve seen satellites move from simply collecting data to actually thinking in space. A brand‑new study out this week showcases just how far the hardware‑and‑algorithm stack has come—and why the next wave of intelligent devices, from CubeSats to wearables, will look very different.
What’s new?

* ResNet‑GLUSE: a lightweight convolutional network that marries squeeze‑and‑excitation blocks with gated linear units.
* 94 – 98 % accuracy on EuroSAT & PatternNet while slashing parameters (‑33×), FLOPs (‑27×) and model size (‑33×) vs MobileViT.
* On‑chip inference on Brainchip’s Akida neuromorphic processor draws just ~0.85 W—orders of magnitude below typical GPUs. arXiv
* Open‑source SENTRY repo (link in the paper) lets teams retrain or port the model to their own edge hardware in minutes.

Why it matters:

* Always‑on autonomy
* Continuous links from Starlink & OneWeb solve the connectivity gap, but not the latency gap. Pushing compute into the payload means a satellite can re‑task itself on the fly—think disaster response or precision agriculture updates in the same orbit.
* Millijoule‑class power budgets
* Early missions like ESA’s Φ‑Sat‑1 proved the concept with Intel’s Movidius Myriad‑2 VPU. eoportal.org

The leap to neuromorphic silicon like Akida (sub‑watt inference) takes that efficiency several steps further. Brainchip's "Design Once, Deploy Everywhere"
A GLUSE‑style backbone can run on classic CPUs, tiny MCUs, VPUs, or spiking neural nets—so the same model scales from cloud pre‑training to the harsh radiation of low‑Earth orbit.

Bigger picture

Edge AI isn’t just about satellites. The same architectural principles—tiny models, local learning, neuromorphic acceleration—unlock smarter drones, industrial sensors, autonomous vehicles and next‑gen wearables. Brainchip’s role here is just one example of how neuromorphic IP is slipping quietly into commercial designs; expect a lot more silicon players to follow.

The age of “compute‑everywhere” is here. Whether you’re building the next CubeSat constellation, retro‑fitting factory equipment, or reimagining consumer devices, start designing for extreme efficiency today—the toolchain and the chips have finally caught up.

 

[Iseki]

Imagine how we would go with a CEO from the defense industry. A CEO who could wrap up a couple of profitable deals, and who could earn themselves, and their board, the bonuses they want.

Remember all those sales staff who were acquired from Arm, then subsequently fired, to protect the boards' remuneration package a couple of years ago..

Remember all those changes made to Akida based on reccommendations by potential customers?

Once the board licenses either Akida2, or TENNS, I'd be happy to follow the board's reccomendation on how to vote. Their performance bonuses would be paid because they had performed, and not just worked hard. But not until that time.

 

[Getupthere]

Mhaaaaaa…. It’s always the same game, and always the same people who try to stir things up right at the finish line—most likely because they’re short and trying to squeeze out a few more bucks. I can tell who’s genuinely invested and who’s just throwing around emotional garbage. No one who’s truly invested talks negatively about the management non-stop. Anyone who’s followed the progress over the past years knows that there’s been real, hard work toward success.

What do you expect from people sitting on their couch, stuffing their face with pizza, criticizing high-performance football players for not scoring a goal every time? (Yes yes, I know, here come the replies: “I’m the CEO of a huge company,” or “I manage 2,000 people” etc… sure, I’ve heard it all.) Sean isn’t a magician. And if the market isn’t ready to fully adopt the product yet, he has to find other ways to finance the company. That’s called entrepreneurial thinking—something a few people here could use.

And as already mentioned—he doesn’t set his own salary. He was hired ( lol Sean was hired) and given a compensation package that was approved. Jealous? That’s up to you.

But as I said: vote as you see fit, but spare us these shady attempts to influence others with your negative attitude!

Alright. Let’s imagine you personally employed Sean as the CEO of your company and were conducting his annual performance review. How would you rate his performance over the past 12 months, on a scale of 1 to 10?

 

[jtardif999]

I agree 100% with you Realinfo.

I have worked as a CEO of a company, being “directed” by a board of Directors. All had different qualifications and personal qualities that added diverse thinking and decision making abilities in their process of “directing” me. However, they weren’t “people facing” or working “on the shop floor” so to speak. They were simply observers of the business. They required input from myself and on occasion others level management for information about the day to day operations, specialised skill consultation and direct business community happenings. They were a “third party” to ensure there was no oversight or wrong doing by Management and to “set direction” of the company, after taking into consideration input from all levels of the company. Without that input they were blind. Needless to say the majority thought they were experts in everything and thought they were worth more than anyone involved in the company, despite the worth of their input.

At the end of the day, it is the people from below with the talent that input and carry out the strategic course set by Directors. Directors are worth something, but not mega bucks in the start up phase. They should wait their turn until their strategy is a win win for the company, their employees, their shareholders and lastly themselves.

Anyone who has not had direct involvement in a company management will look from the outside. Some will say Directors are worth every $ they are paid. Other will ask, what are they doing for those mega $. At this stage of a start up phase, I am asking exactly the later. What are they doing to justify this huge fee. Surely they could implement a remuneration structure that benefits them “if” they set the “right direction” that benefits ALL involved with the company! It’s just common sense to preserve capital for more important aspects of the business, like technical talent, like R&D, like compromise is sales situations to kick start engagements.

I’m all in on Brainchip. Have been for many years. In the process of transferring industry super shares to a SMSF holding to ensure “I” have a vote on important matters such as remuneration and US listing. I have learned Industry Fund direct investment does not give certainty in holding shares and certainly no ability to vote.

Open your eyes all 40,000 of you. Keep them honest with your shareholder vote!

“Surely they could implement a remuneration structure that benefits them “if” they set the “right direction” that benefits ALL involved with the company!”

Isn’t that what they’ve already implemented? The directors and CEO are being remunerated with shares in the company. When the company has success and we get rewarded they will also be rewarded.

 

[rgupta]

Gidday Yogi,

I thought it was a very good podcast, both guests spoke plain English, any educated person can see what the company is
trying to do, to me it just reflects the integrity starting from the founders down.

The demand for talent in the AI sector is huge, you can then lift that to another level when you enter the SNN space, as the
pool of talent would have to be considered way smaller, the next wave (the future) is going to be coming from all the Universities
who are promoting Neuromorphic Computing courses for the next generation of engineers and Neuroscientists etc.

It's important to remember that, not all staff are going to be driven by the almighty dollar, a lot of our staff over the years have been
approaching the last part of their careers, so money may not have been the driving force, but I think that incentives for the younger
staff is extremely important, the instant pay rise would always attract some, but being acknowledged for great work, achieving valid
milestones and earning the respect of both Peter and Anil (for example) would be the incentive from my point of view.

Kris Carlson would be a prime candidate in my opinion...he's been a loyal staff member and super talented!

Regards..........Tech

My thoughts right now Get are ones of concern.

For mine, the announcement that Brainchip was considering moving to a US listing was a genuine shock. I contacted the company about my concerns, which centred around the following premise…that unless there was at least one major revenue generating deal announced, that would increase the share price significantly, then it would be grossly premature to move to a US listing. I added that existing shareholders value would be destroyed if the company moved without at least one major revenue producing deal in the kit bag.

The response from the company shocked me even more than their original announcement to consider moving. I was told, that because of the ASX disclosure rules, a major deal may never happen. I was told that the entities Brainchip was dealing with might never do business with us whilst we were listed on the ASX, because they were not prepared to risk being forced to reveal financial details, and information about how they were going to use our IP.

This prompted my discussions with the ASX about their interpretation of their very own disclosure rules, particularly rule 3.1A. They told me very clearly, that if Brainchip and a customer wanted to maintain confidentiality about a deal they were contemplating, then the onus was on both parties to remain silent about it. As long as confidentiality remained, there was no requirement to disclose the deal…it could remain confidential under ASX disclosure rule 3.1A.

When I told the company this, I used both the unnecessary Ford ASX announcement back in May 2020, that most likely caused Ford to end their collaboration with us, and Mercedes self outing themselves with press releases in January 2022, which has caused complete silence from them about us ever since.

Love them or hate them…after my discussions with the ASX, I cannot believe that the company would consider a premature, highly damaging for existing shareholders move to a US listing, because of the ASX disclosure rules.

Call me a conspiracy theorist…but I believe there is another agenda .

So these are my thoughts right now Get.

But they told us many time in the past that they are happy on ASX.
Looks like the only deal on table with them is US DOD, but I don't think we should move to US because of that.
Dyor

 

[rgupta]

My thoughts right now Get are ones of concern.

For mine, the announcement that Brainchip was considering moving to a US listing was a genuine shock. I contacted the company about my concerns, which centred around the following premise…that unless there was at least one major revenue generating deal announced, that would increase the share price significantly, then it would be grossly premature to move to a US listing. I added that existing shareholders value would be destroyed if the company moved without at least one major revenue producing deal in the kit bag.

The response from the company shocked me even more than their original announcement to consider moving. I was told, that because of the ASX disclosure rules, a major deal may never happen. I was told that the entities Brainchip was dealing with might never do business with us whilst we were listed on the ASX, because they were not prepared to risk being forced to reveal financial details, and information about how they were going to use our IP.

This prompted my discussions with the ASX about their interpretation of their very own disclosure rules, particularly rule 3.1A. They told me very clearly, that if Brainchip and a customer wanted to maintain confidentiality about a deal they were contemplating, then the onus was on both parties to remain silent about it. As long as confidentiality remained, there was no requirement to disclose the deal…it could remain confidential under ASX disclosure rule 3.1A.

When I told the company this, I used both the unnecessary Ford ASX announcement back in May 2020, that most likely caused Ford to end their collaboration with us, and Mercedes self outing themselves with press releases in January 2022, which has caused complete silence from them about us ever since.

Love them or hate them…after my discussions with the ASX, I cannot believe that the company would consider a premature, highly damaging for existing shareholders move to a US listing, because of the ASX disclosure rules.

Call me a conspiracy theorist…but I believe there is another agenda .

So these are my thoughts right now Get.

I think that should had been considered before listing the company on ASX. But believe me this management is not standing on their words from day one. There are so many loop holes all the time to tweak around, but yes if there is nothing on the table it is better to say table is not sturdy enough to put something on table.
Dyor

 

[Food4 thought]

Alright. Let’s imagine you personally employed Sean as the CEO of your company and were conducting his annual performance review. How would you rate his performance over the past 12 months, on a scale of 1 to 10?

The only trouble with this is we as shareholders don’t know shit.
Sure lots of dot joining and media announcements but we don’t have a clue of what’s really going on behind closed doors.
I just hope that everything is kosher and we are in the winners circle.
I try to keep the faith in the BOD and our CEO but sometimes it runs thin.
I think if we were to get more positive news from Sean things would be much better. But he is always hiding

 

[rgupta]

You seem to forget that he (Sean) did not employ himself, he was picked by our board and founders of the company.
He would have been offered a package by them, considered commensurate with the requirements of the position.
So who is really responsible. And who first termed the phrase "rapid commercialization".

We need to stop thinking like peanuts, and realize this is not an Australia centric company, it's trying to be an international company, and should not be run by a bunch of geriatric Aussies.

You have a point, but how much he proved himself in last 3&1/2 years. We cannot keep on paying for our mistakes repeatedly.
Had Sean shown us the results, we need not be discussing this remuneration report at all, we could easily be sleeping and enjoying our holdings. But we are discussing the same we the results are not what we were expected from him.
So may be it was a wrong decision and it should be changed.
Dyor

 

[Rach2512]

Sign Up | LinkedIn

500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.

www.linkedin.com

Andrew Wright SVP New Product Introduction from Efabless Corporation, showing interest.

Company

About Efabless

efabless.com

 

[Bravo]

I wonder if this whole redomiciling proposal is just a distraction from the uncomfortable reality that widespread usage and meaningful revenue may still be 5 to 10 years away?





Trends in Neuromorphic Computing CIOs Should Know
Neuromorphic computing technology is advancing rapidly. What does this mean for CIOs and their organizations?

John Edwards, Technology Journalist & Author
April 14, 2025
5 Min Read

Science Photo Library via Alamy Stock Photo

Neuromorphic computing is the term applied to computer elements that emulate the way the human brain and nervous system function. Proponents believe that the approach will take artificial intelligence to new heights while reducing computing platform energy requirements.
"Unlike traditional computing, which incorporates separate memory and processors, neuromorphic systems rely on parallel networks of artificial neurons and synapses, similar to biological neural networks," observes Nigel Gibbons, director and senior advisor at consulting firm NCC Group in an online interview.

Potential Applications​

The current neuromorphic computing application landscape is largely research-based, says Doug Saylors, a partner and cybersecurity co-lead with technology research and advisory firm ISG. "It's being used in multiple areas for pattern and anomaly detection, including cybersecurity, healthcare, edge AI, and defense applications," he explains via email.
Potential applications will generally fall into the same areas as artificial intelligence or robotics, says Derek Gobin, a researcher in the AI division of Carnegie Mellon University's Software Engineering Institute. "The ideal is you could apply neuromorphic intelligence systems anywhere you would need or want a human brain," he notes in an online interview.


"Most current research is focused on edge-computing applications in places where traditional AI systems would be difficult to deploy, Gobin observes. Many neuromorphic techniques also intrinsically incorporate temporal aspects, similar to how the human brain operates in continuous time, as opposed to the discrete input-output cycles that artificial neural networks utilize." He believes that this attribute could eventually lead to the development of time-series-focused applications, such as audio processing and computer vision-based control systems.

Current Development​

As with quantum computing research, there are multiple approaches to both neuromorphic hardware and algorithm development, Saylors says. The best-known platforms, he states, are BrainScaleS and SpiNNaker. Other players include GrAI Matter labs and BrainChip.

Neuromorphic strategies are a very active area of research, Gobin says. "There are a lot of exciting findings happening every day, and you can see them starting to take shape in various public and commercial projects." He reports that both Intel and IBM are developing neuromorphic hardware for deploying neural models with extreme efficiency. "There are also quite a few startups and government proposals looking at bringing neuromorphic capabilities to the forefront, particularly for extreme environments, such as space, and places where current machine learning techniques have fallen short of expectations, such as autonomous driving."

Next Steps​

Over the short term, neuromorphic computing will likely be focused on adding AI capabilities to specialty edge devices in healthcare and defense applications, Saylors says. "AI-enabled chips for sensory use cases are a leading research area for brain/spinal trauma, remote sensors, and AI enabled platforms in aerospace and defense," he notes.
An important next step for neuromorphic computing will be maturing a technology that has already proven successful in academic settings, particularly when it comes to scaling, Gobin says. "As we're beginning to see a plateau in performance from GPUs, there's interest in neuromorphic hardware that can better run artificial intelligence models -- some companies have already begun developing and prototyping chips for this purpose."

Another promising use case is event-based camera technology, which shows promise as a practical and effective medium for satellite and other computer vision applications, Gobin says. "However, we have yet to see any of these technologies get wide-scale deployment," he observes. "While research is still very active with exciting developments, the next step for the neuromorphic community is really proving that this tech can live up to the hype and be a real competitor to the traditional hardware and generative AI models that are currently dominating the market."

Looking Ahead​

Given the technology's cost and complexity, coupled with the lack of skilled resources, it's likely to take another seven to 10 years before widespread usage of complex neuromorphic computing occurs, Saylors says. "However, recent research in combining neuromorphic computing with GenAI and emerging quantum computing capabilities could accelerate this by a year or two in biomedical and defense applications."

Mainstream adoption hinges on hardware maturity, cost reduction, and robust software, Gibbons says. "We may see initial regular usage within the next five to 10 years in specialized low-power applications," he predicts. "Some of this will be dictated by the maturation of quantum computing." Gibbons believes that neuromorphic computing's next phase will focus on scaling integrated chips, refining and spiking neural network algorithms, and commercializing low-power systems for applications in robotics, edge AI, and real-time decision-making.
Gibbons notes that neuromorphic computing may soon play an important role in advancing cybersecurity. The technology promises to offer improved anomaly detection and secure authentication, thanks to event-driven intelligence, he explains. Yet novel hardware vulnerabilities, unknown exploit vectors, and data confidentiality remain critical concerns that may hamper widespread adoption.

Trends in Neuromorphic Computing CIOs Should Know

Neuromorphic computing technology is advancing rapidly. What does this mean for CIOs and their organizations?

www.informationweek.com

 

[7für7]

Alright. Let’s imagine you personally employed Sean as the CEO of your company and were conducting his annual performance review. How would you rate his performance over the past 12 months, on a scale of 1 to 10?

Look… Without knowing what’s happening in the background, there’s no point in speculating about what I would do. There are contracts that have to be respected. As long as these contracts are still in effect, I wouldn’t take any action. None of us knows whether he’s being strung along daily because there’s no revenue or contracts in sight … or whether they know something we don’t. I’m not taking part in the yearly small talk.

Everything else I already mentioned in my previous posts.

 

[Smoothsailing]

Sign Up | LinkedIn

500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.

www.linkedin.com

Andrew Wright SVP New Product Introduction from Efabless Corporation, showing interest.

View attachment 82750
View attachment 82751 View attachment 82751

Company

About Efabless

efabless.com

Unfortunately

 

[Smoothsailing]

I wonder if this whole redomiciling proposal is just a distraction from the uncomfortable reality that widespread usage and meaningful revenue may still be 5 to 10 years away?





Trends in Neuromorphic Computing CIOs Should Know
Neuromorphic computing technology is advancing rapidly. What does this mean for CIOs and their organizations?

John Edwards, Technology Journalist & Author
April 14, 2025
5 Min Read

Science Photo Library via Alamy Stock Photo

Neuromorphic computing is the term applied to computer elements that emulate the way the human brain and nervous system function. Proponents believe that the approach will take artificial intelligence to new heights while reducing computing platform energy requirements.
"Unlike traditional computing, which incorporates separate memory and processors, neuromorphic systems rely on parallel networks of artificial neurons and synapses, similar to biological neural networks," observes Nigel Gibbons, director and senior advisor at consulting firm NCC Group in an online interview.

Potential Applications​

The current neuromorphic computing application landscape is largely research-based, says Doug Saylors, a partner and cybersecurity co-lead with technology research and advisory firm ISG. "It's being used in multiple areas for pattern and anomaly detection, including cybersecurity, healthcare, edge AI, and defense applications," he explains via email.
Potential applications will generally fall into the same areas as artificial intelligence or robotics, says Derek Gobin, a researcher in the AI division of Carnegie Mellon University's Software Engineering Institute. "The ideal is you could apply neuromorphic intelligence systems anywhere you would need or want a human brain," he notes in an online interview.


"Most current research is focused on edge-computing applications in places where traditional AI systems would be difficult to deploy, Gobin observes. Many neuromorphic techniques also intrinsically incorporate temporal aspects, similar to how the human brain operates in continuous time, as opposed to the discrete input-output cycles that artificial neural networks utilize." He believes that this attribute could eventually lead to the development of time-series-focused applications, such as audio processing and computer vision-based control systems.

Current Development​

As with quantum computing research, there are multiple approaches to both neuromorphic hardware and algorithm development, Saylors says. The best-known platforms, he states, are BrainScaleS and SpiNNaker. Other players include GrAI Matter labs and BrainChip.

Neuromorphic strategies are a very active area of research, Gobin says. "There are a lot of exciting findings happening every day, and you can see them starting to take shape in various public and commercial projects." He reports that both Intel and IBM are developing neuromorphic hardware for deploying neural models with extreme efficiency. "There are also quite a few startups and government proposals looking at bringing neuromorphic capabilities to the forefront, particularly for extreme environments, such as space, and places where current machine learning techniques have fallen short of expectations, such as autonomous driving."

Next Steps​

Over the short term, neuromorphic computing will likely be focused on adding AI capabilities to specialty edge devices in healthcare and defense applications, Saylors says. "AI-enabled chips for sensory use cases are a leading research area for brain/spinal trauma, remote sensors, and AI enabled platforms in aerospace and defense," he notes.
An important next step for neuromorphic computing will be maturing a technology that has already proven successful in academic settings, particularly when it comes to scaling, Gobin says. "As we're beginning to see a plateau in performance from GPUs, there's interest in neuromorphic hardware that can better run artificial intelligence models -- some companies have already begun developing and prototyping chips for this purpose."

Another promising use case is event-based camera technology, which shows promise as a practical and effective medium for satellite and other computer vision applications, Gobin says. "However, we have yet to see any of these technologies get wide-scale deployment," he observes. "While research is still very active with exciting developments, the next step for the neuromorphic community is really proving that this tech can live up to the hype and be a real competitor to the traditional hardware and generative AI models that are currently dominating the market."

Looking Ahead​

Given the technology's cost and complexity, coupled with the lack of skilled resources, it's likely to take another seven to 10 years before widespread usage of complex neuromorphic computing occurs, Saylors says. "However, recent research in combining neuromorphic computing with GenAI and emerging quantum computing capabilities could accelerate this by a year or two in biomedical and defense applications."

Mainstream adoption hinges on hardware maturity, cost reduction, and robust software, Gibbons says. "We may see initial regular usage within the next five to 10 years in specialized low-power applications," he predicts. "Some of this will be dictated by the maturation of quantum computing." Gibbons believes that neuromorphic computing's next phase will focus on scaling integrated chips, refining and spiking neural network algorithms, and commercializing low-power systems for applications in robotics, edge AI, and real-time decision-making.
Gibbons notes that neuromorphic computing may soon play an important role in advancing cybersecurity. The technology promises to offer improved anomaly detection and secure authentication, thanks to event-driven intelligence, he explains. Yet novel hardware vulnerabilities, unknown exploit vectors, and data confidentiality remain critical concerns that may hamper widespread adoption.

Trends in Neuromorphic Computing CIOs Should Know

Neuromorphic computing technology is advancing rapidly. What does this mean for CIOs and their organizations?

www.informationweek.com

Clearly he doesn’t know the current position of Brainchip, so I take this with a grain of salt. We are moving into commercial this year