BRN Discussion Ongoing

Diogenese

Top 20
Korea
Sapeon Korea Inc (SK Telecom)
CEO talking about "SNN" "Neuromorphic"

CEO Ryu suggested SNN and convolutional neural network (CNN) technologies as neuromorphic technologies that will significantly improve memory operation processing performance. SNN is a technology that creates spikes, which resembles electrical signals, and transmits the signals to the opposite side. This reduces power consumption compared to existing neuromorphic technologies. SNN performs information processing when a signal is expressed.
Hi mia,

"SK Group artificial intelligence (AI) semiconductor fabless Sapeon Korea introduced the next-generation memory semiconductor Compute-in-Memory (CIM) technology trend with enhanced computing functions. CIM is a technology that implements logic and memory semiconductors on a single chip."

Compute-in-memory (CIM) usually refers to analog ReRAM/MemRistor, although it may sometimes be loosely used in reference to Akida. But Akida uses compute-near-memory.

In CIM, the memory does the computing by adding currents from the several inputs to change the value of the MemRistor impedance.
 
  • Like
  • Love
  • Fire
Reactions: 25 users
“When I joined, we were only focussed in the US, we now have sales footprint in Japan, Korea, Israel, Europe…” (from 37:43 min) “

Hi Fran , maybe Sean is informing us that we now have Brainchip employees in these locations to help drive adoption .


All in my opinion .
That's exactly what I thought he meant Doz, or at least along those lines..

I didn't think he meant we had only just begun dealing with customers (potential or not) in those countries.

But that we had a more substantial "presence" in them now.
 
  • Like
  • Love
  • Fire
Reactions: 22 users

IloveLamp

Top 20
Screenshot_20230913_054532_LinkedIn.jpg
 
  • Like
  • Love
  • Thinking
Reactions: 23 users

cosors

👀
Hi mia,

"SK Group artificial intelligence (AI) semiconductor fabless Sapeon Korea introduced the next-generation memory semiconductor Compute-in-Memory (CIM) technology trend with enhanced computing functions. CIM is a technology that implements logic and memory semiconductors on a single chip."

Compute-in-memory (CIM) usually refers to analog ReRAM/MemRistor, although it may sometimes be loosely used in reference to Akida. But Akida uses compute-near-memory.

In CIM, the memory does the computing by adding currents from the several inputs to change the value of the MemRistor impedance.
Thank you! Short and to the point. You help me as a layman a lot to understand this better.
Now I wish you that you will be rewarded for your tireless work for us with a good heavy SP at BRN and TLG together - quite unselfishly.)
 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 14 users

cosors

👀
View attachment 44476
🤘😅 Have you seen this yet?!!

The Ultimate LiDAR SLAM Robot Lawn Mower
https://www.kickstarter.com/projects/hookii/neomow-x-the-ultimate-lidar-slam-robot-lawn-mower

___
Wasn't there once propagate that LiDAR is far too expensive 🤔
 
Last edited:
  • Like
  • Fire
Reactions: 7 users

Tothemoon24

Top 20
According to a tweet from a trustable leaker and X user, @Tech_Reve, the Samsung XR headset will display an ultra-high 3,000 ppi pixel density. Like Apple"s Vision Pro headset, it is said to support hand tracking. The leaker claims the XR headset will incorporate smell in some capacity.


 
  • Like
  • Thinking
  • Fire
Reactions: 20 users

chapman89

Founding Member
That's exactly what I thought he meant Doz, or at least along those lines..

I didn't think he meant we had only just begun dealing with customers (potential or not) in those countries.

But that we had a more substantial "presence" in them now.
It’s exactly what Sean said and he is right, under his watch we now have head of sales in other countries, and the head of sales in Japan for Brainchip just happens to be the former Head of Sales guy for ARM in Japan!
 
  • Like
  • Fire
  • Love
Reactions: 65 users

GStocks123

Regular
Worth having a squiz through here. Definitely smells like Akida! 👃

 
  • Like
  • Fire
  • Love
Reactions: 27 users
Has anyone watched the new Apple Watch presentation? They talk about processing at the edge, reduced power... all the words that excite us! I'm not getting my hopes up too much yet but I am very curious and feel like some people on this forum would be able to delve deeper into this than I can.

Quotes from the presentation:
From the 12 minute mark in the presentation

"S9 [the apple silicon chip] also has a new 4-core Neural Engine that can process machine learning tasks up to twice as fast.
This powerful custom silicon is what allows us to maintain all-day 18-hour battery life while adding new features and system-wide improvements to Series 9."

"Thanks to the powerful new neural engine, Siri requests are now processed on-device, making them faster and more secure. This means that the most common requests like... no longer have to go to the cloud, so they can't be slowed down by a poor Wi-Fi or cellular connection. "

"And because the Neural Engine can run a more advanced transformer model, dictation on Series 9 is up to 25% more accurate than on Series 8"

From the 16 minute mark in the presentation
"Double Tap controls the primary button in an app..." [you touch your index and thumb together on your watch hand twice to perform a double tap]
"The new gesture is enable by the powerful neural engine in Series 9, which processes data from the accelerometer, gyroscope, and optical heart sensor in a completely new way. It uses a machine learning algorithm to detect the unique signature of tiny movements and changes in blood flow when you hand and fingers perform a double tab"

Go to the 12 minute mark.
 
  • Like
  • Thinking
  • Fire
Reactions: 46 users

DK6161

Regular
Has anyone watched the new Apple Watch presentation? They talk about processing at the edge, reduced power... all the words that excite us! I'm not getting my hopes up too much yet but I am very curious and feel like some people on this forum would be able to delve deeper into this than I can.

Quotes from the presentation:
From the 12 minute mark in the presentation

"S9 [the apple silicon chip] also has a new 4-core Neural Engine that can process machine learning tasks up to twice as fast.
This powerful custom silicon is what allows us to maintain all-day 18-hour battery life while adding new features and system-wide improvements to Series 9."

"Thanks to the powerful new neural engine, Siri requests are now processed on-device, making them faster and more secure. This means that the most common requests like... no longer have to go to the cloud, so they can't be slowed down by a poor Wi-Fi or cellular connection. "

"And because the Neural Engine can run a more advanced transformer model, dictation on Series 9 is up to 25% more accurate than on Series 8"

From the 16 minute mark in the presentation
"Double Tap controls the primary button in an app..." [you touch your index and thumb together on your watch hand twice to perform a double tap]
"The new gesture is enable by the powerful neural engine in Series 9, which processes data from the accelerometer, gyroscope, and optical heart sensor in a completely new way. It uses a machine learning algorithm to detect the unique signature of tiny movements and changes in blood flow when you hand and fingers perform a double tab"

Go to the 12 minute mark.

At the edge, reduce power.. Yep definitely us!
 
  • Haha
Reactions: 1 users

DK6161

Regular
Has anyone watched the new Apple Watch presentation? They talk about processing at the edge, reduced power... all the words that excite us! I'm not getting my hopes up too much yet but I am very curious and feel like some people on this forum would be able to delve deeper into this than I can.

Quotes from the presentation:
From the 12 minute mark in the presentation

"S9 [the apple silicon chip] also has a new 4-core Neural Engine that can process machine learning tasks up to twice as fast.
This powerful custom silicon is what allows us to maintain all-day 18-hour battery life while adding new features and system-wide improvements to Series 9."

"Thanks to the powerful new neural engine, Siri requests are now processed on-device, making them faster and more secure. This means that the most common requests like... no longer have to go to the cloud, so they can't be slowed down by a poor Wi-Fi or cellular connection. "

"And because the Neural Engine can run a more advanced transformer model, dictation on Series 9 is up to 25% more accurate than on Series 8"

From the 16 minute mark in the presentation
"Double Tap controls the primary button in an app..." [you touch your index and thumb together on your watch hand twice to perform a double tap]
"The new gesture is enable by the powerful neural engine in Series 9, which processes data from the accelerometer, gyroscope, and optical heart sensor in a completely new way. It uses a machine learning algorithm to detect the unique signature of tiny movements and changes in blood flow when you hand and fingers perform a double tab"

Go to the 12 minute mark.

Actually, I can't see a like from Rob Telson. So definitely not us.
 
  • Haha
  • Like
  • Thinking
Reactions: 25 users

rgupta

Regular
At the edge, reduce power.. Yep definitely us!
That was exactly what Antonio was saying. These things are possible without akida1000 as well and that is why he said it was not robust enough.
 

Luppo71

Founding Member
Hi FMF,

I was in the auto repair industry for a while. If you look at new car prices vs what it would cost to build a car from spare parts you will see the monster mark up that car companies make compared to initial sales. The vehicle you buy just gets you on the hook for credit, maintenance & parts.

We are moving to a world of EV cars (less moving parts, less maintenance), that just leaves credit. Add to that, less accidents from AI driving solutions and car companies have a big problem.

Mercedes seem to be ahead of the curve here in offering the vehicles as software as a service. Gives me the screaming you know whats but that is where we seem to be heading.
And the biggest cost for insurance companies is nose to tail accidents which should drop 90% with all new cars
with collision assist but will they drop premiums, i doubt it very much just huge profits incoming.
 
  • Like
Reactions: 10 users

Balliwood

Member
That's exactly what I thought he meant Doz, or at least along those lines..

I didn't think he meant we had only just begun dealing with customers (potential or not) in those countries.

But that we had a more substantial "presence" in them now.



Why was India not mentioned in the International reach?

The Tata/Brainchip link could be very important :-

Data Centres use prodigious amounts of water and energy

India is short of water

Rapid electrification of transport will test India’s power grid capacity. {One factory is making a electric motorcycle every 90 seconds.}

Tata is a very large Indian company interested in AI development

Tata knows about Brainchip

Brainchip’s neuromorphic system-on-a-chip at the edge can radically reduce the dependence on Dats Centres.


Surely worth a mention.





.
 
  • Like
  • Fire
  • Love
Reactions: 20 users

Cartagena

Regular
Actually, I can't see a like from Rob Telson. So definitely not us.
Yes it's expected, the big guys are snapping at our heels and other companies like SynSense and Hailo are in the edge space with similar tech, so our lead time appears to have diminished the longer we take in securing more IP deals and contracts. Mid September now and deadline fast approaching so Sean bring on this Gen 2 release already....
 
  • Like
  • Fire
Reactions: 10 users

Xray1

Regular
Very painful to see . Every time i thought it lowest , i bought more and it sinks futher. I have a lot now but the value currently has been halved.
Just imagine what S/P loss values that Peter and the other Top 20 s/holders would be experiencing ......?????
 
  • Like
Reactions: 3 users
That was exactly what Antonio was saying. These things are possible without akida1000 as well and that is why he said it was not robust enough.
This is why I am no longer playing the dot joining game. Unless it is a company that we have an announced partnership with (ie joint development agreement), or a licence has been sold, I'm no longer subscribing to the 'we might be in XYZ' club. Instead I will do what Sean has advised and watch the financials. So far lumpy doesn't even describe them. Flatline seems more appropriate. BUT we've got some potentially big prospects on the horizon later this year - see Renesas for instance.

I haven't seen anyone discussing one aspect of Chris' departure. We have to pay our people comparably to other companies in the tech world, I agree. But if the other companies are delivering results while we remain in limbo sales wise, the sales team will burn out / succumb to the stress of poor performance and jump ship like CS. He knew accepting his role with Brainchip, what the pay packet was (a decrease to his norm, surely) but jumped on board anyway. So I highly doubt his leaving is anything but due to the sales performance (lack of) so far.

Gen 2 availability announcement can't come quick enough. Then there's no more excuses.
 
  • Like
  • Fire
  • Love
Reactions: 20 users

TECH

Regular
Maybe the company (Brainchip) is patiently awaiting the Board of a major company to find an inkwell to dip their feather into
to sign off on an agreement, I personally feel that if we do get the anticipated announcement surrounding AKD II being formally
released onto the world stage, well a double-headed announcement would be the way to go to launch our second child.

Footnote: I am not referring to Tasmania when I say "double-headed" :ROFLMAO::ROFLMAO::ROFLMAO::ROFLMAO::ROFLMAO::whistle::rolleyes:

Have a great Wednesday ! Tech.
 
  • Like
  • Haha
  • Love
Reactions: 38 users

IloveLamp

Top 20
  • Like
  • Fire
  • Wow
Reactions: 13 users
Top Bottom