BRN Discussion Ongoing

[Tothemoon24]

Liking what this young lady has been working on at Tata

Exciting News!

I’m thrilled to share that I’ve just completed an incredible internship with the neuromorphic computing team at Tata Consultancy Services - Research and Innovation, Kolkata, focused on Machine Learning and Generative AI!
Over the past three months, I had the amazing opportunity to dive deep into the world of image reconstruction and compression, and spike neural networks. This experience allowed me to work on cutting-edge projects, leveraging advanced algorithms and models to push the boundaries of what’s possible with generative AI.

Here’s what I’ve been up to:
1. Developed and optimized transformer models for high-quality lossy image reconstruction and compression. Worked on further developing those models for spike neural networks (ongoing).
2. Explored and implemented techniques with swin transformer blocks for lossy image compression.
3. Collaborated with a talented team of professionals, gaining invaluable insights into real-world applications of ML and AI.

A huge thank you to my mentors Sayan Kahali, project supervisor Sounak Dey, research head Arpan Pal and team at TCS for their guidance and support throughout this journey. Your expertise and encouragement made all the difference!

I’m excited to continue exploring the intersection of AI and image processing, and I’m looking forward to the next chapter of my career. If you have any questions about my internship experience or want to connect on similar topics, feel free to reach out!

Here’s to the future of AI and the endless possibilities it holds!

 

[rgupta]

I don't know why Sean has to comment it this way, the present CR is after relationship building in last two years, now some institutional investors will be on brainchip book.
Results will be out soon on that statement but one thing is sure that shorts are down more than 2.5% since then. Is that mean that company was building relationship with the shorters for last 2 years.
I really don't know how much involved the management is but if we see that shorters had taken those shares that will mean Sean is not fit for his job immediately.

 

[Diogenese]

Great post, by HunterD at the place that shan't be named, so I've copied and pasted it here in full, with full credit to him/her.


"I’ve seen a lot of comments lately about BrainChip’s lack of revenue and how it’s supposedly hurting the stock price. Let’s take a step back and look at the bigger picture.

BrainChip is in a crucial phase of its business cycle. They’ve developed groundbreaking AI technology with substantial patents backing it. The focus right now isn’t 100% on generating immediate revenue—it’s on solidifying deals that will unlock long-term value. The company has been in talks with potential partners, and while we haven’t seen a deal finalised yet, the groundwork is being laid. When these deals are signed, it’s going to be BIG!

For companies like BrainChip, the revenue cycle diverges from that of traditional industries. The emphasis initially lies on securing licensing deals, partnerships, and strategic alliances, which are critical for establishing future revenue streams. This stage focuses on validating the technology, penetrating the market, and building a robust partner network. Once these foundational elements are solidified, the revenue will begin to flow more steadily, and that's when the true impact on the stock price will become evident.

Call me crazy, maybe I’m seeing it differently but I do feel it’s undervalued right now because the market is waiting for that big announcement. When BrainChip announces its first major deal, the market will react quickly, and those who are already in will likely see significant gains. We’re talking about a company with a unique technology that’s poised to become a standard in the AI space. The market isn’t ignoring this, it’s waiting for confirmation.

BrainChip isn’t just another tech company. They’ve got solid technology and patents that give them a significant edge in the market. These patents protect their innovations and ensure they have a competitive advantage. When deals are finalised and the technology starts being widely adopted, those patents become even more valuable. This isn’t just about today’s revenue, it’s about securing the future of AI technology.

So, instead of worrying about short-term revenue, let’s focus on where BrainChip is heading. We know the company is actively working on deals, and when these come through, we’ll see the stock price reflect the true value of the technology and its potential. Stay patient and watch how the market responds once everything falls into place.

My opinion only.. DYOR"

Hi DB,

Thanks for reposting this. I agree with Hunter's sentiments.

Akida is the biggest thing to happen to electronics since the invention of the integrated circuit by Kilby and Noyce circa 1958.

Certainly there have been attempts to make analog SNNs using capacitors or ReRAM/MemRistors, but the manufacturing variability frustrated these attempts.

I think PvdM's genius was the realization that digital was not subject to the effects of manufacturing variation and that digital was the practical solution to SNNs. I'd imagine digital's tolerance for this variation is an order of magnitude or more greater than analog. A bit like Edison's lightbulb invention, the lead seal between the wires and the glass to contain the inert atmosphere was the final tweak which made a practicable device.

And now the tech is evolving at an unprecedented rate which makes it impossible for customers commit to silicon. hence my suspicion that EAPs such as MB, Valeo, etc are using Akida 2 simulation software for signal processing. In particular, TeNNs is ideally suited for moving image and speech analysis as the temporal element is incorporated in the silicon process in a streamlined manner, relieving the associated CPU of the task. Even in software simulation, I imagine TeNNs provides significant advantages in power consumption and latency, but we have not seen any performance figures to support my conjecture.

I'm hoping that TeNNs development will soon reach a plateau of development sufficient for customers to commit to silicon, but I do not wish to see a premature halt to development. However, the next round of development may depend on feedback from the cloud-based FPGA.

Now who was it recently suggested measuring the string from the centre to the end?

 

[CHIPS]

I hate how this SP reacts or is even manipulated.
This morning I was so happy to see that BRN had an SP increase of more than 7% in Germany.
Only a short while later it was and still is a minus of 1.22%

Wake me up once we are at +50%, please.

 

[7für7]

I don't know why Sean has to comment it this way, the present CR is after relationship building in last two years, now some institutional investors will be on brainchip book.
Results will be out soon on that statement but one thing is sure that shorts are down more than 2.5% since then. Is that mean that company was building relationship with the shorters for last 2 years.
I really don't know how much involved the management is but if we see that shorters had taken those shares that will mean Sean is not fit for his job immediately.

It's a bit far-fetched and quite risky to accuse the management of such an act, don't you think? Just more gossip and an attempt to elevate yourself as a small investor... And don't forget, the internet is not a lawless space. Be careful with your accusations and demands. Just some friendly advice.

 

[CHIPS]

SiFive Announces New High-performance RISC-V Datacenter Processor for Demanding AI Workloads
SiFive Performance P870-D brings high compute density and scalability to datacenters, vehicles, and embedded systems

Complete article here:

SiFive Announces New High-performance RISC-V Datacenter Processor for Demanding AI Workloads

Design And Reuse - Catalog of IP Cores and Silicon on Chip solutions for IoT, Automotive, Security, RISC-V, AI, ... and Asic Design Platforms and Resources

www.design-reuse.com

 

[Smoothsailing]

FF
After listening to the Quarterly Update it seems that the AKIDA product portfolio now includes:

1. AKD1000 SoC

2. AKD1000 IP

3. AKD1500 Chip

4. AKD1500 IP

5. AKIDA 2.0 IP

6. AKIDA 2.0 TENNS IP

7. TENNS software

8. TENNS Pleiades software

9. VVDN AKIDA Edge Box

10. EDGX-1 Brain

11. (To be released) Unigen AKIDA Ai Cupcake Edge Server

12. (Under Development) Cloud based AKIDA FPGA Development Environment.

13. Models for Noise Cancellation and Keyword Spotting.

14. Optimised models for GENAi applications at the Edge including ASR.

My opinion only DYOR

Fact Finder

 

[Space Cadet]

FF
After listening to the Quarterly Update it seems that the AKIDA product portfolio now includes:

1. AKD1000 SoC

2. AKD1000 IP

3. AKD1500 Chip

4. AKD1500 IP

5. AKIDA 2.0 IP

6. AKIDA 2.0 TENNS IP

7. TENNS software

8. TENNS Pleiades software

9. VVDN AKIDA Edge Box

10. EDGX-1 Brain

11. (To be released) Unigen AKIDA Ai Cupcake Edge Server

12. (Under Development) Cloud based AKIDA FPGA Development Environment.

13. Models for Noise Cancellation and Keyword Spotting.

14. Optimised models for GENAi applications at the Edge including ASR.

My opinion only DYOR

Fact Finder

I probably would have stretched that to say AKD 2.0 E, S & P.

SC

 

[rgupta]

It's a bit far-fetched and quite risky to accuse the management of such an act, don't you think? Just more gossip and an attempt to elevate yourself as a small investor... And don't forget, the internet is not a lawless space. Be careful with your accusations and demands. Just some friendly advice.

You are right but then equally important is the fact how Sean wants to credit for recent CR as relationship building for last 2 years and telling us that new institutions will join brainchip ledger with this CR while actually shorters closed positions of more than 50 million shares in 2 days after CR.
If it proves me small person then Sean is smaller than me when he tries to take credit on wrong end and got caught red handed.
Anyway I still doubt my judgement and said results will be out soon.
But on the other hand it always happen shorters were given shares from CR by almost all the companies. Money power is much more bigger than any other power on this planet.
As a retail investor we get into more trouble when there is no regulation to protect a retail investor and a game between big players can always be hidden from broader market.

 

[Dijon101]

People need to chill out.

Read the history of wifi if you need to understand time lines. (Edit: Sorry what i should have said was, "if you need a bit of perspective ") It took 2 years for a commercial breakthrough for wifi.

from wiki

"The first version of the 802.11 protocol was released in 1997, and provided up to 2 Mbit/s link speeds. This was updated in 1999 with 802.11b to permit 11 Mbit/s link speeds."

"The major commercial breakthrough came with Apple Inc. adopting Wi-Fi for their iBook series of laptops in 1999.[11] It was the first mass consumer product to offer Wi-Fi network connectivity, which was then branded by Apple as AirPort"

 

[manny100]

When Sean addressed TENNs use cases he said (circa 17 mins plus) that TENNs was definitely going to support LLM' s (large language models) on the Edge. He said that this was an undeniable trend in the industry right now as LLM's move outside the Data Centre to the Edge. Sean said they are looking at a very strong offering in that area in the very near future.

 

[Jacob Kiraz]

When Sean addressed TENNs use cases he said (circa 17 mins plus) that TENNs was definitely going to support LLM' s (large language models) on the Edge. He said that this was an undeniable trend in the industry right now as LLM's move outside the Data Centre to the Edge. Sean said they are looking at a very strong offering in that area in the very near future.

Bottom line by AGM 25 deals need to be done, All this and that, Deals done

 

[TECH]

Gidday All..I thought Tony's questions would have been a fair representation of shareholders concerns during the previous quarter..Sean was obviously reading a prepared, controlled response sheet..very robotic..that I don't like, not this far into his 5 year plan..as some others have already mentioned, rather too guarded.

We have the funds, we have the strong patent portfolio, we have the ever growing IP portfilio, we have added the software to the hardware offering, we have and still are bending over backwards to appease the EAP's...it really is OUT OF OUR CONTROL NOW...they, it or whomever will either commit, or just keep kicking tyres, it is frustrating..BUT as 2025 draws closer things may just surprise many of us...I'm still very positive but am honest enough to say I'm not satisfied at this point in our journey...TENN's X TENN's

 

[Learning]

guarded anodyne close to chest response/answer

Hi

Hi Learning,

It's good to see things are progressing according to plan.

However, I feel Sean is still playing his cards close to his chest, due, no doubt, to NDAs, so I felt some responses were guarded or anodyne, scripted rather than spontaneous - responses rather than answers. I would like some of the nitty-gritty, like whether Mercedes and Valeo are planning on using Akida simulation software commercially.

We have a growing portfolio of offerings:

We have IP for:-
A. Akida 1 and all its nodal variations
B. Akida 1500 and all its nodal variations
C. Akida 2 and all its nodal variations
D. TeNNs (I'm guessing it also comes in various multiples)
E. C + D

On top of that EAPs are using Akida 2/TeNNs simulation software for testing and will eventually be able to access the cloud-based FPGA.

The problem with the cloud-based FPGA is that its response time will be swamped by the transmission delay over the internet. Luckily, we have Akida 1 silicon performance figures to show customers the actual latency of Akida.

Akida 1 is used for classification (eg, images/words) which are then passed on to a processor for further processing and analysis. Akida 2 adds the temporal element for video/natural language processing/noise cancelling etc.

Hi Dio,

I totally agree with you. All of Sean's answers are very scripted, and haven't added anything we shareholders don't know.

However, I do like the answer, where he said Brainchip had currently have engagements with all the offerings. So from AKD1000, ADK1500, ADK2000 & TeNNs. Just hopefully a contract or two can be sign soon. To boost investors confidence.

I do think the proprietary algorithm TeNNs would be quicker to implement than Akida silicon. That just me thinking out loud.

Learning

 

[DingoBorat]

I probably would have stretched that to say AKD 2.0 E, S & P.

SC

I think FactFinder, did a pretty good job of stretching it out, as it is ..

Pretty sure AKD1500 uses AKIDA 1.0 IP, for starters..

 

[Cirat]

When Sean addressed TENNs use cases he said (circa 17 mins plus) that TENNs was definitely going to support LLM' s (large language models) on the Edge. He said that this was an undeniable trend in the industry right now as LLM's move outside the Data Centre to the Edge. Sean said they are looking at a very strong offering in that area in the very near future.

Please Disregard and thankyou to Frangipani


GLTA, DYOR, This not financial Advice

 

[Frangipani]

During the Investor Podcast, Sean talks about the transition of Anil (see 6.55 mark onwards) and says, i think, Don Thompson (Distinguished Engineer) ?? is already working with Tony Lewis.
Is this the same Donald Thompson who is presenting at the "AI Hardware and Edge AI Summit" on September 10th 2024. His topic at 1.45pm - 2.10pm is [EFFICIENT MODEL TRAINING & ARCHITECTURE TRACK]: Micro-Prompting LLMs and SLMs for Efficiency and Performance Increase.
I'm may be way off track.

Sorry if already posted or discussed.


GLTA, DYOR, This not financial Advice

Actually, he was referring to Jon(athan) Tapson instead…

 

[Jacob Kiraz]

Gidday All..I thought Tony's questions would have been a fair representation of shareholders concerns during the previous quarter..Sean was obviously reading a prepared, controlled response sheet..very robotic..that I don't like, not this far into his 5 year plan..as some others have already mentioned, rather too guarded.

We have the funds, we have the strong patent portfolio, with have the ever growing IP portfilio, we have added the software to the hardware offering, we have and still are bending over backwards to appease the EAP's...it really is OUT OF OUR CONTROL NOW...they, it or whomever will either commit, or just keep kicking tyres, it is frustrating..BUT as 2025 draws closer things may just surprise many of us...I'm still very positive but am honest enough to say I'm not satisfied at this point in our journey...TENN's X TENN's

Good luck Tech at some point IP's need to be signed or excuses will run out, let's hope we don't get to that point

 

[Cirat]

Actually, he was referring to Jon(athan) Tapson instead…

View attachment 68059

thanks

 

[charles2]

Chip politics US v China

Second thoughts with comments

California Democrats fear US tech firm 'death spiral' with more China curbs

California Democrats are calling on the Biden administration to freeze reported plans to impose fresh restrictions on U.S. technology exports to China, arguing unilateral curbs benefit foreign rivals at the expense of U.S. businesses. Washington has imposed a raft of restrictions on exports of...

www.yahoo.com