I thought I would appeal to the independent arbiter:Hi @Diogenese
I just checked my post and have made a correction which I have marked in orange text. (I am certain I used "neuro" first time around but I think the correction was made for me to neural.)
The tell is that when you listen closely to what he says he does not say "neuromorphic' network engine nor does he say 'neural' network engine he labours in his speech and uses the words "neuro" network engine pulling up short of either 'neural' or 'neuromorphic'.
My opinion only DYOR (listening)
FF
AKIDA BALLISTA
PS: I have now listened multiple times on both my phone and on my PC and still hear the same word 'neuro'.
30:52 with the German company called baloof and ready robotics for machine learning on manufacturing you know adding machine 30:58 learning to the robots so then you know exactly the positioning and you know for 31:03 a wearable application they are a company called No watch that literally doing no watch it's showing how much you 31:11 are stressed and again you need to train a lot of models that you knows taking 31:16 all of the sensor from your hand if you are too stressed or not and this is a a mental well-being application and the so 31:24 you see industrial you see wearable Edge compute application high-end in this 31:29 case the request was we want to count people in a traffic light in India Okay 31:35 so a lot of things you need an Nvidia GPU for that so this is where cost is 31:41 doesn't matter you just need to execute with supporting Nvidia as well and for example a non-invasive glucose think 31:48 about it this is revenue revolutionary application from a company called No Labs how much models you need to train 31:54 but still need to be handled application low cost low power and in this case it 32:00 was no* network engine so you see different type of application exactly the same is I shared before 32:06 unfortunately I cannot tell who is who but in then each of them won one of that 32:12 okay all multiple and this is good example other companies so we mentioned already 32:18 a wearable or ring I believe some of you know the ordering and just today Samsun
Well it looks like we were both mistaken - after all, when is a computer ever wrong?!
It is true that an industrial robot will have power laid on, so it does not need Akida's parsimonious power usage, whereas the wearable glucose meter does.
The glucose meter needs to be trained, whereas the robot learns. I doubt that the glucose meter uses ML, because the models would need to be verified in the lab, and ad hoc changes would not have FDA approval.
So there are arguments both ways, so I'm going with the Scottish courts - not proven.
* Late extra: I see Murphy hasn't retired yet.
Last edited: