BRN Discussion Ongoing

alwaysgreen

Top 20
I'll chime in on the discussion about the GPUs with my limited knowledge.

Can Akida replace GPUs in cars for running the sensors that are currently run on GPUs? Yes, absolutely as per Anil's statement.

Can Akida run graphic intensive applications such as modern video games, 3D architectural modelling software that need graphics cards (GPUs) that cost $2,500 and require ridiculous amounts of water cooling. Not to my knowledge and this is not the intended market for Akida. Would love @Diogenese to prove me wrong but I'm pretty sure I'm right on this otherwise Akida would be flying off shelves in every video game shop on Earth.
 
Last edited:
  • Like
  • Love
Reactions: 12 users

Diogenese

Top 20


"NVIDIA Jetson Orin Nano utilizes the Ampere-based GPU, along with eight streaming multiprocessors containing 1,024 CUDA cores and 32 Tensor Cores, which will be used for processing artificial intelligence workloads. The Ampere-based GPU Tensor Cores offer improved performance per watt support for sparsity, allowing for twice the Tensor Core throughput."

US2022327101A1 INCREASING SPARCITY IN DATA SETS (Priority 20210409)

1667038555686.png



[0522] ... AI services 3818 may leverage AI system 3824 to execute machine learning model(s) (e.g., neural networks, such as CNNs) for segmentation, reconstruction, object detection, feature detection, classification, and/or other inferencing tasks.

NVIDIA goes fast because it uses lots of TOPS. They reduce power by sparsity, in their case omitting multiplications by zero.

The power figures are 5W to 15W
 
  • Like
  • Fire
  • Love
Reactions: 11 users
"NVIDIA Jetson Orin Nano utilizes the Ampere-based GPU, along with eight streaming multiprocessors containing 1,024 CUDA cores and 32 Tensor Cores, which will be used for processing artificial intelligence workloads. The Ampere-based GPU Tensor Cores offer improved performance per watt support for sparsity, allowing for twice the Tensor Core throughput."

US2022327101A1 INCREASING SPARCITY IN DATA SETS (Priority 20210409)

View attachment 20568


[0522] ... AI services 3818 may leverage AI system 3824 to execute machine learning model(s) (e.g., neural networks, such as CNNs) for segmentation, reconstruction, object detection, feature detection, classification, and/or other inferencing tasks.

NVIDIA goes fast because it uses lots of TOPS. They reduce power by sparsity, in their case omitting multiplications by zero.

The power figures are 5W to 15W
Wow that’s amazing only 5 to 15 Watts that’s less than the human brain.

What’s that Blind Freddie but at least 50 to 150 times more power hungry than AKIDA.

Sorry Freddie @Diogenese did not say if the 5 to 15 watts includes the electricity to run the cooling fan. Maybe but he has been up all day and it is Saturday.

Oh well back to the drawing board.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Haha
  • Like
Reactions: 12 users

Diogenese

Top 20
I'll chime in on the discussion about the GPUs with my limited knowledge.

Can Akida replace GPUs in cars for running the sensors that are currently run on GPUs? Yes, absolutely as per Anil's statement.

Can Akida run graphic intensive applications such as modern video games, 3D architectural modelling software that need graphics cards (GPUs) that cost $2,500 and require ridiculous amounts of water cooling. Not to my knowledge and this is not the intended market for Akida. Would love @Diogenese to prove me wrong but I'm pretty sure I'm right on this otherwise Akida would be flying off shelves in every video game shop on Earth.
Hi ag,

I'm not going to risk giving @Fact Finder another chance to rub my nose in it by saying that Akida can't do GPU, but Akida does not run programs. The Cortex MCU is used only for configuration, but does not play any part in the core business of the SNN - classification/inference.
However, there are many cases where Akida can run as a co-processor with a stored program computer.
Given that games are audio-visual, and given that user input is part of the game, there may be scope for Akida in processing some of the user inputs.

However, I've never played one of these new-fangled contraptions:

 
  • Like
  • Love
  • Fire
Reactions: 17 users

RobjHunt

Regular
Hi ag,

I'm not going to risk giving @Fact Finder another chance to rub my nose in it by saying that Akida can't do GPU, but Akida does not run programs. The Cortex MCU is used only for configuration, but does not play any part in the core business of the SNN - classification/inference.
However, there are many cases where Akida can run as a co-processor with a stored program computer.
Given that games are audio-visual, and given that user input is part of the game, there may be scope for Akida in processing some of the user inputs.

However, I've never played one of these new-fangled contraptions:


The most remarkable thing about BrainChip is …….

BrainChip!

Pantene peeps
 
  • Like
  • Fire
  • Love
Reactions: 14 users
Goodness me there is that nasty 2024 year being thrown around again.

That is a whole five 4C’s away.

“We’ll all be ruined said Hanrahan”

What’s that Blind Freddie Brainchip has over $US50 million dollars available to meet a cash burn of around $15 million per annum but is that enough to get us to 2024 that’s over 12 months away?

Why are you putting on your hat? Where are you going? Speak to me? Why did you slam the door?

All I asked was is $US50 million enough to keep the lights on at Brainchip for another 12 months.

My opinion only DYOR
FF

AKIDA BALLISTA
Fact finder
All I see is everyone on this site wants things to go only one way, clearly they know very little about business and how things work in the real world.
It’s not a straight line
Obstacles get in the way
Things change that’s out of your control
It would be very interesting to see all the negative people on here actually run a business, a start up, build something from scratch. But most of them won’t do anything of the sort because they don’t have the balls or just don’t know where to start. But they are very quick to attack the people who do, quite frankly they shit me to tears.
How many business start from the beginning with nothing and go up up ⬆️
Without any set backs or problems, I personally don’t know of any.

But if you believe in the science and the product then please hang on hold tight because this is going to go in one direction once we start to move away from the platform
 
  • Like
  • Love
  • Fire
Reactions: 29 users

JDelekto

Regular
I'll chime in on the discussion about the GPUs with my limited knowledge.

Can Akida replace GPUs in cars for running the sensors that are currently run on GPUs? Yes, absolutely as per Anil's statement.

Can Akida run graphic intensive applications such as modern video games, 3D architectural modelling software that need graphics cards (GPUs) that cost $2,500 and require ridiculous amounts of water cooling. Not to my knowledge and this is not the intended market for Akida. Would love @Diogenese to prove me wrong but I'm pretty sure I'm right on this otherwise Akida would be flying off shelves in every video game shop on Earth.

Convolutional Neural Networks (CNNs) rely on many mathematical operations using matrices to build a model. The feature of the GPU that makes it ideal for CNN processing is that many processing cores can do these matrix operations concurrently or a lot of math operations simultaneously.

The Akida processor is geared more toward Spiking Neural Networks (SNNs), from my understanding, does not require the matrix math processing used by a CNN and only processes the information from sensors that change. For example, only the portion of pixels in a video stream that change needs to be processed.

So the role of GPUs and Akida in AI processing is generating the mathematical models that allow them to recognize patterns. These patterns can be images, gestures captured in a video, or voice commands (not just the command, but the specific user's voice), similar to how our brains learn and recognize patterns of these same inputs.

Akida has the advantage that: a) it can train its model on chip (using its one-shot learning) more efficiently, and b) this training is done faster and with less computational power. In many cases today, the models trained for CNNs are done on machines with powerful GPUs or shifted to the Cloud, where many of these GPUs (usually specialized for the task) can generate these in parallel.

Akida does not do any rendering, ray tracing, application of textures to polygons, etc., which are things that gamers and content creators expect a GPU to do. We can still expect to see customers like Mercedes still using GPUs in their User Interface, though they will probably go for a lower-power solution, like the ARM-based Qualcomm Snapdragon XR2 family.

I don't believe the Qualcomm solution will replace Akida but instead work in tandem with it. There are many types of sensors that go beyond video and audio. Several examples are temperature, proximity, infrared, ultrasonic, light, smoke & gas, humidity, motion, pressure, etc. If these sensors can produce spike trains, then Akida can learn from them.

While several of these types of sensors are employed in informational systems today (for example, your car beeps when too close to an obstacle or can tell you that your tire pressure is too low), Akida would be able to learn from a sensor's input and be able to recognize patterns which can result in more reactive systems, as opposed to just presenting the information.

I think there will be a lot of applications for Akida that are yet to discover.
 
  • Like
  • Love
  • Fire
Reactions: 23 users
On a different note, Why its important to keep the secret sauce secret.

Four Samsung employees charged with semiconductor tech theft


It is reported that four current and former Samsung employees have been charged with theft of proprietary semiconductor technology. These employees reportedly stole highly valued semiconductor chip technology from Samsung and leaked them to an overseas firm.

The Seoul Central District Prosecutors Office has indicted those four employees with physical detention for violating the unfair competition prevention act and the industrial technology protection act. Two of those employees are former engineers, while the remaining two work as researchers for Samsung Engineering.

One former employee, who worked for Samsung’s semiconductor division, obtained blueprints and operation manuals for the ultrapure water system and other critical technical data while searching for a job. He then leaked that to a Chinese semiconductor consulting firm while searching for a job there. And when he got a job there, he used the stolen data and ordered an ultrapure water system.

Prosecutors said another former Samsung employee was charged with stealing a file containing key semiconductor technology. He allegedly gave that file to the company’s competitor Intel while still working for Samsung. He stole data by capturing images of the data in the file.


Learning
Even thought I´m a Samsung investor, I don´t feel bad for them, Samsung is stealing technology from everybody. Their QLED technology is based on stolen technology, some of their server RAM is stolen directly from Netlist, they even infringed on Apple. They are thieves, but good at it and I´ve given up having morals when investing.
 
  • Like
  • Fire
  • Wow
Reactions: 8 users
A difficult day replaced by a new one. So I am resetting my attitude as I became too confrontational for which I apologise.

I believe in the big picture and one 4C is like a zero in a scene being monitored by AKIDA it just does not rate a spike and certainly not the one seen yesterday.

There are at least 21 reasons why this is so and I posted them yesterday but for the moment ignore every single one of those points and consider how a few million dollars in the 4C can out weighs this publicly, Edge Impulse advertised FACT about AKIDA 1.0.

AKIDA 1.0 out performs GPUs and CPUs.

In the following Techradar surveys the best cheap GPUs are discussed and they are all priced in the hundreds of dollars.


In the tests conducted by Nviso it was able to run AKIDA at 1,000 fps and privately it stated they had achieved 1,670 fps.

In Gaming I am informed by another poster here that frames per second performance is very important and I have even found an article which explains this:


Now if you have opened the above link and read the article you will have found it includes a chart of the frames per second performance of better GPUs and they are sitting in the low hundreds of fps.

By now you might have started to see the very, very big picture and why one 4C or a dozen 4C’s means absolutely nothing against the FACT that:

1. Rob Telson has secured a partnership with Edge Impulse;

2. Edge Impulse has over 50,000 developers using their platform;

3. Edge Impulse has recently completed the incorporation of Meta TF into its platform;

4. Edge Impulse is encouraging these 50,000 plus developers to use Meta TF;

5. Edge Impulse is publicly stating that AKIDA 1.0 out performs GPUs and CPUs;

6. AKIDA 1.0 costing $25.00 runs at more than tens of hundreds of fps;

7. GPU’s costing $100’s run at less than a couple of hundred fps; and

7. The bonus FACT MEGACHIPS largest customer is Nintendo and Brainchip is a contracted supplier of AKIDA technology to MEGACHIPS.

My opinion only I will take the above FACTS over a sugar hit in a 4C every day of the week.

Remember the Data Scientist PhD student who described AKIDA 1.0 as a BEAST? Well that is exactly what it is.

Now add back the 21 other reasons to ignore a 4C result and if an investor cannot now see the very, very big picture staring them in the face then there is nothing more to say.

My opinion only DYOR
FF

AKIDA BALLISTA

I think it´s relevant to compare nVidias graphics cards to Akida, but we have to be sure we´re measuring the same thing. Generated 3D pictures FPS or AI inference of pictures FPS. If Akida can beat nVidias A100 (around 30.000 USD) in different inferencing tasks, that would be a huge win for Brainchip and they could go straight to many datacenters :)


B.t.w. the new nVidia 4090 is able to get close to 1000 FPS generated pictures in some games at lower settings, not that it´s relevant at all :D
 
  • Like
  • Love
Reactions: 7 users

Learning

Learning to the Top 🕵‍♂️
Hi FF, If I didn't know better, I would assume a bit of a Dorithy Dixer was going on here.o_O:)
Has to Google who is Dorithy Dixer to understand your reply. 😅😅😅
chrome_screenshot_1667045764269.png

Learning 🤓😂🤣
 
  • Like
Reactions: 9 users

Chris B

Regular
Ok. First of all... yes, I was very disappointed with the latest 4C report... many investers are getting frustrated, But taking the emotion out of it, I never expected revenue to kick off until the first quarter of 2023. I just had unrealistic expectations after a good 2nd quarter. Everyone has their own strategies and expectations and I respect that. For me... Nothing has changed my opinion for the end game. I still strongly believe that I will reap a fantastic Harvest from what I have Sown. Just my opinion...
 
  • Like
  • Love
Reactions: 13 users

Deadpool

Did someone say KFC
  • Haha
  • Love
  • Like
Reactions: 4 users

M_C

Founding Member
“It took Tesla 17 years to turn a profit when it announced that 2020 was the first full year of profitability in the company's history. While the company generates substantial revenue from automotive sales and regulatory credits, it took some time to profit due to production costs and supply chain issues.8 Sept 2022
View attachment 20527
https://www.forbes.com › 2022/09/08

By The Numbers, How Does Tesla Make Money In 2022?

When I read someone state that the next 4C’s will be ‘make or break’ for Brainchip which at the most extreme assessment of its commercialisation phase has only been at it full time since June, 2021 I have to say based on Tesla what a complete load of old rubbish this statement is and completely miss states both the financial situation of the company and the development time lines for adoption of new technology into products.

So I say beware of wolves in sheep clothing.

What angers me as past evidence proves is long term genuine holders who do not panic and sell because in their hearts and minds they know the opportunity Brainchip presents but yet vent and spew negativity without regard to the impact their statements have on the less informed retail investor looking for facts and reassurance about the fundamentals of their investments in difficult times.

I remember a series of posts I had with @MC🐠 over at the other place about the influence he had on others which he initially did not accept. These posts related to dot joining not being negative by the way.

It is now the weekend and we have the second wave of more articulate manipulators spreading negativity that by itself does not withstand logical analysis but woven into backhanded positive statements will eat away at shareholder confidence.

An emerged poster throwing in a little “I think that is being misleading” comment to attempt to diminish a posts importance or at the very least inject doubt. There are lots of ways to ask for more information about a point which would not have this effect.

So remember Tesla but also take comfort from the fact that Brainchip sells IP using a developing ecosystem so does not have all the infrastructure and labour costs that Tesla had to deal with before reaching profitability but reach it, it did after 17 years.

My opinion only so do not believe a word I or anyone else says until you have DYOR
FF

AKIDA BALLISTA
True. I am a stubborn bastard
 
  • Like
  • Haha
Reactions: 9 users

RobjHunt

Regular
This quarterly was a little less of what was in the last, revenue wise.

Let’s see what happens over the next couple of quarterly’s.

Pantene.
 
  • Like
  • Love
Reactions: 5 users
Thanks to @alwaysgreen @Diogenese , @JDelekto , @FrederikSchack and @RobjHunt

Your input and knowledge is greatly appreciated.

I will leave these two thoughts of Peter van der Made:

1. In a Pitt Street Research interview “We are just getting started.”

2. At the after AGM he said to me “I wish I could have a 100 more years, there is so much more I have to do.”

I have no doubt that the tech world has barely scratched the surface of the AKIDA Technology magic as envisaged by Peter van der Made.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Love
Reactions: 39 users
  • Like
  • Love
  • Haha
Reactions: 7 users

RobjHunt

Regular
  • Like
  • Haha
Reactions: 6 users

M_C

Founding Member
  • Like
  • Love
  • Haha
Reactions: 9 users

JDelekto

Regular
Ok. First of all... yes, I was very disappointed with the latest 4C report... many investers are getting frustrated, But taking the emotion out of it, I never expected revenue to kick off until the first quarter of 2023. I just had unrealistic expectations after a good 2nd quarter. Everyone has their own strategies and expectations and I respect that. For me... Nothing has changed my opinion for the end game. I still strongly believe that I will reap a fantastic Harvest from what I have Sown. Just my opinion...

I tend to think that it will take a little longer. When I began to acquire BrainChip shares in 2020, I planned to hold them for a 5-year minimum, but with the expectation, it would more realistically be a 5 to 10-year investment.

While two different markets, consider consumer VR. What I think is good consumer VR started selling at the beginning of 2016, saw explosive growth in 2021, and continues to grow in popularity. A different niche, and probably can't be as prolific as BrainChip's IP, has still been six years since its introduction, and they're estimating revenues at $25bn US in 2022.

The revenue estimates in the AI arena are much higher, and Akida has only been available commercially for about a year. Customers are still evaluating Akida and are probably trying to curtail their expenditures during the current economic climate. I believe we will slowly see revenue generated from hardware that includes BrainChip's IP from MegaChips and Renesas.

I am personally holding for another three years minimum, as I feel the company has a good head start and reduced competitive risk through its protective patents. I believe that this technology will have a profound impact on consumer goods and services in the future. However, devices that use Akida need to get into the hands of people, just like VR headsets, before we start seeing widespread acceptance and growth.
 
  • Like
  • Love
  • Fire
Reactions: 23 users
Thanks to @alwaysgreen @Diogenese , @JDelekto , @FrederikSchack and @RobjHunt

Your input and knowledge is greatly appreciated.

I will leave these two thoughts of Peter van der Made:

1. In a Pitt Street Research interview “We are just getting started.”

2. At the after AGM he said to me “I wish I could have a 100 more years, there is so much more I have to do.”

I have no doubt that the tech world has barely scratched the surface of the AKIDA Technology magic as envisaged by Peter van der Made.

My opinion only DYOR
FF

AKIDA BALLISTA
I should have also mentioned that last year I found a NASA research paper where a neuromorphic processor running SNN was proposed to run quantum annealing. @ Diogenese will remember it anyway I referred it to Tony Dawe and asked if they were aware of this research and received back an email from Tony which said, “I asked Peter about this and he is aware.’

The depth of what AKIDA can do is like peeling an onion. There are so many layers.

Old technology thinking has little to any relevance.

In a recent Prophesee video they exposed that they are working on an event based sensor using two lenses side by side which will allow continuous real vision to be processed as events but displayed as a conventional video would on a screen but without frame breaks.
The ability to do this required a neuromorphic brain able to fuse the signals.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 23 users
Top Bottom