BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
7für7
Top 20
Interesting what happens in Greece recently
www.ekathimerini.com
Greece to host EU-backed AI factory, PM announces | eKathimerini.com
Greece is among seven European Union nations selected to establish one of Europe’s first artificial intelligence (AI) factories with funding from the European Commission and national sources, Prime Minister Kyriakos Mitsotakis announced Tuesday.
Just out of interest Income and State taxes in California outlined below.
Mentioning because of all the crap about Sean and the BOD selling converted vested RSU shares to pay tax- mainly on the crapper.
Aussies have this 'idea' that taxes in the USA are generally low because the Fed rate is lower but its not widely understood that other taxes apply. They get really peed off on the crapper when i post about Sean and Tony V holdings and the whopping size of inside holdings.
Mentioning because of all the crap about Sean and the BOD selling converted vested RSU shares to pay tax- mainly on the crapper.
Aussies have this 'idea' that taxes in the USA are generally low because the Fed rate is lower but its not widely understood that other taxes apply. They get really peed off on the crapper when i post about Sean and Tony V holdings and the whopping size of inside holdings.
- Federal income tax: 22% (37% once your supplemental wages exceed $1,000,000)
- California income tax: 10.23%
- Medicare tax: 1.45%
- Additional Medicare tax: 0.9% (once your total wages exceed $200,000)
- Social Security tax: 6.2% (up to the wage limit)
- California State Disability Insurance: 1.2% (up to the wage limit). If you don’t live in California, your locality may have its own version of state disability or other payroll taxes.
Boab
I wish I could paint like Vincent
California is No 4 in the list of highest overall taxes and on income tax they are number 1.
Smoothsailing
Regular
Asking ChatGpt
Emerging Developments in Neuromorphic Chips
Recent advancements in neuromorphic chip designs indicate a strong push towards their adoption in mobile technology:
Emerging Developments in Neuromorphic Chips
Recent advancements in neuromorphic chip designs indicate a strong push towards their adoption in mobile technology:
- BrainChip’s Akida Pico: This chip has been specifically designed for use in power-constrained environments such as mobile devices. Its architecture supports efficient neural processing while maintaining low energy consumption.
BrainShit
Regular
Use Perplexity instead... because, it scores with high factual accuracy, the ability to search the Internet in real time and provides citations for its answers. It is particularly useful for precise research and up-to-date information queries.
ChatGPT is ... lets say more versatile in content creation, including creative texts such as poems and scripts, and offers a fluid conversational experience. However, it is not always up to date with the latest information. So kind of more creative than fact based....
But .... BrainChip envisions that key applications for the Akida Pico will include voice wake detection, keyword recognition, noise suppression, presence detection, wearables, and personal voice assistants. It can also be used in appliances for voice control and is particularly effective for wearable AI and smart home devices. As BrainChip puts it, ultra-low-power consumption of the Akido Pico is optimized for 'extreme edge AI,' possibly implying on emerging devices that do not exist yet.
"Like all of our Edge AI enablement platforms, Akida Pico was developed to further push the limits of AI on-chip compute with low latency and low power required of neural applications," said Sean Hehir, CEO at BrainChip. "Whether you have limited AI expertise or are an expert at developing AI models and applications, Akida Pico and the Akida Development Platform provides users with the ability to create, train and test the most power and memory efficient temporal-event based neural networks quicker and more reliably."
Source: https://www.tomshardware.com/tech-i...an-ai-npu-that-consumes-less-than-a-milliwatt
Source: https://www.embedded.com/brainchips...-ai-processing-with-event-based-architecture/
Smoothsailing
Regular
I can paint your house champ however Iam not very technically. Well dun on this up date tho I appreciate it…cheers
Last edited:
A little off topic but there is not a means test for Social Security in the US.
The prevailing view in the US is that everyone earns it and the costs of maintaining a system fairly would cancel out and maybe cost more than just a simple everyone qualifies.
Also the value of concessions here is not added to individual pension incomes so they are effectively 'tax free perks'.
Exactly Dio.........Intel are hanging onto our coat tails, we are actually commercial, they are from my vantage point, still wearing their
lab coats, Intel as a company have so many issues at play, even with the new CEO (did Pat finally fall on his sword ?) she has a mammoth
task ahead of her, the old Intel carried a lot of weight years ago, but I personally don't view Mike and his team a threat to us at all.
I do know that our key leaders within Brainchip have a lot of respect for Mike (Davies) but we are ahead of them, that I feel comfortable
saying, you can only ride the wave of yesteryear for so long....the power of the name Intel, is a fading star, purely my view, many may
totally disagree with me, and that's all good.
Cheers....Tech
Last edited:
Yes. I don't bear any ill will towards Intel. They mead great computers last millenium.
I think their futurists got the AI revolution date wrong. Is seems like Mike's group were given the wrong riding instructions - solid research and development, but misreading the pace of AI change and perhaps perhaps a touch of incumbency huberis.
On the other hand, they have been working on SLMs for a few years ...
US2022036123A1 MACHINE LEARNING MODEL SCALING SYSTEM WITH ENERGY EFFICIENT NETWORK DATA TRANSFER FOR POWER AWARE HARDWARE 20211020
Machine learning model swap (MLMS) framework for that selects and interchanges machine learning (ML) models in an energy and communication efficient way while adapting the ML models to real time changes in system constraints. The MLMS framework includes an ML model search strategy that can flexibly adapt ML models for a wide variety of compute system and/or environmental changes. Energy and communication efficiency is achieved by using a similarity-based ML model selection process, which selects a replacement ML model that has the most overlap in pre-trained parameters from a currently deployed ML model to minimize memory write operation overhead.
... so they aren't completely behind the 8-ball.
Sean as the BOD 'mouthpiece' has stated that the aim is in a decade, is that BRN will be one of the top 3 leaders of 'Edge' providers.
That is why we need to move 'big' in 2025. One decent deal and we are off and running on the way to build something really special.
No real fear of competition right now as that will come later when the market is huge and we are well on the way to being a large business with established clients.
Large markets attract new entrants. The market is not large ATM.
Fortunately our whopping level of inside ownership will either block any takeover attempts or force a huge to good to refuse premium.
That is why we need to move 'big' in 2025. One decent deal and we are off and running on the way to build something really special.
No real fear of competition right now as that will come later when the market is huge and we are well on the way to being a large business with established clients.
Large markets attract new entrants. The market is not large ATM.
Fortunately our whopping level of inside ownership will either block any takeover attempts or force a huge to good to refuse premium.
rgupta
Regular
You have a point Manny that market is not big enough and that is why not enough competition.
But another matter of fact is that they want to come as close as possible to a human brain.
Regarding getting into top 3 players at edge AI means there are and will be similarly technologies like our, who are or will be working at the edge. Right now Qualcomm and apple are one of the biggest company working towards the edge. Let us see what the future holds.
Dyor
Pom down under
Top 20
I’ll be just happy with any deal that propels us past the $1 mark atm.
Pom down under
Top 20
Hey @TECH have you extended your Jan 2025 timeline by anychance?
FJ-215
Regular
manny has pushed his out a decade so the dollar party may be a way off.
BrainShit
Regular
Custom hardware tailored to specific models can unlock performance gains and energy savings that generic hardware cannot achieve, but there are tradeoffs.
HW and SW Architecture Approaches For Running AI Models
State space models
“SSMs essentially take control theory and apply deep learning to it so that you hold state like in an RNN, that feeds back results,” said Steve Brightfield, chief marketing officer of BrainChip. “Power has always been an issue with AI models, but it really accelerated with LLMs and transformers. Transformers have a squared computational complexity. When you have a longer contact window, the computations square. It’s a logarithmic power increase, and that’s what everybody’s been challenged by. It’s a scaling problem. What now you’re seeing is, even at the data center level, they’re looking at modifying a transformer to be a hybrid model. It will use a transformer for some pieces of it, but it will replace other portions of the transformer with either a state space model or some other type of model approach.”
semiengineering.com
HW and SW Architecture Approaches For Running AI Models
State space models
“SSMs essentially take control theory and apply deep learning to it so that you hold state like in an RNN, that feeds back results,” said Steve Brightfield, chief marketing officer of BrainChip. “Power has always been an issue with AI models, but it really accelerated with LLMs and transformers. Transformers have a squared computational complexity. When you have a longer contact window, the computations square. It’s a logarithmic power increase, and that’s what everybody’s been challenged by. It’s a scaling problem. What now you’re seeing is, even at the data center level, they’re looking at modifying a transformer to be a hybrid model. It will use a transformer for some pieces of it, but it will replace other portions of the transformer with either a state space model or some other type of model approach.”
HW and SW Architecture Approaches For Running AI Models
Custom hardware tailored to specific models can unlock performance gains and energy savings that generic hardware cannot achieve, but there are tradeoffs.
semiengineering.com
I never said $1 for a decade target or made any $ prediction al all in my post.
I do not know why you would say that?
I am not into SP predictions except I do believe that short term we will close chart gaps above if we get a decent deal.
From memory the highest gap above is around $1.90.
From there the next target will be the last highest high circa $2.34 and blue sky from there.
The above are 2025 aspirations on a decent deal.
Tothemoon24
Top 20
Hopefully when the senior Scientist from Tata says “ our Spiking Neural Network “ he’s meaning our partners Brainchip
Where it can be used? -- In stabilizing a drone in high wind, docking a space ship perfectly on a space station, controlling and balancing an automated car etc. Stay tuned as we delve into exploring potential power efficiencies on neuromorphic boards and market opportunities. Exciting times ahead!
Where it can be used? -- In stabilizing a drone in high wind, docking a space ship perfectly on a space station, controlling and balancing an automated car etc. Stay tuned as we delve into exploring potential power efficiencies on neuromorphic boards and market opportunities. Exciting times ahead!
Last edited:
Similar threads
