New Rob Telson Podcast January 19, page-14

Hi all, have just had the opportunity to listen to today’s Podcast from Utilising AI and for those that haven’t here is a few of the key points I picked up.

Rob is a champion, our Sales & Marketing could seriously not be in better hands. With the recent price dynamics it is hard to believe we are only getting started yet here we are! Mind blowing stuff really.

It’s funny you know, out of all the information and excitement presented there was one quote which really stuck out for me, which I think captures the essence and simplicity of what Akida and one-shot learning can do but others can’t. Have a listen, it was in relation to the factory floor and I think you will understand. It was this:

“bananas go left, oranges go right”

Note, I was typing while listening so please excuse any errors/omissions, context and quote inaccuracies etc. This is intended as a summary on my end only and I encourage everyone to tune in and form their own opinions when they have a spare half an hour. Most importantly do your own research.

Summary points as follows:

·BrainChipreceived a lot of attention from the previous Podcast

·Coupleof subsequent AI Field Days where BrainChip stole the show

·Thebig question coming from the Podcast and Field Days was ‘when can we get ourhands on one of these chips’

·Robis responsible for sales, has been with BrainChip for just over a year and ahalf and has seen this company flourish

·Neuromorphicdifferent to the AI architecture which is available today (Accelerators, DeepLearning, GPU etc)

·EdgeAI Inference Market - $46b potential as of today (growing)

·Robestablishing channels from product development to commercialisation

·Robadvised that BrainChip has been on a ‘rocketship of excitement’ and a lot haschanged within BrainChip since the last Podcast (These are exciting times, goodtimes for BrainChip)

·Startedby licencing IP to a couple of key majors in the industry – Renesas & MegaChip

·Beyondthat - Chip validated (now a full production chip – AKD1000) and introduceddevelopment systems (Shuttle/Raspberry PI – plug and play technology)

·Alot going on from a business standpoint as we move through commercialisation

·Excesspower, time, energy via traditional computation, Akida extremely low power(microwatts only)

·Neuromorphicarchitecture - 3, 4, 5 functionalities going on at any one time via Akida

·Industrywants lower power, faster processing – Rob sees BrainChip really impacting the $46bmarket in this space

·Latestdemos available on BrainChips You Tube channel are going to ‘blow your mind’ –designed to show you how we can apply this practically on a daily basis

·Machines(ie. Coffee machines) will have the ability to recognise a person andrequirements based on profile, refrigerators which can smell etc

·Earlystage, no availability to consumers yet, but ‘we’ve all been round the block sowhen this turns the corner it’s going to turn into a tidal wave, not a tidepool’

·Whatis it that makes BrainChip special as opposed to AI everywhere? It is theneuromorphic architecture, traditional AI not capable of doing so, especiallyfrom a power consumption perspective

·Bytraining on the chip and not the cloud or re-training a network you are savinga tonne of time, a tonne of energy and a tonne of cost – having the capabilityto do this on the device becomes a tremendous advantage

·Robadvised that we are at the very early stage of this entire process ‘tip of theiceberg and I strongly believe that’

·Factoryfloor - One shot learning simplifies the entire industrial applicationenvironment - “bananas go left, oranges go right”

·Vibrationalanalysis – sound recognition and vibration recognition is unique and critical(some machines worth hundreds of millions), easily implemented with Akida – we seethe industrial environment taking off

·Medicaldevices – Covid detection; breath markers and other analysis (Akida showedaccuracy levels of over 95%). There is going to be devices out there at somepoint that are going to be able to recognise disease – to learn on the edgewill mitigate scientist involvement

·Akidaoperates the way people wish AI operated (incremental learning)

·Eventbased system (real time spikes – very different from other machine learning)

·Neuromorphicarchitecture - zeroes are non-events, no requirement to process

·Whatif every sensor is a smart sensor? Really exciting stuff

·BeneficialAI – BrainChip want to see good things come from this (remote analysis, remotemedicine examples). Very excited about these types of discussions.

·Productcontinuing to be validated, growing level of acceptance, eventually widelyaccepted and adopted

·Requireactive academic involvement through world of AI

·Todayis an exciting day for BrainChip – full commercialisation of AKD1000 ProductLine, full stack of technology available for all different levels of peoplewanting access to AI

·MiniPCIE board – we want this into hands of as many people as possible ‘before wesell out, come get it’ This is a BIG deal as per Rob’s comments

·Volumeproduction of development boards available – contact us to purchase in volume

·BrainChipis the first neuromorphic architecture to be fully commercialised and this isthe tip of the iceberg (stated again!)

·Futuretechnology really exciting

·ThreeQuestions:

-Chris:When will we see a full self-driving car that can drive anywhere, any time? Rob– Good question, I know they’re on it! Some companies have started grid citiesall over the world to start self-driving environments. 5-10 years we will havefull self-driving vehicles in some cities around the world

-Stephen:Are there any jobs that will be completely eliminated by AI in the next fiveyears? Rob – Good question as well. Rob looks at it differently, doesn’tbelieve AI will eliminate jobs in entirety, human element still needs to beaddressed (people buy from people). Broad perspective – emotion, rationalness,AI can’t duplicate.

-GirardKavelines: What is it that scares you about AI in today’s industry? Rob – wantsto see it used in a productive environment. Technology saturation in general.