Tothemoon24
Top 20
Can Stable Diffusion generative AI run on a smartphone? Qualcomm says — Yes it can — VentureBeat
The most advanced generative AI models in the world, like Stable Diffusion, generally run only in the cloud.
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Learning
Learning to the Top 🕵♂️
Hi MadMayHam,
Good Morning to you.
Fantastic find, thanks for sharing
It's great to be a shareholder 🏖
Fact Finder
Top 20
Great research and would explain how the head of Accenture would be in a position to say in a podcast that Brainchip was already successful.
And here I was thinking that he came on the podcast to help out an old mate and was reading from a script prepared beforehand.
Clearly Accenture as an organisation has in-depth knowledge of Brainchip and its AKIDA technology as they even correctly identify in the paper that AKIDA and Loihi are on different technology pathways SOMETHING WHICH even Quantum Ventura Inc did not fully understand in there paper. Hopefully their intern will help to get them on track.
My opinion only DYOR
FF
AKIDA BALLISTA
jtardif999
Regular
I would suggest based on the @Diogenese find that any tech using foveated rendering (the spell checker doesn’t even like the word) would be a good chance to also involve Neuromorphic computing.I had never heard of the word foveated before in my life and within an hour of reading your post, this came up in my feed. Cookies and targeted marketing at its best/worst.
Likely nothing to do with Akida but PSVR2 tracks your eyes and only renders the area in the scene you are looking at in high def. The areas you are not looking at are rendered at a lower resolution. Pretty incredible tech.
What is foveated rendering, and why can it make PSVR 2 streams blurry? | TheSixthAxis
Foveated Rendering is making PSVR 2 streams and videos look a bit blurrier and lower resolution than they feel to players. We dive in to see why.www.thesixthaxis.com
What is foveated rendering, and why can it make PSVR 2 streams blurry?
Stefan L 22/02/23
Pushing high-end graphics in virtual reality is a demanding endeavour for any VR headset processor, whether it’s a standalone headset, a high-end PC or the PS5 and PSVR 2. It’s made harder still for PSVR 2 by the high-end specs of the headset and the new expectation for high resolution gaming. So, how does the PS5 pull it off, and why doesn’t it always translate to what you can see in YouTube and Twitch streams?
One of the tricks that PSVR 2 leans upon when render virtual reality is called foveated rendering. You’ll see this term used in a lot of bullet point lists when games are being upgraded or ported to PSVR 2, alongside noting support for 4K and HDR. It’s a lot of attention for a rather technical feature to get, but what is it?
To put it simply, foveated rendering is a way to reduce image quality for parts of the screen that are in your peripheral vision. The game can prioritise where it’s spending the most hardware resources to get the best performance and visual effect for the player.
Focusing on the target with my bow, Horizon Call of the Mountain uses foveated rendering to lower the resolution elsewhere, most noticeable on the blurrier barrels and settlement in the top left.
It’s all based on real life. When you focus your vision on a cup on a coffee table, for example, the rest of the room that is outside that central focal point is less clear to your eyes and brain. Foveated rendering simulates that effect in VR, allowing the game engine to reducing detail and resolution in places that you aren’t looking.
This has been used since the very start of the current era of virtual reality gaming in 2016, but was initially done in a more rudimentary fashion by locking the focus point to the direction of the headset. Some headsets since the Vive Pro Eye in 2019 have included an internal camera to track eye movements – a key feature for PSVR 2 – allowing for games to more dynamically shift the rendering focal point to better match what you’re actually looking at.
This effect is practically unnoticeable within PSVR 2 itself. The very edges of your field of view are always going going to be a bit blurred and distorted because of the image passing through the Fresnel lenses in the headset, but within that central area there’s still plenty of space to optimise what is being given the most attention – wherever you move your eyes to will then have the game’s highest resolution and level of detail.
But foveated rendering is laid bare to people watching PSVR 2 streams and video capture, and adds to the challenge of making a pleasant viewing experience – the other being that the view can move, shake and judder even if you think you’re keeping your head still. Different parts of the screen are going in and out of higher and lower resolution, matching the player’s eye movements, and that can naturally lead to the snap judgement that the game resolution is low and the graphics fuzzy. Now, there is a natural softness to the player as the image is passing through a lens – even when in the sweet spot, the headset won’t be quite as sharp as getting up close to a 4K TV – but VR games on YouTube will always have parts of the screen look blurrier and lower resolution than they do to the person playing.
In some ways this makes the 4K claim that games make a bit of a lie – similar to the dynamic resolutions that are so often used for flat screen games. Claiming 4K is really just using an easily understood term to convey something much more complex, the notion that you’re getting 4K-like rendering at the things you are specifically looking at.
It’s all part of the difficult challenge that developers and users face in trying to show and explain what VR gamers are like, without putting a headset on someone and sending them on their way. Just rest assured that foveated rendering might look a bit rough in a stream, but it’s all in aid of making the actual games look and feel as good as possible in VR.
toasty
Regular
Ive been thinking about yesterday's Mercedes reveal. I'm wondering if AKD1500 might be for Merc? My reasoning is this. Merc have obviously determined that centralised processing is appropriate for their deployment, evidenced by their use of NVidia in particular. However, we know that they continue to work with BRN on deployment of AKIDA. My guess is that AKIDA 1500 will provide the "smarts" at the sensor level whilst MB.OS/NVIDIA will be used as the processing/decision making engine. Just my musings so make sure you DYOR.
Dhm
Regular
Hasn't the Akida 1500 been (more quickly) developed at the insistance of a client? Who knows which one.
Funny you post this cause I was thinking on Wednesday what ever eventuated with Rob from Dell and his chat on podcast 6 from 3rd of May 21.
Learning
Learning to the Top 🕵♂️
Hi toasty,
I too have a similar thought, since the news released of AKD1500.
Mercedes are working with NVIDIA, so it would be more cost effective for Mercedes to eliminate the Arm Cortex processor and just have Akida working with NVIDIA as the main processor. (Just my un technically thought)
Learning 🏖
jtardif999
Regular
Interesting development in the parking area. Apologies if it has already been posted:
NTT DATA, Valeo and Embotech Form Consortium to Provide Automated Parking Solutions
VEN.AI aims to be the go-to solution provider of automated parking solutions for car manufacturers
January 4, 2023, 12:05 PM Eastern Standard Time
NTT DATA, a global digital business and IT services leader, Valeo, and Embotech, a software scale-up for autonomous driving systems, announce the next big step to provide automated parking solutions. Together, the three companies have built a consortium, VEN.AI, that aims to be the go-to solution provider for production ready parking automation with global roll out capabilities. The consortium combines each company’s core competencies including owned IP, the latest technology, sales and support structure and strong operations offerings.
LAS VEGAS--(BUSINESS WIRE)--VEN.AI offers an infrastructure-based solution that has very few requirements from the vehicle side as it guides vehicles via the use of sensors, connectivity (e.g., 5G) and offboard computing to a dedicated parking spot. Automated parking solutions can be implemented in a variety of use cases including the assembly plants where vehicles are produced, outbound-logistics distribution parks, depots for vehicle fleets operators, retail outlets as well as in parking garages as an automated valet service.
In the initial phase, VEN.AI is focusing on manufacturing-specific use cases to help car manufacturers increase the efficiency of their assembly lines by automatically guiding vehicles from one production station to another. From here, it will then drive the vehicle from the end of the assembly line to its dedicated spot on the large launch areas. The outcome is a more efficient assembly line, saving on time and costs.
“VEN.AI is an example of the investments NTT DATA is making to drive efficiency and productivity toward Smart Manufacturing for our clients in the automotive industry,” said Prasoon Saxena, President of Manufacturing, NTT DATA Services. “With innovations from NTT DATA including advances in Edge computing, network monitoring and advanced connectivity, VEN.AI will make mobility much easier and energy efficient for consumers, car manufacturers and more.”
“This global partnership has brought about a cutting-edge innovation that swiftly addresses the digital transformation our global automotive clients urgently need in streamlining their production process,” said Robb Rasmussen Global CMO, NTT DATA.
As the world leader in ADAS sensors and related detection algorithms, Valeo is responsible for supplying the necessary technology to make the parking system function properly. The sensors supplied are used to detect and locate cars in the parking area, as well as to understand the surrounding environment. This information is then used by the algorithms to accurately guide the car to an available parking space.
“We are excited to participate in this automated valet parking project alongside NTT DATA and Embotech,” said Marc Vrecko, President, Comfort and Driving Assistance Systems Business Group, Valeo. “Fully automated parking is growing, due to the increasing demand of our customers, and Valeo has all the assets to respond to this market: we have the most comprehensive sensor portfolio on the market, offering the vehicle a complete vision. And having precise perception is key to ensuring vehicles are safely and efficiently parked.”
Safety is a key component of VEN.AI by integrating the latest technology innovations including the use of cameras, light detection, low-latency connectivity and sensors. The system has been designed to fulfill the highest requirements and standards on safety and availability and has the best-in-class building components including Embotech PRODRIVER®. This SAE L4 virtual driver is designed to provide fully flexible, human-like driving, which is a key enabler of automated driving in busy logistics areas with mixed traffic as well as parking into very tight spaces, increasing the space efficiency.
“VEN.AI proves that autonomous driving, when employed in the proper context, has an immediate business case with the potential to disrupt the future of automotive manufacturing and distribution,” said Andreas Kyrtatos, CEO, Embotech. “This venture has innovation at its core and is dedicated to bringing sustainable autonomy to automotive manufacturing and distribution.”
Additionally, VEN.AI is integrating the parking automation solution within global production systems, adding automated charging for electric vehicles, and using internal onboard sensors to improve efficiency and flexibility in large parking areas. The global roll out also includes 24/7 operations support.
NTT DATA, Valeo and Embotech Form Consortium to Provide Automated Parking Solutions
VEN.AI aims to be the go-to solution provider of automated parking solutions for car manufacturers
January 4, 2023, 12:05 PM Eastern Standard Time
NTT DATA, a global digital business and IT services leader, Valeo, and Embotech, a software scale-up for autonomous driving systems, announce the next big step to provide automated parking solutions. Together, the three companies have built a consortium, VEN.AI, that aims to be the go-to solution provider for production ready parking automation with global roll out capabilities. The consortium combines each company’s core competencies including owned IP, the latest technology, sales and support structure and strong operations offerings.
LAS VEGAS--(BUSINESS WIRE)--VEN.AI offers an infrastructure-based solution that has very few requirements from the vehicle side as it guides vehicles via the use of sensors, connectivity (e.g., 5G) and offboard computing to a dedicated parking spot. Automated parking solutions can be implemented in a variety of use cases including the assembly plants where vehicles are produced, outbound-logistics distribution parks, depots for vehicle fleets operators, retail outlets as well as in parking garages as an automated valet service.
In the initial phase, VEN.AI is focusing on manufacturing-specific use cases to help car manufacturers increase the efficiency of their assembly lines by automatically guiding vehicles from one production station to another. From here, it will then drive the vehicle from the end of the assembly line to its dedicated spot on the large launch areas. The outcome is a more efficient assembly line, saving on time and costs.
“VEN.AI is an example of the investments NTT DATA is making to drive efficiency and productivity toward Smart Manufacturing for our clients in the automotive industry,” said Prasoon Saxena, President of Manufacturing, NTT DATA Services. “With innovations from NTT DATA including advances in Edge computing, network monitoring and advanced connectivity, VEN.AI will make mobility much easier and energy efficient for consumers, car manufacturers and more.”
“This global partnership has brought about a cutting-edge innovation that swiftly addresses the digital transformation our global automotive clients urgently need in streamlining their production process,” said Robb Rasmussen Global CMO, NTT DATA.
As the world leader in ADAS sensors and related detection algorithms, Valeo is responsible for supplying the necessary technology to make the parking system function properly. The sensors supplied are used to detect and locate cars in the parking area, as well as to understand the surrounding environment. This information is then used by the algorithms to accurately guide the car to an available parking space.
“We are excited to participate in this automated valet parking project alongside NTT DATA and Embotech,” said Marc Vrecko, President, Comfort and Driving Assistance Systems Business Group, Valeo. “Fully automated parking is growing, due to the increasing demand of our customers, and Valeo has all the assets to respond to this market: we have the most comprehensive sensor portfolio on the market, offering the vehicle a complete vision. And having precise perception is key to ensuring vehicles are safely and efficiently parked.”
Safety is a key component of VEN.AI by integrating the latest technology innovations including the use of cameras, light detection, low-latency connectivity and sensors. The system has been designed to fulfill the highest requirements and standards on safety and availability and has the best-in-class building components including Embotech PRODRIVER®. This SAE L4 virtual driver is designed to provide fully flexible, human-like driving, which is a key enabler of automated driving in busy logistics areas with mixed traffic as well as parking into very tight spaces, increasing the space efficiency.
“VEN.AI proves that autonomous driving, when employed in the proper context, has an immediate business case with the potential to disrupt the future of automotive manufacturing and distribution,” said Andreas Kyrtatos, CEO, Embotech. “This venture has innovation at its core and is dedicated to bringing sustainable autonomy to automotive manufacturing and distribution.”
Additionally, VEN.AI is integrating the parking automation solution within global production systems, adding automated charging for electric vehicles, and using internal onboard sensors to improve efficiency and flexibility in large parking areas. The global roll out also includes 24/7 operations support.
Dhm
Regular
Most probably not us, but....
https://www.smh.com.au/national/nsw...be-tracked-at-mardi-gras-20230223-p5cn1a.html
Don't you think the colouring in the last graphic is suspiciously similar to how Brainchip shows things on their website?
https://www.smh.com.au/national/nsw...be-tracked-at-mardi-gras-20230223-p5cn1a.html
Don't you think the colouring in the last graphic is suspiciously similar to how Brainchip shows things on their website?
VictorG
Member
alwaysgreen
Top 20
In all likelihood, we will be out of the 200 in the next rebalance. Currently sitting out of the 300.
Looking at this positively, hopefully there's less instos and therefore, less shorters.
Yes agreed but the psvr2 has just been released to market. I think it's too early to include Akida but again, would love to be proven wrong.
alwaysgreen
Top 20
D
Deleted member 118
Guest
alwaysgreen
Top 20
BaconLover
Founding Member
Unfortunately this also means some of those funds will off load us from their portfolio so a bit selling pressure can be expected.View attachment 30445
In all likelihood, we will be out of the 200 in the next rebalance. Currently sitting out of the 300.
Looking at this positively, hopefully there's less instos and therefore, less shorters.
Yes agreed but the psvr2 has just been released to market. I think it's too early to include Akida but again, would love to be proven wrong.
Unless ofcourse some magic happens around the AGM time, which I'm still hopeful for.
jtardif999
Regular
MB had a good 2022. They look after their shareholders:
DOW JONES NEWSWIRES
Mercedes-Benz to return nearly $6 billion to shareholders a strong fourth quarter, but warns of flat sales in 2023
Mercedes-Benz Group AG is planning to distribute 5.6 billion euros ($5.98 billion) to shareholders after revenue and net profit surged in the fourth quarter and the whole of 2022, though the company warned that sales should be flat this year as supply snags remain.
The German luxury-car maker XE:MBG on Friday posted quarterly revenue of EUR41.00 billion, up from EUR35.24 billion the previous year.
Quarterly net profit jumped 63% to EUR4.03 billion, bringing profit for the year to EUR14.81 billion, up 34%.
Earnings before interest and taxes–a key profitability metric–rose to EUR5.41 billion in the quarter from EUR4.12 billion.
Analysts polled by FactSet had forecast quarterly revenue of EUR38.44 billion on profit of EUR3.73 billion and EBIT of EUR5.14 billion.
The company said it would propose a dividend of EUR5.20 a share for 2022, or EUR5.6 billion in total, topping the EUR5.00 a share it paid for the previous year.
Mercedes-Benz said the global shortage of semiconductors was easing, though supply constraints remained.
For 2023, Mercedes-Benz expects revenue at the 2022 level, and EBIT slightly below it.
Write to Mauro Orru at mauro.orru@wsj.com; @MauroOrru94
Tracking the pulse of the markets
Copyright © 2023 MarketWatch, Inc. All rights reserved.
DOW JONES NEWSWIRES
Mercedes-Benz to return nearly $6 billion to shareholders a strong fourth quarter, but warns of flat sales in 2023
Mercedes-Benz Group AG is planning to distribute 5.6 billion euros ($5.98 billion) to shareholders after revenue and net profit surged in the fourth quarter and the whole of 2022, though the company warned that sales should be flat this year as supply snags remain.
The German luxury-car maker XE:MBG on Friday posted quarterly revenue of EUR41.00 billion, up from EUR35.24 billion the previous year.
Quarterly net profit jumped 63% to EUR4.03 billion, bringing profit for the year to EUR14.81 billion, up 34%.
Earnings before interest and taxes–a key profitability metric–rose to EUR5.41 billion in the quarter from EUR4.12 billion.
Analysts polled by FactSet had forecast quarterly revenue of EUR38.44 billion on profit of EUR3.73 billion and EBIT of EUR5.14 billion.
The company said it would propose a dividend of EUR5.20 a share for 2022, or EUR5.6 billion in total, topping the EUR5.00 a share it paid for the previous year.
Mercedes-Benz said the global shortage of semiconductors was easing, though supply constraints remained.
For 2023, Mercedes-Benz expects revenue at the 2022 level, and EBIT slightly below it.
Write to Mauro Orru at mauro.orru@wsj.com; @MauroOrru94
Tracking the pulse of the markets
Copyright © 2023 MarketWatch, Inc. All rights reserved.
buena suerte :-)
BOB Bank of Brainchip
I agree .... and with all the leg work/trials/testing done for the production and successful launch of our flagship chip AKD1000 has helped I think in being able to adjust for customers requirements to produce AKD1500 at a more rapid pace!!?
Have a positive Friday and weekend Chippers
Last edited:
Similar threads
