BRN Discussion Ongoing

[Proga]

Perhaps Rob should call Sam Altman to remind him of an available resource that isn't nuclear fusion which can help with this issue. Would be the perfect opportunity to discuss the benefits of neuromorphic computing in helping to reduce AI's massive energy consumption.


EXTRACT ONLY

Sam Altman: Age of AI will require an ‘energy breakthrough’​

Speaking at Davos, OpenAI's CEO spoke of a vague AI future made possible only by currently unavailable resources.

BY MACK DEGEURIN | PUBLISHED JAN 18, 2024 2:09 PM EST

• TECHNOLOGY

Sam Altman, chief executive officer of OpenAI, attends the World Economic Forum (WEF) in Davos, Switzerland. Halil Sagirkaya/Anadolu via Getty


Open AI CEO Sam Altman believes long-awaited nuclear fusion may be the silver bullet needed to solve artificial intelligence’s glutinous energy appetite and pave the way for an AI revolution. When that revolution does arrive, however, it might not seem quite as shocking as he once claimed.

Altman touched on AI’s growing demands earlier this week while speaking at a Bloomberg event outside of the annual World Economic Forum meeting in Davos, Switzerland. The CEO said powerful new AI models would likely require even more energy consumption than previously imagined. Solving that energy deficit, he suggested, will require a “breakthrough” in nuclear fusion.

“There’s no way to get there without a breakthrough,” Altman said at the event according to Reuters. “It motivates us to go invest more in [nuclear] fusion.”

AI’s energy problem​

Though some AI proponents believe insights gleaned from advanced models could help fight climate change in novel ways, a growing body of research suggests the up-front energy required to train these complex models is taking a toll of its own. Experts expect the vast amounts of data needed to train models like OpenAI’s GPT and Google’s Bard could increase the global data server industry, which the International Energy Agency (IEA) estimates already accounts for around 2-3% of global greenhouse gas emissions.

Researchers estimate training a single large language model like GPT-4 could use around 300 tons of CO2. Others estimate a single image spit out by AI image generator tools like Dall-E or Stable Diffusion requires the same amount of energy as charging a smartphone. The massive server farms needed to facilitate AI training also require vast amounts of water to stay cool. GPT-3 alone, recent research suggests, may have consumed 185,000 gallons of water during its training period.

Sam Altman: Age of AI will require an 'energy breakthrough'

Speaking at Davos, OpenAI's CEO spoke of a vague AI future made possible only by currently unavailable resources.

www.popsci.com

AI is moving faster than nuclear fusion @Bravo. AI needs to go to the edge

 

[DerAktienDude]

Appreciate your view and I agree rewards before performance did not seam like a great result maybe this is what the prices of these individuals are in the USA for the tech industry.
Question
I have never seen Samsung linked as a partner do you have info on this that can be shared?

Valeo is a interesting one the probability of us in that system is actually pretty high but yes there has been no other comments beyond this then that announce other then the original press release that we are in a joint development partnership. We still advise that we are partners and the ASX announcement would need to be put out if we are not working with them.

Samsung announced two major things. The first is the roadmap for 2024 with ultra low power NPUs included for the first time.
Samsung was even earlier into neuromorphic that in that link I made a mistake the last time posting it.

BrainChip showcases Akida neural processing capabilities, opens developer environment | Biometric Update

Akida is a licensable IP technology that can be embedded into ASIC devices and will be applicable as an integrated SoC for neural processing on the edge.

www.biometricupdate.com

From the Samsung roadmap for 2024:

- The company plans to empower its home appliances with generative AI capabilities, ensuring more intuitive interactions between users and their devices. Consequently, Samsung plans to equip its home appliances with an ultra-low-power Neural Processing Unit (NPU) chipset. This chipset will play a vital role in facilitating the efficient functioning of generative AI, ensuring a harmonious and user-friendly experience.

- The generative AI technology will not be confined to a single facet but will encompass voice recognition, visual processing, and display functionality. “Generative AI technologies will be applied to voice, vision and display,”

-Samsung’s vision is to have home appliances that deeply understand user behaviour and preferences, resulting in context-aware and intelligent responses

- The AI-driven enhancements won’t be confined to a select range of devices; they will be widespread across Samsung’s appliance lineup, covering premium and entry-level products. The same can be seen in Samsung’s operating system, ‘Tizen,’ which will receive substantial upgrades. Tizen, optimized for voice recognition and video processing, will be strengthened with AI capabilities.

- Further, Samsung is engineering an energy-efficient chipset, addressing smart homedevices’ growing complexity and data intensity. The aim is to keep power consumption to a minimum, targeting an impressive goal of less than 0.1 watts of energy utilization over 24 hours for AI-enabled appliances.

Second major thing was their announcement for the new Galaxy.

Samsung Galaxy S24: Unser Live-Ticker zum Nachlesen

Heute findet die Vorstellung der Samsung Galaxy S24-Reihe statt. Hier könnt ihr das Galaxy AI-Event live mitverfolgen.

www.gamestar.de

It’s in German so write “lokal” for the good stuff. Npu was mentioned too along with some other interesting stuff. Just translate into English.

- The issue is. If Samsung is not using us through ARM or whoever else it might be we compete against a company with its own foundries. Then the whole so and so much years lead is gone and partnering with Edge impulse and other start ups isn’t worth anything anymore. It has to be use or we’re into enormous trouble. The lead is gone and we’re competing against one of the biggest chip producers worldwide.

- Some other posts with Mercedes came as a reply. There’s nothing confirmed so far. Edge impulse is nice but a start up founded 5 years ago is 3 steps back from licensing with Renesas and megachips but nice to have nevertheless. Valeo publicly stated that they have orders around the 1 billion €/£/$ mark im not sure which currency. We have not received a penny of preliminary payments so we’re out. We’d at least have seen 10-15% of our share in preliminary payments and they’re surely not around the range of what a kebap store ran by two people make in a month. Let’s hope it’s us that Samsung uses, watch the financials and actually experience a pleasant surprise for once.

 

[DingoBorat]

Nvidia partners with Luminar.

MB dumps Valeo for Luminar.

MB partners with Nvidia

Is there a cart/horse thing going on here?

Luminar has foveated lidar which focusses the laser dots on a point of interest (POI), and this gives it longer range for detecting/classifying POIs.

Luminar use software NNs in processing lidar data:

US11361449B2 Neural network for object detection and tracking 20200506

View attachment 54679

US2022309685A1 NEURAL NETWORK FOR OBJECT DETECTION AND TRACKING


[0014] FIG. 3 is a block diagram of an example software architecture for an autonomous vehicle in which the techniques of this disclosure can be implemented;

[0081] The classification module 112 is generally configured to determine classes (labels, categories, etc.) for different objects that have been identified by the segmentation module 110 .

Now I wouldn't presume to tell Luminar their business, but ...

It may not be the "best" way to skin a cat, but that's what they're doing..

 

[Proga]

"It comes across as you have nothing of quality to counter other than an attack the messenger.. What quality of person does that?"

Even the fact, that you consider yourself a "messenger" delivering unbiased news, is a complete joke.

I don't believe Schnitzel lover is stupid. Only the dingbats who reply to his posts. And the TSE has a lot of dingbats who believe they can win an argument against an idiot.

Seriously. Schnitzel cares more about that you reply, not what you reply. The dingbats need to grow up and stop encouraging him

 

[Makeme 2020]

LOLLLLL... your rubbish posts are what give these threads a flavour of poor in the first place. Why are you even here LMAO.

View attachment 54669

BIN BOY

 

[McHale]

It still does not remove him as CEO.
The Current Board has on its face been happy with his performance and approved his decisions.

Re-election of Directors is a simple majority vote. If they stand again highly likely they will be elected again as 51% can be garnered from the top 1,000 shareholders who hold 70% of the shares on issue.

To remove CEO it would require a majority of new Board members not aligned with the CEO to be elected.

Where will those alternate Directors come from the ranks of HC.

We do not know of any friendly institution or company sitting amongst the ranks of shareholders.

I am sure those who are calling for this vote will continue to do so and offer no answers to these questions.

My opinion only DYOR
Fact Finder

A question here @Fact Finder, I have asked this here previously but am not aware of anyone answering my question, so I ask again.

Do we know who was responsible for the vote against the remuneration report specifically, as I understand it there would not have been enough retail people present at the AGM to support that vote, so my thinking it must have been via a lot of retail proxies.

Usually from what I have observed with regard to remuneration strikes, for retail investors to be coherent enough to support a strike there needs to be widespread dissatisfaction and some activist investors who promote and disseminate info to promote said strike, I am not aware of any such being widely discussed at that time, either here or HC.

Or could it have been an institutional holder(s), who had unknown motives. I have asked you because I believe it has been mentioned that you were present at the meeting.

You also state that the top 1000 shareholders own 70% of SOI, so 51% of votes could be garnered from top 1000, we know PVDM & AM have significant holdings but not enough to cover 51%, and I am curious about the intents of some institutional holders because of a pattern of block voting at AGMs. I have no definitive info to support this whatsoever, but there is a pattern which I illustrate below
2023 remuneration vote below

Further I am now going back to the 2022 AGM where there was a large block of votes which opposed the re-election of PVDM to the Board. This shocked me at the time and I believe others here felt the same way, Peter had been installed as acting CEO after the resignation Lou Di Nardo in March 2021, but stepped down from that role in Nov 2021 when Sean Hehir was appointed to that role.

So again this was another significant block of votes which must have been an organized retail group, or powerful organic retail reaction against PVDM, who to all intents and purposes was a respected Board member and CTO. Or was it an institutional actor?
Below is 2022 vote targeting PVDM

I am also aware that there was another significant block of votes opposed to the remuneration package at the 2021 AGM, so there has been ongoing significant blocks of votes opposing remuneration for some time.
2021 vote below

So if anyone else other than Fact Finder has any thinking about these consistent and significant blocks of votes at our AGMs, I would appreciate feedback you may have. I have no trust whatsoever with regard to the ASX and their total lack of respect or regulation toward the fair treatment of retail market participants, or the absence of any apparent disposition on their part to prevent corrupt and malicious SP manipulation.

 

[Slade]

BIN BOY

For god sake grow up.

 

[DingoBorat]

I don't believe Schnitzel lover is stupid. Only the dingbats who reply to his posts. And the TSE has a lot of dingbats who believe they can win an argument against an idiot.

Seriously. Schnitzel cares more about that you reply, not what you reply. The dingbats need to grow up and stop encouraging him

That's gold Proga
So you think he's not stupid, but still an idiot

Schnitzel lover, isn't an attention seeking troll, in my opinion.

His agenda is simple, whether he is acting on his own accord, or for someone else.

He seeks to sow doubt and fear, for his own or others benefit.
So that people will sell and either he or other party's, can get their orders filled/stock returned.

He will say, that he has no intention of buying, until he sees a change in trend.
Although, by his own accounts, he has already been in and out several times.

The fact that he is not a complete simpleton, is why I reply to "some" of his posts, when he says things that are deliberately misleading, or untrue.

His post about ARM, having a "loss making" business model, which BrainChip had adopted, took me all of a 20 second Google search, to disprove.

He is smart enough to work that out, if he puts in even a "little" delving, before he posts.

But Truth/honesty, is not a good tool for him.

 

[ForrestofHope]

Hi DAD,

Yes, as you correctly point out, Samsung have been interested in developing neuromorphic chips for a while. They make no secret of it. This is from an article dated Feb 2020.

Taken from the article:
Both Samsung Electronics and SK hynix are pouring huge amounts of money into computer chips processing information in the way human brains do….For what it called “on-device artificial intelligence,” Samsung Electronics said in July of last year that it was spurring its research institutes to develop NPUs that were optimized for deep-learning algorithms.

“It said it was investing heavily in the development of NPUs and other system semiconductors, aiming at positioning itself in the No. 1 slot by 2030 when it came the manufacture of system semiconductors. In this regard, it said the number of personnel working on NPUs will increase 10-fold to 2,000 by 2030.

But the ultimate goal of Samsung Electronics, SK hynix and other chip makers is to develop neuromorphic chips. SK hynix says, “By imitating the human brain’s structure, neuromorphic chips can recognize more diverse information such as atypical text, images, sounds and videos, as patterns.”

Samsung Electronics and SK hynix, as well as the Korean government, believe they are viable as global efforts to develop neuromorphic chips are in an initial stage.

The Ultimate Goal Is to Develop Neuromorphic Chips

It will not be Homo sapiens alone that will determine what their future will be like. Artificial intelligence will have a fair share of influence on shaping the future for Homo sapiens as well.That is the belief held by the Electronics and Telecommunications Research Institute (ETRI), which says art

www.wired.kr

 

[DerAktienDude]

Hi DAD,

Yes, as you correctly point out, Samsung have been interested in developing neuromorphic chips for a while. They make no secret of it. This is from an article dated Feb 2020.

Taken from the article:
Both Samsung Electronics and SK hynix are pouring huge amounts of money into computer chips processing information in the way human brains do….For what it called “on-device artificial intelligence,” Samsung Electronics said in July of last year that it was spurring its research institutes to develop NPUs that were optimized for deep-learning algorithms.

“It said it was investing heavily in the development of NPUs and other system semiconductors, aiming at positioning itself in the No. 1 slot by 2030 when it came the manufacture of system semiconductors. In this regard, it said the number of personnel working on NPUs will increase 10-fold to 2,000 by 2030.

But the ultimate goal of Samsung Electronics, SK hynix and other chip makers is to develop neuromorphic chips. SK hynix says, “By imitating the human brain’s structure, neuromorphic chips can recognize more diverse information such as atypical text, images, sounds and videos, as patterns.”

Samsung Electronics and SK hynix, as well as the Korean government, believe they are viable as global efforts to develop neuromorphic chips are in an initial stage.

The Ultimate Goal Is to Develop Neuromorphic Chips

It will not be Homo sapiens alone that will determine what their future will be like. Artificial intelligence will have a fair share of influence on shaping the future for Homo sapiens as well.That is the belief held by the Electronics and Telecommunications Research Institute (ETRI), which says art

www.wired.kr

And that’s us showcasing Akida at Samsung Strategy & Innovation Center in San Jose, California, in 2020

BrainChip showcases Akida neural processing capabilities, opens developer environment | Biometric Update

Akida is a licensable IP technology that can be embedded into ASIC devices and will be applicable as an integrated SoC for neural processing on the edge.

www.biometricupdate.com

 

[Kachoo]

Samsung announced two major things. The first is the roadmap for 2024 with ultra low power NPUs included for the first time.
Samsung was even earlier into neuromorphic that in that link I made a mistake the last time posting it.

BrainChip showcases Akida neural processing capabilities, opens developer environment | Biometric Update

Akida is a licensable IP technology that can be embedded into ASIC devices and will be applicable as an integrated SoC for neural processing on the edge.

www.biometricupdate.com

From the Samsung roadmap for 2024:

- The company plans to empower its home appliances with generative AI capabilities, ensuring more intuitive interactions between users and their devices. Consequently, Samsung plans to equip its home appliances with an ultra-low-power Neural Processing Unit (NPU) chipset. This chipset will play a vital role in facilitating the efficient functioning of generative AI, ensuring a harmonious and user-friendly experience.

- The generative AI technology will not be confined to a single facet but will encompass voice recognition, visual processing, and display functionality. “Generative AI technologies will be applied to voice, vision and display,”

-Samsung’s vision is to have home appliances that deeply understand user behaviour and preferences, resulting in context-aware and intelligent responses

- The AI-driven enhancements won’t be confined to a select range of devices; they will be widespread across Samsung’s appliance lineup, covering premium and entry-level products. The same can be seen in Samsung’s operating system, ‘Tizen,’ which will receive substantial upgrades. Tizen, optimized for voice recognition and video processing, will be strengthened with AI capabilities.

- Further, Samsung is engineering an energy-efficient chipset, addressing smart homedevices’ growing complexity and data intensity. The aim is to keep power consumption to a minimum, targeting an impressive goal of less than 0.1 watts of energy utilization over 24 hours for AI-enabled appliances.

Second major thing was their announcement for the new Galaxy.

Samsung Galaxy S24: Unser Live-Ticker zum Nachlesen

Heute findet die Vorstellung der Samsung Galaxy S24-Reihe statt. Hier könnt ihr das Galaxy AI-Event live mitverfolgen.

www.gamestar.de

It’s in German so write “lokal” for the good stuff. Npu was mentioned too along with some other interesting stuff. Just translate into English.

- The issue is. If Samsung is not using us through ARM or whoever else it might be we compete against a company with its own foundries. Then the whole so and so much years lead is gone and partnering with Edge impulse and other start ups isn’t worth anything anymore. It has to be use or we’re into enormous trouble. The lead is gone and we’re competing against one of the biggest chip producers worldwide.

- Some other posts with Mercedes came as a reply. There’s nothing confirmed so far. Edge impulse is nice but a start up founded 5 years ago is 3 steps back from licensing with Renesas and megachips but nice to have nevertheless. Valeo publicly stated that they have orders around the 1 billion €/£/$ mark im not sure which currency. We have not received a penny of preliminary payments so we’re out. We’d at least have seen 10-15% of our share in preliminary payments and they’re surely not around the range of what a kebap store ran by two people make in a month. Let’s hope it’s us that Samsung uses, watch the financials and actually experience a pleasant surprise for once.

DaD

The part on Valeo they have pre orders correct. But in business they likely have not been paid until the unit is shipped or delivered. There could have been a pre payment likely covering production costs and later on when they get paid if we are in Scala 3 we will then get paid does that make sense?

So example at my work I get an order from a customer then I order through my vendor. They deliver to me I pay vendor.

Then I deliver the product to my customer and then invoice them and after that I get paid.

So if I had a small partner they would be paid last as they did not do the work or any risk in buying product building cost and shipping it then hoping they get all the money from the customer.

 

[Tothemoon24]

 

[Makeme 2020]

For god sake grow up.

Sorry Bin Boys DADDY

 

[CrabmansFriend]

View attachment 54610

No real point or obvious connection, I just think InstaDeep is a very interesting company (AI software & solutions, research).
They got aquired by BioNTech at the beginning of 2023. Areas of InstaDeep's focus and solutions are biology (e.g. drug discovery etc.), logistics, electronic design and energy.

They are not working exclusivly for BioNTech but do also work for other customers (e.g. Deutsche Bahn). But given BioNTech's focus on cancer treatments and an anticipated shift to more individualization as well in tech/AI as in medical treatments/devices (remember Nandan's examples about his wife or pacemakers) I expect more and more growth in early diagnosis of illnesses/cancer and more regular health monitoring. This could mean we will see a lot of new devices in the future (or feature enhancements in future iterations of devices we are already using) used for these puposes. What will waerable devices like smartwatches, ear buds, hearing aids, glasses, rings, fitness trackers etc. be able to monitor and recognize in 5 to 10 years from now?

But you're right, I am not joining a dot here. As I wrote before - I just think it's "nice" (it = an interesting company like InstaDeep at the same event as Brainchip).

And just another example for portable, medical diagnosis devices (still in it's early stages, but we are pros in waiting, aren't we? )

FDA approves AI-powered skin cancer-screening device

Testing for this can can involve taking a chunk out of your flesh, making this a case of AI tech being kinder to humans

www.theregister.com

 

[Dr E Brown]

It's just not BRN. Every lithium and graphite stock is being over run with trolls. Look at the SP price not the fundamentals lol

This might end HC. It's going to be left with trolls fighting each other. Investors are leaving in droves. They're sick of wasting time filtering through the crap

And biotechs

 

[Easytiger]

Just a few comments:

1. A second strike, a spill and election of Directors does not touch the position of CEO. His/Her position is unaffected.

2. It has been stated more than once and in some detail by the Chair at the last AGM that as Brainchip cannot afford to match market salaries dollar for dollar to attract the right people they provide shares as part of the salary package.

3. Work is such a ‘sh-t’ you have to pay people to do it.

4. I cannot name anyone from my life experience who refused to take their pay on pay day.

My opinion only DYOR
Fact Finder

The only point I would add to the above is that a second strike gives a clear message to the board that shareholders want change, do you think that the board will hold the line.

 

[Dr E Brown]

I suspect there might be a reluctance to provide an investor update currently, possibly due to a shift in the business model, which could be a sensible decision.

The following is my interpretation based on recent partnership details and CES content. I'm not a tech industry expert, but this is my attempt to read between the lines. If you prefer to avoid speculation, you may want to skip the rest.

Plan A was to sell IP licenses, allowing the license owner to independently develop products. However, this plan did not unfold as expected, not going down that rabbit hole.

Plan B has emerged as a necessary pivot, directing us towards product development and application-specific solutions in collaboration with multiple partners.

This strategy offers a risk-free scenario for the companies we collaborate with as "partners." They gain access to the IP and receive support either at no cost or through hourly consulting fees. This allows them to develop and market a product to an end-user customer or OEM with the potential for success. If the product doesn't sell, they face no financial loss, avoiding the risk of investing millions in a license for IP they might struggle to deploy, while we invest possible years in development without a return. This arrangement is not favorable for us.

However, these corporations, i.e Microchip and Onsemi, boast multi-billion-dollar revenues and extensive resources. The reality is that we have encountered challenges in independently penetrating the market. The most efficient path to bringing our products to market now and mainstreaming SNNs, paving the way to return to plan A, involves riding the coattails of these major players as they compete in the currently hyped edge AI space. The commercial details surrounding the IP in all these partnerships are still unclear.

Even though it represents a tangible step forward with actual signs of progress, this doesn't result in an inspiring update or an outcome from their own efforts that management is likely to emphasise.

Again, this is my current interpretation and could be entirely off the mark.

I don’t think the business model has changed, more the path to the promised land of IP sales has become more apparent.

 

[AI_Inquirer]

Found a Brainchip dot and 1500 reference! There is a lot of leads to delve into with this one.

https://www.linkedin.com/feed/update/urn:li:activity:7153439471911190528/

 

[DerAktienDude]

DaD

The part on Valeo they have pre orders correct. But in business they likely have not been paid until the unit is shipped or delivered. There could have been a pre payment likely covering production costs and later on when they get paid if we are in Scala 3 we will then get paid does that make sense?

So example at my work I get an order from a customer then I order through my vendor. They deliver to me I pay vendor.

Then I deliver the product to my customer and then invoice them and after that I get paid.

So if I had a small partner they would be paid last as they did not do the work or any risk in buying product building cost and shipping it then hoping they get all the money from the customer.

When we’re talking about sums that could be in the range of 100 million USD per customer you want some kind of security/ bond in form of preliminary payments in case the customer falls out. (Insolvency, other product whatever might be the case)

Correct these payments are mostly made to at least cover production costs, costs that went into the development of the product or even for wages if you’re that risky. Highly unlikely though.

And you’re most likely correct that we’d receive a fraction of these payments because we’re at the end of the food chain at this point. We’re even receiving them fairly late I guess.

So if we’re calculating the lowest amount, of 10% preliminary payments of the 1bn USD there will be 100 million USD left. Let’s say Valeo keeps half of it because other vendors who participated want a security as well as and calculate that we’d receive 1% of the rest. 50 million USD would be left so 500k for us.
That’d be the absolute bottom of the barrel that I’m able to come up with. It’s most likely much higher than that though.

 

[DingoBorat]

When we’re talking about sums that could be in the range of 100 million USD per customer you want some kind of security/ bond in form of preliminary payments in case the customer falls out. (Insolvency, other product whatever might be the case)

Correct these payments are mostly made to at least cover production costs, costs that went into the development of the product or even for wages if you’re that risky. Highly unlikely though.

And you’re most likely correct that we’d receive a fraction of these payments because we’re at the end of the food chain at this point. We’re even receiving them fairly late I guess.

So if we’re calculating the lowest amount, of 10% preliminary payments of the 1bn USD there will be 100 million USD left. Let’s say Valeo keeps half of it because other vendors who participated want a security as well as and calculate that we’d receive 1% of the rest. 50 million USD would be left so 500k for us.
That’d be the absolute bottom of the barrel that I’m able to come up with. It’s most likely much higher than that though.

"so 500k for us"

Now we're talking more "Barnacle Bill's" if that comes through