davidfitz
Regular
I got an email yesterday with a link and an access code.
website is https://www.votingonline.com.au/BRNAGM23
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
website is https://www.votingonline.com.au/BRNAGM23
I'm looking forward to the 4C and then hopefully a decent announcement before the AGM. Personally holding until December 2025 at an average of 61 cents in my super. 7 years from retirement in 2025 and rule of thumb is super doubles every 7 years. So will re-evaluate then, unless a black swan event of course.
So am happy to hold and ignore the short term noise, heard it all before
Frangipani
Top 20
I would like to believe all of you that have been using ChatGPT for the purpose of dot joining (as well as other things, yet hopefully excluding AI-generating love letters) are by now very much aware of its limitations - and by that I not only mean that its so-called “knowledge cutoff date” was September 2021 or that it can struggle to grasp the subtle nuances of human communication such as sarcasm or humour. (Just imagine how boring this forum would be with merely bots posting…)
Nevertheless, it can’t hurt to be reminded from time to time that - while undoubtedly fascinating and inspirational - ChatGPT’s generated content should be taken with a grain of salt, as the purported “facts“ are at times merely fabricated falsehoods. I am not saying you won’t occasionally unearth hitherto buried treasures; just keep in mind that even OpenAI itself is warning users on its website that “ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers”, something which has been referred to as “hallucinations“. What may sound harmless, can actually have serious consequences in real life, be it harmful medical or legal advice or the potential to destroy an innocent person’s reputation, like in the recent shocking case when “ChatGPT invented a sexual harassment scandal and named a real law prof as the accused (…) even citing a fake Washington Post article as evidence”. Now that is truly scary!
On a lighter note, I wanted to share with you the following witty article about ChatGPT’s capabilities (and shortcomings) as a travel agent - mind you, I have no idea how different the author’s experience would have been with the AI chatbot based on the newly released GPT-4 model - my guess is probably somewhat improved, but certainly still far from perfect.
The last paragraph is the essence of the author’s experiment (do read the whole article, though, if you have the time - trust me, it will magically take your mind off forum quarrels) and is really hammering the message home - at least for the time being. Who knows what the future holds?
„After conversing with the chatbot for two hours, I’m nowhere near finalizing our vacation plans. It has given me some useful information—namely the best time of year to travel and a few hotels and restaurants to consider—but nothing it recommended can be taken as gospel. Will its ability to provide helpful trip-planning services improve exponentially in the years to come? No doubt.
But for now, nothing beats good old-fashioned, human-powered research.“
www.cntraveler.com
Nevertheless, it can’t hurt to be reminded from time to time that - while undoubtedly fascinating and inspirational - ChatGPT’s generated content should be taken with a grain of salt, as the purported “facts“ are at times merely fabricated falsehoods. I am not saying you won’t occasionally unearth hitherto buried treasures; just keep in mind that even OpenAI itself is warning users on its website that “ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers”, something which has been referred to as “hallucinations“. What may sound harmless, can actually have serious consequences in real life, be it harmful medical or legal advice or the potential to destroy an innocent person’s reputation, like in the recent shocking case when “ChatGPT invented a sexual harassment scandal and named a real law prof as the accused (…) even citing a fake Washington Post article as evidence”. Now that is truly scary!
On a lighter note, I wanted to share with you the following witty article about ChatGPT’s capabilities (and shortcomings) as a travel agent - mind you, I have no idea how different the author’s experience would have been with the AI chatbot based on the newly released GPT-4 model - my guess is probably somewhat improved, but certainly still far from perfect.
The last paragraph is the essence of the author’s experiment (do read the whole article, though, if you have the time - trust me, it will magically take your mind off forum quarrels) and is really hammering the message home - at least for the time being. Who knows what the future holds?
„After conversing with the chatbot for two hours, I’m nowhere near finalizing our vacation plans. It has given me some useful information—namely the best time of year to travel and a few hotels and restaurants to consider—but nothing it recommended can be taken as gospel. Will its ability to provide helpful trip-planning services improve exponentially in the years to come? No doubt.
But for now, nothing beats good old-fashioned, human-powered research.“
I Let ChatGPT Plan My Vacation—Here’s What It Got Right and Oh-So Wrong
The buzzy chatbot's lack of specificity—and factual errors—can be exhausting when it comes to travel recommendations.
Deena
Regular
I think you are barking up the wrong tree Newk. You see management don't control the share price. The questions you should be asking are; Is management making the right connections with new and existing customers and partners? YES! Are they meeting performance goals with regard to product development and use of those products with appropriate developmental software and tools? Another big YES! Mmmm ... In fact I am thinking they are doing an outstanding job here.
Are they getting out and spruiking our IP and chip use at conferences and field days, seminars etc? A big YES again.
There is an old saying: Energy flows where attention goes. If you focus on all the negatives, then that will grow in importance to you. That circle gets bigger. If you focus on the progress that is being made with customers and partners you would be very encouraged. All be it that the company is bound by NDAs which are important to our customers and partners as well as ourselves.
As for someone's suggestion that the company should ditch all NDAs. Yeah, right. If you want to lose 90% of our customers who insist on the NDAs in order to maintain secrecy and stay ahead of their competition. How stupid a suggestion is that? Clearly from someone with no business acumen at all. I'm sure glad they are not running our business.
It is about time some posters got real instead of making ridiculous suggestions to feed into their negative sentiments. (Or is there some other underlying motive?)
Another interesting point; I remember some years ago my brother's broker said "Don't invest in Brainchip because they are not big enough and companies like Intel have deep pockets and will simply poach all the best talent from the company by offering them more money." (Or words to that effect.) How do you think we'd go if we weren't offering performance rights to our top personnel? Do you get the picture? By the way, I don't think share price is one of those performance indicators.
In summary:
1. This company is in sound hands with many of the top personnel and best brains in the industry.
2. Almost weekly new partnerships and customers are popping up. Well done and thank you to all the great researchers here who give their time so freely to keep us ahead of other less informed investors. In fact far better than many of the so called analysts who only seem to analyse their own navels ... and don't do that very well.
3. Our list of partners, customers, and developers continues to expand along with new industries that we are evolving to.
4. Performance goals are outstanding as we churn out the Akida 1000; 1500; 2000 ... with many more on the drawing board and in development. These are all technical masterpieces.
5. We are well cashed up to continue the development until revenue exceeds expenditure.
6. Our ground-breaking technology is protected by a growing list of patents and copyright registrations.
7. More staff are constantly being recruited to meet the growing customer demand (while other big tech firms are laying off thousands of personnel). That gives me lots of confidence
8. And for those shareholders who don't seem to be getting enough feedback from management I have a few suggestions for you:
a. Get off your arse and look at the company web site. Just about everything you want to know (that the company is allowed to tell you) is there.
b. If there is something specific you need to know then ring or send an email to our investor relations person Tony. It's not that hard you know.
c. Stop whingeing here over and over, and crapping all over our wonderful research exchange.
9. We KNOW what's coming and we know it takes time for the commercial products to hit the streets so remember, patience is a virtue. (Lord give me patience ... right bloody now!) LOL
And on a final note. I will be attending the AGM in Sydney where I hope to meet a lot of other LT shareholders. Stay positive everyone. You will live longer, and be happier.
Cheers
Deena
Last edited:
jtardif999
Regular
Then you didn’t see where BRN announced compatibility in integrating with the M85?
McHale
Regular
I just took a look at this vid link, which Walhik Joe put up over at HC.
It's an interview with Nandan Nayampally, I don't believe I have seen it here, what I liked was the way Nandan articulates what AIoT is and a bit of an exploration of AI market.
I like the well considered way Nandan comes across.
Something I would like BRN executives focus upon more in their various interviews, is the one or two shot on chip learning capability of the various Akida iterations i.e. would have been good to see Sean slip the one shot learning thing somewhere in the conversation he had with Zach Shelby the other day.
One shot learning is a major differentiator, Akida is way more than an accelerator.
The link: https://lnkd.in/eRXqqdcx
It's an interview with Nandan Nayampally, I don't believe I have seen it here, what I liked was the way Nandan articulates what AIoT is and a bit of an exploration of AI market.
I like the well considered way Nandan comes across.
Something I would like BRN executives focus upon more in their various interviews, is the one or two shot on chip learning capability of the various Akida iterations i.e. would have been good to see Sean slip the one shot learning thing somewhere in the conversation he had with Zach Shelby the other day.
One shot learning is a major differentiator, Akida is way more than an accelerator.
The link: https://lnkd.in/eRXqqdcx
JDelekto
Regular
I do not consider Akida to be an accelerator. From my understanding, AI processors termed "accelerators" improve the speed at which large numbers of matrix computations are done for inferencing. While Akida can perform these classifications at a very high rate of speed, it is due to the nature of the neuromorphic architecture despite the efforts to optimize traditional convolutional neural networks.
While NVISO has found that Akida can perform over 1000 FPS when inferencing, I would love to see some performance metrics on the "one or more" shot learning. The enormous power requirements and costs incurred today are spent training the various models used by these AI processors, typically using large GPU or dedicated AI hardware in the cloud.
Akida has the benefit of being able to train or update its models in the field. I would be curious to know what type of time, power, and costs would be involved in creating a small farm of Akida processors used to take sensor input (video, audio, text, etc.) and build these models as opposed to current means. Perhaps this is one area that BrainChip could focus on improving, as reducing these costs would impact the AI industry as a whole.
As a separate fun thought, I imagine a robot with two subsystems using Akida processors: a passive system that takes different sensor inputs to generate models and another active motor system that uses these models to react to this input. I would like to see the rate at which it could learn versus the latency at which it reacts based on what it has learned.
Sirod69
bavarian girl ;-)
Magnus Östberg
Hello from China to all auto enthusiasts - always great to be here. I have been busy at #AutoShanghai meeting with partners, participating in valuable interviews and, of course, visiting the trade fair to personally experience all the latest technology in the world of mobility.
It is our aim at #MercedesBenz to create an unsurpassed in-car digital experience that integrates all vehicle functions with our own MB.OS operating system. We must deliver many high-quality software products quickly to remain ahead of our competition.
Therefore, developer productivity has become a leading priority as we transition to software-first vehicles. The first step is giving our developers tools that help them improve agile efficiencies, including the ability to quickly pivot direction based on consumer demand and our product review processes.
This dedication to regional software development is made clear by our new Localization and R&D Centre right here in #Shanghai, which specializes in smart car features and automated driving. Here we are building dedicated software with real-time 3D content and an enriched user interface (UI) for automated driving functions based on Alibaba’s AMAP data, including lane-level map views specific to our customers here in China.
A precursor to our upcoming MB.OS will also be available early next year in China. It is a rich entertainment ecosystem, including an integrated MBUX API that makes it easy to directly install Android apps. Dedicated apps will include content from Tencent Video, WeChat, QQ Music and Ximalaya internet radio, with many more to come.
To further enable our digital future, we will also be introducing Abacus, a scenario-based AI integration platform that provides a highly customized experience to meet the demands of localized customers. And we are just getting started!
Sirod69
bavarian girl ;-)
Christopher Seidl
Looking forward to my Arm TechTalk with Ture Nielsen about Renesas Electronics DA16200. It's true ultra-low power and comes with #CMSIS drivers for easy integration into your next #embedded project.
cassip
Regular
Does he refer to this website?
Magnus Östberg
Hello from China to all auto enthusiasts - always great to be here. I have been busy at #AutoShanghai meeting with partners, participating in valuable interviews and, of course, visiting the trade fair to personally experience all the latest technology in the world of mobility.
It is our aim at #MercedesBenz to create an unsurpassed in-car digital experience that integrates all vehicle functions with our own MB.OS operating system. We must deliver many high-quality software products quickly to remain ahead of our competition.
Therefore, developer productivity has become a leading priority as we transition to software-first vehicles. The first step is giving our developers tools that help them improve agile efficiencies, including the ability to quickly pivot direction based on consumer demand and our product review processes.
This dedication to regional software development is made clear by our new Localization and R&D Centre right here in #Shanghai, which specializes in smart car features and automated driving. Here we are building dedicated software with real-time 3D content and an enriched user interface (UI) for automated driving functions based on Alibaba’s AMAP data, including lane-level map views specific to our customers here in China.
A precursor to our upcoming MB.OS will also be available early next year in China. It is a rich entertainment ecosystem, including an integrated MBUX API that makes it easy to directly install Android apps. Dedicated apps will include content from Tencent Video, WeChat, QQ Music and Ximalaya internet radio, with many more to come.
To further enable our digital future, we will also be introducing Abacus, a scenario-based AI integration platform that provides a highly customized experience to meet the demands of localized customers. And we are just getting started!
Abacus.AI - About
Automate all your work using this all-in-one AI platform. Our general purpose agent, Abacus AI Deep Agent will create apps, videos, presentations and workflows
Sirod69
bavarian girl ;-)
Does he refer to this website?
Abacus.AI - About
Automate all your work using this all-in-one AI platform. Our general purpose agent, Abacus AI Deep Agent will create apps, videos, presentations and workflowsabacus.ai
No, I can't find anything from AbacusAI, not even on LinkedIn. I don't think he means the company.
This is the Link:
#autoshanghai #mercedesbenz #shanghai #leadincarsoftware #automotivesoftware | Magnus Östberg
Hello from China to all auto enthusiasts - always great to be here. I have been busy at #AutoShanghai meeting with partners, participating in valuable interviews and, of course, visiting the trade fair to personally experience all the latest technology in the world of mobility. It is our aim...
Last edited:
Sirod69
bavarian girl ;-)
A low cost neuromorphic learning engine based on a high performance supervised SNN learning algorithm
- Article
- Open Access
- Published: 18 April 2023
......
SNN-based neuromorphic architectures
Most of the work related to SNNs is for Von Neumann systems only. Only a few researchers have focused on building dedicated hardware systems using SNNs. The most recent example is Akida, developed by BrainChip that contains 1.2 million neurons45. Akida is a neural chip that has the capability to process and analyze images, sound, and data patterns. Akida, like many other neuromorphic processors, works in an event-driven manner.
A low cost neuromorphic learning engine based on a high performance supervised SNN learning algorithm - Scientific Reports
Spiking neural networks (SNNs) are more energy- and resource-efficient than artificial neural networks (ANNs). However, supervised SNN learning is a challenging task due to non-differentiability of spikes and computation of complex terms. Moreover, the design of SNN learning engines is not an...
Stable Genius
Regular
No, I can't find anything from AbacusAI, not even on LinkedIn. I don't think he means the company.
This is the Link:
#autoshanghai #mercedesbenz #shanghai #leadincarsoftware #automotivesoftware | Magnus Östberg
Hello from China to all auto enthusiasts - always great to be here. I have been busy at #AutoShanghai meeting with partners, participating in valuable interviews and, of course, visiting the trade fair to personally experience all the latest technology in the world of mobility. It is our aim...www.linkedin.com
The Abacus website is quite interesting. They do use NN although I haven’t found any references to SNN they talk about RNN (recurrent NN).
They have some big names involved, ex AWS, ex Google CEO and Chairman etc.
My understanding of what I have read is they are similar to Edge Impulse by providing a training service. Discusses transferring learnt data, which could come in handy when learning from driving situations.
If it’s not them they could be quite relevant down the track. I don’t see them as a threat to Akida as I haven’t found them proving the hardware.
It looks as though they could be beneficial in training and updating our hardware which could increase our uptake.
Happy to be corrected by those more knowledgeable as I’ve only had a 15 min look at them.
GazDix
Regular
Can't be arsed worrying too much about SP at this stage. Been on a downtrend for so long and I am boarding amongst my average price now it just feels normal. I bet many retail have sold or are nearly down like me and thinking about breaking even and looking at the next shiny thing.
All I know is that we are a new tech/AI stock and volatility is normal. The pump to $2 plus was to cleanse the fairweather holders and get more institutional money in. As is always the case.
This money playing with us now has all the patience in the world and will bugger every weak hand there is if there is value. Why? It pumped to $2 and above 15 months ago and has gone down over 76% or so. The same thing happened in 2021. From touching $1 back to 30 cents or so. I am not technical analyst but ho hum on a 'meme' 'penny' share.
We are either all insane to keep holding for this long or genius. Time will tell. But my timeframe is different from yours and we all have different ones.
All I know is that we are a new tech/AI stock and volatility is normal. The pump to $2 plus was to cleanse the fairweather holders and get more institutional money in. As is always the case.
This money playing with us now has all the patience in the world and will bugger every weak hand there is if there is value. Why? It pumped to $2 and above 15 months ago and has gone down over 76% or so. The same thing happened in 2021. From touching $1 back to 30 cents or so. I am not technical analyst but ho hum on a 'meme' 'penny' share.
We are either all insane to keep holding for this long or genius. Time will tell. But my timeframe is different from yours and we all have different ones.
Thanks Diogenese, doesn't that mean they will need to sign an IP contract first? (Unless it's somehow flowing through MegaChips or Renasas)
alwaysgreen
Top 20
It has been discussed that Valeo and Brainchip are in partnership so potentially, they don't need to sign an IP agreement but once sales commence, we should just see revenue flowing through.
I hope this is the case.
Tothemoon24
Top 20
Great to see AI / machine learning, Reaching the approval of the FDA today .
Helping to save lives that’s bloody cool .
RapidAI Receives First and Only FDA 510(k) Clearance of Non-Contrast CT Imaging Product to Accelerate Acute Stroke Triage
MACHINE LEARNINGNEWS
By Business WireOn Apr 19, 2023Share
Part of RapidAI’s industry-leading stroke solution, Rapid NCCT Stroke supports faster treatment and transfer decisions while driving greater equity of care
RapidAI, the global leader in neurovascular and vascular AI-enhanced clinical decision support and patient workflow, announced it has received FDA 510(k) clearance for Rapid NCCT Stroke – a major addition to the RapidAI suite of non-contrast based solutions for stroke and trauma care, and the first and only FDA cleared medical device to detect suspected intracranial hemorrhage (ICH) and large vessel occlusion (LVO) from value-based CT imaging.AiThority Interview Insights: AiThority Interview with at Brian Sathianathan, Co-Founder and CTO at Iterate.ai
Rapid NCCT Stroke uses artificial intelligence to analyze non-contrast CT (NCCT) images to determine suspicion of ICH and LVO of the distal internal carotid artery (ICA) and middle cerebral artery (MCA-M1). The fully automated system then delivers triage and prioritization notifications through PACS, email, and the Rapid mobile app. Because NCCT imaging is readily available and is the initial imaging modality used for stroke and trauma patients, care teams can make time-sensitive workflow and transfer decisions faster. For hospitals that do advanced imaging, Rapid NCCT Stroke may significantly reduce the time between CT and CTA scans.
Read More about AiThority Interview: AiThority Interview with Thomas Kriebernegg, Managing Director & Co-Founder at App Radar
Rapid NCCT Stroke supports teams in:
- Accelerating door-to-imaging and door-to-decision time
- Making faster transfer decisions
- Improving equity of care by giving centers of all levels greater access to advanced value-based CT imaging technology that may improve decision making and impact patient outcomes
Sirod69
bavarian girl ;-)
Development and Validation of ADAS/AD Features Using MATLAB Solutions
Virtual Speaking Session at MATLAB Expo 2023- Wed, May 10 | 11:30 pm - 12:00 am (CDT)
- Thu, May 11 | 4:00 am - 4:30 am (CDT)
- Thu, May 11 | 11:00 am - 11:30 am (CDT)
Join any of the above sessions where Dr. Khan will focus on:
- Sensor modeling, scenario construction, environment modeling, control algorithm development, vehicle dynamics modeling, closed-loop integration, and validation.
- Demonstrations of the interoperability of MATLAB solutions with 3D simulation tools for better visualization.
- How ISO safety of the intended functionality can be validated in a virtual space using MathWorks solutions?
- MathWorks solutions for ADAS/AD hardware-in-the-loop simulation validation for electronic control units.
TATA ELXSI at MATLAB EXPO'23
Learn from Tata Elxsi expert about the benefits of simulating AD/ADAS algorithms using MATLAB® and Simulink® products.
asset.tataelxsi.com
Similar threads
