BRN Discussion Ongoing

TheFunkMachine

seeds have the potential to become trees.
https://brainchipinc.com/brainchip-holdings-momentum-keeps-building/ new research


https://brainchipinc.com/brainchip-holdings-welcome-to-the-revolution/ old research


New Pitt street resesrch article. Valuation stays the same @$1.50 slightly disappointed with that as a few key things have happens since their last valuation @$1.50 but I guess they are trying to be realistic.

Good summery, but not as good as their first one in my opinion. I think they are just building on the momentum of their last research and has not included a lot of the earlier achievements in this article. So I recommend reading both together to get a complete perspective on Brainchip.
 
  • Like
  • Fire
Reactions: 21 users
  • Like
  • Thinking
Reactions: 10 users
I thought this was interesting. This fellow worked at Hewlett-Packard Back then for around 12 years and as you will see has a belief in neuromorphic computing. If you only give him a credibility rating of 50% to account for the issues around social media what Brainchip has achieved to date is well beyond astounding.
My opinion only DYOR
FF

AKIDA BALLISTA
1648420855753.png
https://www.quora.com/How-far-along-is-neuromorphic-computing
How far along is neuromorphic computing?
How far along is neuromorphic computing?
It’s no less than 20 years out. I regularly attend IEEE meetings related to the core technologies and its manufacture. I also work with organizations likely to be the first to adopt.
We haven't figured out the basic details of the devices required to make neuromorphic neural nets yet. We have ideas of what likely will work but these devices are still Big-Research, Little-Product Development (R & d), rather than Little-R, Big-D.
The devices that have been built are not reliable enough to be used commercially yet. There are issues with manufacturability and quality. There is infrastructure uncertainty and market uncertainty still - everyone promises they’ll buy product but until they actually do, they could be blowing smoke up your posterior.
We don’t yet have test equipment capable of testing many of the fundamental performance metrics of these devices. Developing the latter will still take 5–10 years if we could get a firm spec for what is required and the numbers we have so far tell us WE need several years of R&D ourselves as upstream suppliers to these researchers just to figure out how to make such equipment and make it useable.
We have at least started toward this technology but there are many, many critical path issues unsolved and undefined still. It’s not an economically viable reality yet. It’s still pure research. This is actually 100% normal for early stage pre-technologies. In general, you have a 20 year timer that only starts once you’ve proven all the key underlying technologies.
The short answer: getting a PhD in the area (which takes 5–10 years) is still a good idea if you want to do that. Many PhD-level questions still need to answered first/yet!
 
  • Like
  • Fire
  • Thinking
Reactions: 23 users

Foxdog

Regular
So one might surmise, from this article, that Brainchip are up to 18 years ahead of the competition 🤔 I wonder what the share price will be in 18 years...... 😂
 
  • Like
  • Haha
  • Love
Reactions: 14 users
I thought this was interesting. This fellow worked at Hewlett-Packard Back then for around 12 years and as you will see has a belief in neuromorphic computing. If you only give him a credibility rating of 50% to account for the issues around social media what Brainchip has achieved to date is well beyond astounding.
My opinion only DYOR
FF

AKIDA BALLISTA
View attachment 3265
https://www.quora.com/How-far-along-is-neuromorphic-computing
How far along is neuromorphic computing?
How far along is neuromorphic computing?
It’s no less than 20 years out. I regularly attend IEEE meetings related to the core technologies and its manufacture. I also work with organizations likely to be the first to adopt.
We haven't figured out the basic details of the devices required to make neuromorphic neural nets yet. We have ideas of what likely will work but these devices are still Big-Research, Little-Product Development (R & d), rather than Little-R, Big-D.
The devices that have been built are not reliable enough to be used commercially yet. There are issues with manufacturability and quality. There is infrastructure uncertainty and market uncertainty still - everyone promises they’ll buy product but until they actually do, they could be blowing smoke up your posterior.
We don’t yet have test equipment capable of testing many of the fundamental performance metrics of these devices. Developing the latter will still take 5–10 years if we could get a firm spec for what is required and the numbers we have so far tell us WE need several years of R&D ourselves as upstream suppliers to these researchers just to figure out how to make such equipment and make it useable.
We have at least started toward this technology but there are many, many critical path issues unsolved and undefined still. It’s not an economically viable reality yet. It’s still pure research. This is actually 100% normal for early stage pre-technologies. In general, you have a 20 year timer that only starts once you’ve proven all the key underlying technologies.
The short answer: getting a PhD in the area (which takes 5–10 years) is still a good idea if you want to do that. Many PhD-level questions still need to answered first/yet!
Assuming you have read the above this fellow in 2020 would clearly have been aware of IBM - True North, Intel - Loihi, HRL - Synapse, ETA - chip, SpinNaker, Human Brain Project etc, etc.

Working at Hewlett-Packard a company of sufficient size and reach that all of the players who had advanced neuromorphic products would have been pitching to yet he is maintaining at least 20 years. At about the time he was posting Brainchip was just starting to move out of stealth mode.

HE WOULD CLEARLY HAVE PLACED BRAINCHIP A BIT MORE THAN THREE YEARS AHEAD BACK THEN IF HE HAD KNOWN.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 18 users
Who is looking forward to Ford breaking cover and saying they have “Akida Inside” 😁.

Surely they want to be looking as cool and EV competitive as Mercedes soon!
 
Last edited:
  • Like
  • Love
Reactions: 20 users

Slymeat

Move on, nothing to see.
Different technology. ResApp detects small but different sounds from the alveoli via a forced cough into an app. Like Shazam in a sense. Diagnoses a raft of respiratory disease like pneumonia, bronchitis, COPD and asthma. Now they also screen for COVID. Fast, accurate, scalable and cheap. I hold this company.
I did hear of the ResApp approach being debunked somewhat due to the fact that many covid sufferers do not have a cough. Many sufferers are completely asymptomatic. DISC: I hold shares in ResApp.

The Nanose /diaNose approach of detecting Volatile Organic Compounds seems, IMHO, to be where the future lies.
 
  • Like
  • Fire
Reactions: 13 users

Slymeat

Move on, nothing to see.
didn't want to confuse anyone. but since we have a partnership with AirForce about the development of a radar system. I thought to share an article where you can see where this can go. Which by the way is new and is currently being discussed a lot on TV.
Thanks for clearing that up @Dallas. I must admit that I did originally fail to see the link.

I do hope Akida is involved in its AI.

Now that WOULD be sweet.
 
  • Like
Reactions: 4 users

Dhm

Regular
I did hear of the ResApp approach being debunked somewhat due to the fact that many covid sufferers do not have a cough. Many sufferers are completely asymptomatic. DISC: I hold shares in ResApp.

The Nanose /diaNose approach of detecting Volatile Organic Compounds seems, IMHO, to be where the future lies.
Debunked by people who don't understand the tech and sensitivity of the lungs when they get the virus. A forced cough still has the alveoli irritation that Covid causes and therefore the algorithms detect. Anyway this isn't the place to comment on other companies.

As you say, the Nanose Brainchip technology will be where much of the future lies.
 
  • Like
Reactions: 4 users

Slymeat

Move on, nothing to see.
Debunked by people who don't understand the tech and sensitivity of the lungs when they get the virus. A forced cough still has the alveoli irritation that Covid causes and therefore the algorithms detect. Anyway this isn't the place to comment on other companies.

As you say, the Nanose Brainchip technology will be where much of the future lies.
The post I read was written by a medical doctor and that was sufficient for me.

Being a long suffering ResApp shareholder, I did have my rose-tinted glasses firmly on at the time too.

And although I was merely replying to your original post on a technology that does not involve Akida, I do believe it is OK to mention other technologies/companies that may be in competition to technologies that incorporate Akida. That creates a healthy discussion.
 
  • Like
Reactions: 6 users

LuWil

Regular

Cancer Diagnoses From Exhaled Breath With Na-nose​


Completion date: 31/12/2022
 

Attachments

  • D7399A04-15B1-4AD3-B0F7-15E939832180.jpeg
    D7399A04-15B1-4AD3-B0F7-15E939832180.jpeg
    251.2 KB · Views: 55
  • 7E82A483-3BCB-4E68-8AD4-2F16967FA773.jpeg
    7E82A483-3BCB-4E68-8AD4-2F16967FA773.jpeg
    265.5 KB · Views: 47
  • 60B38A41-282E-4828-8281-210D2D83DFBE.jpeg
    60B38A41-282E-4828-8281-210D2D83DFBE.jpeg
    190.2 KB · Views: 46
  • Like
Reactions: 2 users

Diogenese

Top 20
I thought this was interesting. This fellow worked at Hewlett-Packard Back then for around 12 years and as you will see has a belief in neuromorphic computing. If you only give him a credibility rating of 50% to account for the issues around social media what Brainchip has achieved to date is well beyond astounding.
My opinion only DYOR
FF

AKIDA BALLISTA
View attachment 3265
https://www.quora.com/How-far-along-is-neuromorphic-computing
How far along is neuromorphic computing?
How far along is neuromorphic computing?
It’s no less than 20 years out. I regularly attend IEEE meetings related to the core technologies and its manufacture. I also work with organizations likely to be the first to adopt.
We haven't figured out the basic details of the devices required to make neuromorphic neural nets yet. We have ideas of what likely will work but these devices are still Big-Research, Little-Product Development (R & d), rather than Little-R, Big-D.
The devices that have been built are not reliable enough to be used commercially yet. There are issues with manufacturability and quality. There is infrastructure uncertainty and market uncertainty still - everyone promises they’ll buy product but until they actually do, they could be blowing smoke up your posterior.
We don’t yet have test equipment capable of testing many of the fundamental performance metrics of these devices. Developing the latter will still take 5–10 years if we could get a firm spec for what is required and the numbers we have so far tell us WE need several years of R&D ourselves as upstream suppliers to these researchers just to figure out how to make such equipment and make it useable.
We have at least started toward this technology but there are many, many critical path issues unsolved and undefined still. It’s not an economically viable reality yet. It’s still pure research. This is actually 100% normal for early stage pre-technologies. In general, you have a 20 year timer that only starts once you’ve proven all the key underlying technologies.
The short answer: getting a PhD in the area (which takes 5–10 years) is still a good idea if you want to do that. Many PhD-level questions still need to answered first/yet!

Well Jeff was clearly talking about analog NNs (ReRAM/MemRistors).

"There are issues with manufacturability and quality."

One of PvdM's great insights eliminated these problems at one stroke - a digital NN that uses standard CMOS manufacturing and binary digital signals. While analog neurons are a close theoretical analogy to real neurons, manufacturing variability is still one of the issues that analog NN makers still have to address because any variations in component dimensions alters the amplitude of the analog signals because the strength of the various current signals are added to determine when the firing threshold is reached. The current strength is determined by the tolerance of the resistors/capacitors. So, by way of an illustrative example only, if there were say a 10% variation in signal strengths, adding just 5 analog signals could produce a chance of up to 50% error in the sum of the signal strengths.

On the other hand, binary digital circuits have a high tolerance for manufacturing variations as each switching element produces either a 1 or a zero output. The supply voltage of a CMOS transistor would be much higher than +/- 10% of the trigger threshold voltage.

As for the test equipment, there are the various image databases against which the performance of NNs can be gauged.

I'll leave it to others, like Socionext, Mercedes, Renesas, MegaChips, Valeo, NASA ... to assess the market uncertainty and economic viability.

I blame ill-conceived KIPs for the siloization of work. In the old days, you could always peek over your neighbour's fence to see what was happening elsewhere. I'm pretty sure Crick & Watson did not have KIPs when they were tossing ideas around in a Cambridge pub.
 
  • Like
  • Fire
  • Love
Reactions: 17 users

GDJR69

Regular
Most certainly there is money in automotive for Brainchip but remember the acting CEO Peter van der Made, the then VP of Sales and Marketing Rob Telson put forward last year that they were going to saturate the Edge market. Not we are going to saturate an Edge market. They also used the word ubiquitous to describe what they had planned for AKIDA and did not qualify it as just being the automotive market where this was going to happen. The new CEO Sean Hehir has not stated otherwise than his one comment about focus.

Then we have both Renesas and MegaChips who cover more than just automotive and MegaChips actually list mobile phones as a potential use case for AKIDA technology.

Then there is the Nanose SniffPhone idea which clearly needs a mobile phone to fulfil the description.

There are so many other Smart Health applications which can be processed through mobile phones but if they cause the phone to need a recharge every 30 minutes are likely not going to be picked up but a 3 to 5 day phone life they become the marketing edge for a phone company.

I was reading an article yesterday probably posted here by someone but it was basically saying that do not expect your new phone to be much more revolutionary than the last one because battery life was greatly restricting added innovations.

I am not saying Brainchip is in bed with a mobile phone company but I am saying it cannot be discounted on the basis of known facts.

Also on the Mercedes EQXX front remember the concept also includes turning it into a technology hub linked to your Smartphone, your Smartoffice, your Smarthome it could be argued that just this one known fact makes it more likely than not that phones using AKIDA technology are already on the table.

My opinion only DYOR
FF

AKIDA BALLISTA
Throwing the question out there, but how much (if any) revenue from these activities do you think we will see in the next quarterly? Or are we still too early for any significant revenue ?
 
  • Like
Reactions: 3 users

Foxdog

Regular
Who is looking forward to Ford breaking cover and saying they have “Akida Inside” 😁.

Surely they want to be looking as cool and EV competitive as Mercedes soon!
I'm looking forward to Ford vs Mercedes in a hot lap. Then imagine an entire grid full of autonomous race cars each with AKIDA - perhaps not so fanciful....
 
  • Like
  • Wow
  • Haha
Reactions: 9 users

MrNick

Regular
I'm looking forward to Ford vs Mercedes in a hot lap. Then imagine an entire grid full of autonomous race cars each with AKIDA - perhaps not so fanciful....
That's sounds about as enjoyable and gonad-fizzy as watching Formula E.
 
  • Like
  • Haha
Reactions: 6 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
Voago Linkedin 2 days ago.



Screen Shot 2022-03-28 at 12.03.07 pm.png





This made me think about FF's post about Nokia being chosen by NASA to deploy the first LTE/4G communications system on the Moon. I'm certainly no expert but I'm assuming that the hardware that Nokia intend to use will need to be space-hardened. Vorago, Nockia and BrainChip - a match made in heaven (or on the moon) one would think. :p



(Extract 1 - BrainChip and VORAGO Technologies Agree to Collaborate through the Akida™ Early Access Program)

Louis DiNardo, BrainChip CEO commented, “We are both excited and proud to participate in this Phase I program with VORAGO Technologies and support NASA’s desire to leverage neuromorphic computing in spaceflight applications. The combination of benefits from the Akida neuromorphic processor and a radiation-hardened process brings significant new capabilities to spaceflight and aerospace applications”.


Bernd Lienhard, VORAGO CEO added, “We are thrilled and honored to partner with BrainChip to harness the radiation hardening capabilities of our patented HARDSIL® technology for the Phase I program with NASA. Our ongoing mission of creating components with increased availability and unmatched solutions in aerospace and defense applications paired with the Akida neuromorphic processor will create unprecedented standards moving forward in the industry.”




(Extract 2 - Thierry Klein, head of Enterprise and Industrial Automation Research Lab at Nokia Bell Labs, is addressing the gap between communication technology on the Moon and technology astronauts have access to on Earth. )


VIA SATELLITE: What are the challenges to building a network on the Moon? It sounds like science fiction. How difficult will this be?​


Klein: The main challenge is the environment. We know how to build the core 4G and LTE systems and how to transmit data wirelessly. The first challenge is all of the space hardening, whether it is for the launch of the rocket, the landing on the Moon, or the extreme mechanical stress that the equipment will be exposed to in terms of shock, vibration, and acceleration.


We also need to harden the equipment for the unique space environment, like the extreme hot and cold temperatures, the vacuum, radiation, and the impact of radiation on the equipment. That is a really difficult problem.


I would say the third challenge comes from the operational environment. We need to be very meticulous about size, weight, and power consumption. We need to worry about integrating our network equipment with the lunar lander and integration with the lunar rover. We need to have a fully operational system that is reliable and robust because we can’t send a technician to the Moon to make configuration changes or reboot the system. The system has to be completely autonomous and we need to be able to remotely monitor, manage, and configure it from mission control on Earth.



 
  • Like
  • Fire
  • Love
Reactions: 30 users
Throwing the question out there, but how much (if any) revenue from these activities do you think we will see in the next quarterly? Or are we still too early for any significant revenue ?
I recently had a private conversation with another poster on this question and extracting the relevant part the following is my view which I referred to Tony Dawe at the time to be sure I was not completely off the mark and he responded that he saw no reason to change my response:

"When Sean Hehir was interviewed earlier this year he answered Yes to a question about would revenue start to ramp up this year.

I personally have never expected explosive income growth this year but there are stages in a company like Brainchip's development.

1. The first is the development phase and during this phase there is no income and going concern status is determined by capital raises.

2. The second stage is becoming a commercial entity and then the income is a mix of capital raises and some income.

3. The third stage is becoming a going commercial concern where its continued operation is being funded from income and capital raises if required are only for special projects or acquisitions and not to meet day to day operating expenses.

4. The fourth stage is when the income increases beyond simply funding ongoing operations and questions of what shall it do with profits arise.

The messages I have been hearing from Brainchip have communicated to me that we are in the second stage and the third stage will start second half of this year.

I personally had no expectation of explosive revenue occurring this year. What I do expect from Peter van der Made's explosive sales statement is that we will see an increasing number of current EAP's breaking cover and more and more logo's of companies appearing in the Early Adopters and Licensee columns and license fees will be seen in the balance sheet. We will also likely see other companies making announcements similar to Mercedes."


My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 65 users

Slade

Top 20
Too long down rabbit holes for me. What I learned was that Driver Monitoring Systems are likely to be mandated just like seat belts. With the need for low power and privacy I would say Akida is in pole position.
From our website:
The Akida AKD1000 Edge AI processor can also monitor the passenger compartment, checking driver alertness with the aim to reduce human error. Privacy is assured without the need to upload images to the internet.
 
  • Like
  • Fire
  • Love
Reactions: 22 users
Top Bottom