supersonic001
Regular
BrainChip on LinkedIn: University of Oklahoma Joins the BrainChip University AI Accelerator…
University of Oklahoma Joins the BrainChip University AI Accelerator Program https://lnkd.in/gvTX9jDT
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Fullmoonfever
Top 20
@Diogenese
Thoughts if you get any time or maybe previously covered? TIA
patents.justia.com
Thoughts if you get any time or maybe previously covered? TIA
US Patent Application for FIELD PROGRAMMABLE GATE ARRAY (FPGA) BASED NEUROMORPHIC COMPUTING ARCHITECTURE Patent Application (Application #20230122192 issued April 20, 2023) - Justia Patents Search
This disclosure relates generally to a method and a system for computing using a field programmable gate array (FPGA) neuromorphic architecture. Implementing energy efficient Artificial Intelligence (AI) applications at power constrained environment/devices is challenging due to huge energy...
patents.justia.com
FIELD PROGRAMMABLE GATE ARRAY (FPGA) BASED NEUROMORPHIC COMPUTING ARCHITECTURE
Mar 2, 2022 - Tata Consultancy Services Limited
Frangipani
Regular
Brainchip getting a mention in this interview with SmartSense by Digi’s Gil Dror:
“So neuromorphic chips are basically trying to take that same concept and apply it to IoT. Now there are already companies in that space like Brainchip that have proven that this is more than just a theory.”
www.iotforall.com
Transcript from 17:53 min onwards:
“[Gil] We’ve also seen some innovation with things like TinyML where they’re trying to take that concept and actually bring it down to battery-powered devices, which I think is great because we’re gonna be basically be extending that possibility and making these small devices even more intelligent. But that- but still that’s still limited to the execution of the model. It’s not really taking data and training a model on the spot. That’s where I think the next jump is gonna be, where we get to the point where edge devices can take input and train models directly on the spot without having to send the data anywhere, just doing that computation on the spot. Think about the potential of that.
Not just the time to action that you have from being able to absorb new information added to the model and actually take action immediately almost in the blink of an eye for a user. But also think about the computing capability that you have. Now, if you have a network of IoT devices, instead of trying to centralize all the processing on the backend, the processing is gonna happen where the data is. And then be sent to the cloud after that. So it is just- it is a different paradigm. I don’t think we could do it necessarily with the current chip architecture that we have. But that’s a different story.
– [Ryan] One thing that came up in prior to us chatting was the term neuromorphic computing and that was something new to me. Can you just as to wrap this up just tell our audience what that is.
– [Gil] Yeah, absolutely. If you think about your typical chipset, it hasn’t changed in eons. It’s still the same binary, one, zero. And if you think about machine learning and their capabilities, they’re basically relying on GPUs, right? So graphic processing, which can process a lot of things at the same time.
Lots of RAM, lots of power. The problem with these ingredients is they don’t scale down very easily. They scale up very easily, but they don’t really scale down very easily. So it’s very difficult to take that and put it on a battery powered device that’s sitting in the middle of a desert.
What neuromorphic chip design is trying to do, and by no means, I’m not an expert in that, more of a hobbyist, but what they’re trying to do is really mimic how the brain works and try to convert that into a chip that can execute commands. If you think about the brain as a computing unit, it’s actually very effective.
Because it uses a network, almost like a graph database to execute very complex computations, with very little battery, very little calorie consumption. So neuromorphic chips are basically trying to take that same concept and apply it to IoT. Now there are already companies in that space, like BrainChip that have proven that this is more than just a theory. You could basically take that and execute it.
– [Ryan] Yeah, we actually spoke to somebody from Brainchip a number of weeks ago, yeah, it’s a very interesting space for sure. But yeah, thank you for coming on and talking about a lot of these different topics, the security side, the LLMs, AI, all that kind of stuff that’s going on.
“So neuromorphic chips are basically trying to take that same concept and apply it to IoT. Now there are already companies in that space like Brainchip that have proven that this is more than just a theory.”
Internet of Things Security: What To Know
With 15 billion connected devices, IoT security isn't easy. Gil Dror joins the podcast to discuss Internet of Things security.
www.iotforall.com
Transcript from 17:53 min onwards:
“[Gil] We’ve also seen some innovation with things like TinyML where they’re trying to take that concept and actually bring it down to battery-powered devices, which I think is great because we’re gonna be basically be extending that possibility and making these small devices even more intelligent. But that- but still that’s still limited to the execution of the model. It’s not really taking data and training a model on the spot. That’s where I think the next jump is gonna be, where we get to the point where edge devices can take input and train models directly on the spot without having to send the data anywhere, just doing that computation on the spot. Think about the potential of that.
Not just the time to action that you have from being able to absorb new information added to the model and actually take action immediately almost in the blink of an eye for a user. But also think about the computing capability that you have. Now, if you have a network of IoT devices, instead of trying to centralize all the processing on the backend, the processing is gonna happen where the data is. And then be sent to the cloud after that. So it is just- it is a different paradigm. I don’t think we could do it necessarily with the current chip architecture that we have. But that’s a different story.
– [Ryan] One thing that came up in prior to us chatting was the term neuromorphic computing and that was something new to me. Can you just as to wrap this up just tell our audience what that is.
– [Gil] Yeah, absolutely. If you think about your typical chipset, it hasn’t changed in eons. It’s still the same binary, one, zero. And if you think about machine learning and their capabilities, they’re basically relying on GPUs, right? So graphic processing, which can process a lot of things at the same time.
Lots of RAM, lots of power. The problem with these ingredients is they don’t scale down very easily. They scale up very easily, but they don’t really scale down very easily. So it’s very difficult to take that and put it on a battery powered device that’s sitting in the middle of a desert.
What neuromorphic chip design is trying to do, and by no means, I’m not an expert in that, more of a hobbyist, but what they’re trying to do is really mimic how the brain works and try to convert that into a chip that can execute commands. If you think about the brain as a computing unit, it’s actually very effective.
Because it uses a network, almost like a graph database to execute very complex computations, with very little battery, very little calorie consumption. So neuromorphic chips are basically trying to take that same concept and apply it to IoT. Now there are already companies in that space, like BrainChip that have proven that this is more than just a theory. You could basically take that and execute it.
– [Ryan] Yeah, we actually spoke to somebody from Brainchip a number of weeks ago, yeah, it’s a very interesting space for sure. But yeah, thank you for coming on and talking about a lot of these different topics, the security side, the LLMs, AI, all that kind of stuff that’s going on.
Tothemoon24
Top 20
Good Morning Chippers,
Markets looking Bullish today.
Dow +1.28%
Nas +1.55%
Sox +1.55%
Vix. -12.76%
BCHPY , $11.51 USD, +0.17% on volume of 3220 shares , volume at 1,843% above daily average.
* Still absolutely NOTHING VOLUME
transacting , though on volume chart looks positive
. With any luck start of a reversal with solid upside.
USA, lifting of US debt ceiling passes House , ( 314 for / 117 against ) now onto the Senate for final clearance.
Also, should see above volume for Brainchip stock on our market today, ASX 200 rebalance, all shall be revealed after 5:00 pm.
.
Looking good.
Regards,
Esq.
Markets looking Bullish today.
Dow +1.28%
Nas +1.55%
Sox +1.55%
Vix. -12.76%
BCHPY , $11.51 USD, +0.17% on volume of 3220 shares , volume at 1,843% above daily average. * Still absolutely NOTHING VOLUME
transacting , though on volume chart looks positive
. With any luck start of a reversal with solid upside.USA, lifting of US debt ceiling passes House , ( 314 for / 117 against ) now onto the Senate for final clearance.
Also, should see above volume for Brainchip stock on our market today, ASX 200 rebalance, all shall be revealed after 5:00 pm.
. Looking good.
Regards,
Esq.
Last edited:
Cardiff Ocean Group - Donald Jones
Scripps Research Translational Institute USC Center for Body Computing
www.cardiffoceangroup.com
3200 shares of BCHPY is an astronomical amount. Typically 0-300 shares trade each day, being equivalent to 40 shares of those offered on the ASX. Someone in the US market is taking a liking to Brainchip.Good Morning Chippers,
Markets looking Bullish today.
Dow +1.28%
Nas +1.55%
Sox +1.55%
Vix. -12.76%
* Still absolutely NOTHING VOLUME
transacting , though on volume chart looks positive
USA, lifting of US debt ceiling passes House , ( 314 for / 117 against ) now onto the Senate for final clearance.
Also, should see above volume for Brainchip stock on our market today, ASX 200 rebalance, all shall be revealed after 5:00 pm.
Looking good.
Regards,
Esq.
Caveat: Unless those were all sales. At the price stated they may have. In any event extraordinary volume.
SharesForBrekky
Regular
Yeah, second biggest amount in one day since the Mercedes announcement back in Jan 22.
Boab
I wish I could paint like Vincent
$11.51 US = AUD17.50 divided by 40 = .4376c (Lets say 44c)
Cyw
Regular
Beware, shorters!
Traders betting against semiconductor stocks are taking a beating this year, as increased interest in artificial intelligence fuelled a blistering rally in some sector names.
Paper losses for semiconductor short sellers have topped $US18 billion ($27 billion) in 2023, according to data from S3 Partners. Put another way, traders betting against semis have lost 92 cents on every dollar they’ve put into short positions this year.
It’s no surprise, as the Philadelphia Stock Exchange Semiconductor Index is up 39 per cent so far this year, closing at its highest since March 2022 on Tuesday.
Nvidia has been the largest source of pain for those betting against the chipmakers, with mark-to-market losses of $US8.6 billion ($13.1 billion) this year amid the stock’s 170 per cent surge. The rally also prompted the secto’s biggest short-covering (when traders exit their contrarian positions by buying the stock back) in the last 30 days.
Short sellers have also covered large positions in Texas Instruments, ON Semiconductor and Microchip Technology.
“These stocks have been bucking broncos on the upside, so it can be intimidating to try to short them, betting on a reversal,” said Alec Young, chief investment strategist at quant-research firm Mapsignals.
That said, short sellers are far from being discouraged. In the last 30 days, traders have added nearly $US1.4 billion in bets against the sector, with the largest increases in short selling in Advanced Micro Devices, Intel, Marvell Technology and Lattice Semiconductor.
Short sellers losing big in bets against chipmakers
Traders betting against semiconductor stocks are taking a beating this year, as increased interest in artificial intelligence fuelled a blistering rally in some sector names.
Paper losses for semiconductor short sellers have topped $US18 billion ($27 billion) in 2023, according to data from S3 Partners. Put another way, traders betting against semis have lost 92 cents on every dollar they’ve put into short positions this year.
It’s no surprise, as the Philadelphia Stock Exchange Semiconductor Index is up 39 per cent so far this year, closing at its highest since March 2022 on Tuesday.
Nvidia has been the largest source of pain for those betting against the chipmakers, with mark-to-market losses of $US8.6 billion ($13.1 billion) this year amid the stock’s 170 per cent surge. The rally also prompted the secto’s biggest short-covering (when traders exit their contrarian positions by buying the stock back) in the last 30 days.
Short sellers have also covered large positions in Texas Instruments, ON Semiconductor and Microchip Technology.
“These stocks have been bucking broncos on the upside, so it can be intimidating to try to short them, betting on a reversal,” said Alec Young, chief investment strategist at quant-research firm Mapsignals.
That said, short sellers are far from being discouraged. In the last 30 days, traders have added nearly $US1.4 billion in bets against the sector, with the largest increases in short selling in Advanced Micro Devices, Intel, Marvell Technology and Lattice Semiconductor.
HopalongPetrovski
I'm Spartacus!
buena suerte :-)
BOB Bank of Brainchip
Those buggers are just not letting us go
!! They keep loading up the sell side and dropping the SP again 
HopalongPetrovski
I'm Spartacus!
Sooner or later there will come an announcement or other news that will lend us such impetus that they will have no option but to jump on board. ala the MB news we all remember so fondly from January '22.Those buggers are just not letting us go
The more they have invested in pushing us down at the time, the greater the reverse thrust they will have to apply to try and salvage their position.
This time, I'm ready.
Bring it, Brainchip.
Tothemoon24
Top 20
We tend to agree "“If you can run a model, inference-wise, on the device…that’s a huge advantage for us…We have the opportunity to expand the capacity of this low-power engine [developed for smartphones] and apply it to large language models" Akash Palkhiwala Qualcomm
To set the stage, large language model-based solutions like ChatGPT and MidJourney are capturing hearts, minds and dollars given the broad applicability of AI-assisted workflows. Two things to note: running large language models (LLMs) like those used by ChatGPT and MidJourney consumes a lot of cloud computing resources which, in turn, costs a lot of money; and these types of LLMs that ingest enormous amounts of data are not really optimized for specific enterprise use cases. To borrow an anecdote from Michael Dell as articulated earlier this week at Dell Tech World, if a bank wants to use AI for fraud detection, it doesn’t need an AI solution built on an LLM that has consumed hundreds of years of literature.
What does that mean for Qualcomm? “This is a rapidly evolving industry…over the last several months and it’s going to continue to be that over the next year or two,” Palkhiwala said. “If you think about the hyperscalers, there’s a tremendous focus and effort on having large language models running in the cloud, but also reducing the size so you can run it on the device.” Given that divergence in the size of LLMs and Qualcomm’s proven capabilities around on-device AI, that’s important “because you could run the smaller models on the device with very good accuracy and performance which we can take across our ecosystem.”
Key point here is that, for Qualcomm, on-device AI maps to the company’s ongoing strategy to diversify beyond smartphones. As the company looks to grow its consumer and industrial IoT businesses, as well as its booming automotive business, on-device AI tech developed for handsets can be tweaked and ported across all lines of business.
Palkhiwala explained: “From an AI perspective, our view is as large language models come into play, a lot of the inference is going to happen on the device rather than in the cloud…The cost is definitely way cheaper on the device side,” not to mention considerations around data privacy and security, and application-specific latency needs. “If you can run a model, inference-wise, on the device…that’s a huge advantage for us…We have the opportunity to expand the capacity of this low-power engine [developed for smartphones] and apply it to large language models…That’s what creates an advantage for us going forward…It’s something that creates a competitive advantage for us across all edge devices.”
In terms of how that drives new revenue for Qualcomm, Palkhiwala said its current AI engine that’s integrated into its mobile platforms needs to evolve in terms of the size of the engine and memory bandwidth to better support LLMs. As that evolves into different products, it increases per-device content opportunity. Another driver relates to the diversification strategy. Palkhiwala gave the example of Qualcomm’s focus on PCs; as that business ramps, having on-device AI content “improves of the probability of success in those areas.” He said the company is also exploring additional avenues for monetization.
Qualcomm views on-device AI as key to reducing cloud costs associated with AI processing
As bullishness, or at least perceived bullishness, around artificial intelligence (AI) drives up stock prices and market caps for companies like NVIDIA, AMD and Marvell, Qualcomm Technologies CFO Akash Palkhiwala this week gave a look at what these movements mean for the San Diego-based firm. With existing expertise and product for on-device AI, Palkhiwala told investors at the JP Morgan Global Technology, Media and Communications Conference that Qualcomm’s technology roadmap puts it in a “unique place.”To set the stage, large language model-based solutions like ChatGPT and MidJourney are capturing hearts, minds and dollars given the broad applicability of AI-assisted workflows. Two things to note: running large language models (LLMs) like those used by ChatGPT and MidJourney consumes a lot of cloud computing resources which, in turn, costs a lot of money; and these types of LLMs that ingest enormous amounts of data are not really optimized for specific enterprise use cases. To borrow an anecdote from Michael Dell as articulated earlier this week at Dell Tech World, if a bank wants to use AI for fraud detection, it doesn’t need an AI solution built on an LLM that has consumed hundreds of years of literature.
What does that mean for Qualcomm? “This is a rapidly evolving industry…over the last several months and it’s going to continue to be that over the next year or two,” Palkhiwala said. “If you think about the hyperscalers, there’s a tremendous focus and effort on having large language models running in the cloud, but also reducing the size so you can run it on the device.” Given that divergence in the size of LLMs and Qualcomm’s proven capabilities around on-device AI, that’s important “because you could run the smaller models on the device with very good accuracy and performance which we can take across our ecosystem.”
Key point here is that, for Qualcomm, on-device AI maps to the company’s ongoing strategy to diversify beyond smartphones. As the company looks to grow its consumer and industrial IoT businesses, as well as its booming automotive business, on-device AI tech developed for handsets can be tweaked and ported across all lines of business.
Palkhiwala explained: “From an AI perspective, our view is as large language models come into play, a lot of the inference is going to happen on the device rather than in the cloud…The cost is definitely way cheaper on the device side,” not to mention considerations around data privacy and security, and application-specific latency needs. “If you can run a model, inference-wise, on the device…that’s a huge advantage for us…We have the opportunity to expand the capacity of this low-power engine [developed for smartphones] and apply it to large language models…That’s what creates an advantage for us going forward…It’s something that creates a competitive advantage for us across all edge devices.”
In terms of how that drives new revenue for Qualcomm, Palkhiwala said its current AI engine that’s integrated into its mobile platforms needs to evolve in terms of the size of the engine and memory bandwidth to better support LLMs. As that evolves into different products, it increases per-device content opportunity. Another driver relates to the diversification strategy. Palkhiwala gave the example of Qualcomm’s focus on PCs; as that business ramps, having on-device AI content “improves of the probability of success in those areas.” He said the company is also exploring additional avenues for monetization.
buena suerte :-)
BOB Bank of Brainchip
Absolutely agree Hoppy ... 'Just one power announcement'Sooner or later there will come an announcement or other news that will lend us such impetus that they will have no option but to jump on board. ala the MB news we all remember so fondly from January '22.
The more they have invested in pushing us down at the time, the greater the reverse thrust they will have to apply to try and salvage their position.
This time, I'm ready.
Bring it, Brainchip.
Damo4
Regular
S&P/ASX 200 and 300
For anyone interested in re-balances see below:
S&P/ASX Australian Indices Methodology
Looks like we might be sitting around 328th ($0.450, this could change), which should mean deletion form both indexes.
Edit: ETF's are included in the below table, which I don't believe should be in there. Could be as high as 270th.
The only unknown is that there has to be a deletion for an addition, and both companies must surpass the buffer as seen below.
Whilst BRN should be deleted, I don't know what happens if there was no-one who could surpass the buffer to be added.
For anyone interested in re-balances see below:
S&P/ASX Australian Indices Methodology
Edit: ETF's are included in the below table, which I don't believe should be in there. Could be as high as 270th.
The only unknown is that there has to be a deletion for an addition, and both companies must surpass the buffer as seen below.
Whilst BRN should be deleted, I don't know what happens if there was no-one who could surpass the buffer to be added.
Last edited:
Similar threads
