World's largest neuromorphic supercomputer is switched on

originally posted by: andy06shake
Would you kill your own creator?

If not then why assume any AI would?

My mother has a few issues and has been known to be inefficient at times, don't mean I'm planning on killing the poor Woman.

Many men in history have wanted to kill their fathers. In fact Freud says this is normal.

a reply to: DISRAELI

Says some interesting things about our relationship towards mommy also if im not mistaken.

Are we feeding this AI copious amounts of Cocaine?

a reply to: andy06shake

Are we feeding this AI copious amounts of Cocaine?

Save some for the "Gods"

Haha, I jest... kind of.

a reply to: CriticalStinker

But not really? LoL

a reply to: andy06shake
Thanks.

Perfect, I hoped you use a word like sentinent. Now, define sentinent. Not for me, think about it what that means.

Semantics. Your language is different from mine and thus you express different thought patterns when you think about something. Sentinent translates to feeling stuff around you. Feelings are not a condition for strong AI. Being aware of it´s surrounding is. Now how can an AI be sentinent and not sentinent at the same time?

Is being able to be sentinent a necessary precursor for emotions or vice versa?

I think different when I think in english. It´s a different construct and more important than most think. I´ll post a snipped from my script for you so I don´t have to write it out (again).

Contrary to the popular and wide spread believe, strong AI does not need to have the ability to feel and express emotions to be counted as strong AI. We also know that feelings and emotions can limit our ability to utilize our intelligence to the full extend possible.

Artifical Intelligence does not need feelings or emotions to be artiffically intelligent. Intelligence is the opposite of emotions.

Intelligence is based on knowledge. Emotions are based on feelings. Feelings do not base on knowledge but can be altered by it. Emotions can alter the way we put our intelligence to use.

Example: You know well that it´s a bad idea to go visit your girlfriend now, because if you drive there, you won´t have money to eat next week. Now our feelings tell us you want to see your girlfriend, but your intelligence tells you better not to.

The intelligent solution would be to just not see your girlfriend but instead call her.
The emotional solution would be to run to the car and drive to your girlfriend.

If you do the first, your intelligence won, as you know you will suffer when you make the decission to go. If you still go, that what´s partly defining love, too.

You see, we can´t use simple words, we have to split them into their true meaning although it makes things complicated. Words bear meanings and concepts. Sometimes those concepts are too blurry.

This is why I brought up thought patterns in different languages. It´s a big topic. Why? Because AI has to learn and train. Since we can´t only communicate in mathematics with it, it needs to establish some kind of thought pattern system.

Analyzing those conditions and relationships was a big part of my former job.

a reply to: verschickter

It's not that AI won't require feelings or have such, just that said feelings will be completely alien to our own given the very different nature of the AI.

Questions such as defining what is sentient, conscious, self-aware, is exactly what machines such as this neuromorphic supercomputer are being built to answer.

Language is just communication most of which it's not even verbal where humans are concerned. AI language will be mathematical, at least to begin with.

Sounds like you had an interesting job.

a reply to: CriticalStinker
It´s their work that inspired me to use drones as the physical sensor and output interface in 2013.

Notice how they use cameras for positioning, I used a xbox kinect that works like a radar and ultrasonic detection on the drones to find the edges, compared with the acceleration sensor. They do need to calibrate now and then by hovering over a lined red laserpointer to get their references correct on the X+Y axis. Z axis was done by a bottom mounted ultrasonic sensor with ranges up to 3m and a accuracy window between 1-3cm/reading.

They are establishing a connection via class1 bluetooth modules for the datastream and WIFI for the video to my workstation which is doing all the "brainwork". Drones itself are dumb as a untouched IC.

PS: I know I constantly write artifical (artificial) wrong.. It´s something I´ll never get out of my vocabulary probably. Yes, I am a bit dyslexic in the terms that when I learned the word wrong, I´ll never get it correctly, even though I recognize after the fact, sometimes.

...nothing to see here (maybe you got mail?)..

a reply to: verschickter

The Kinect platform was an unfortunate piece of technology that just came too soon.

It was the technology Microsoft was working on to take out physical interface like mouse and keyboard and was just too ahead of it's time.

It was the technology that was behind much of the interface in the movie "Minority Report".

I was happy when Microsoft tried to mitigate losses on it by integrating it with video games.

I was thrilled when they encouraged people to use it for anything they could, essentially open sourcing it.

I'm sure you know how it works, but for those who haven't visually seen how it works like radar...

This can be watched on mute, it is a kinect in a room filmed in night vision, capturing the IR.


It can be coded to give AI depth perception and a 3D representation of it's surroundings.

But can it run fortnite?

a reply to: CriticalStinker

It can be coded to give AI depth perception and a 3D representation of it's surroundings.

If you spin it and process the information in tandem with the angle, you can compile a 3D depth map for the whole room and then the fun begins.

Think of the kinect being a rotating radar transceiver. If you introduce a sinewave motion to the rotation, you can elegently scan the whole room in nearle 360° freedom to every side.

a reply to: Lucidparadox

Probably not to be honest.

Depends on the OS.

a reply to: verschickter

Mmhmm.

Also add in other forms of radar and they will be able to perceive environments through certain barriers.

It should come as no surprise the US military has small warfare nano drones that can see through walls.

There are people who theorize that one day our immediate environment will be so saturated with these types of drones their will be a neural network of most of the world, to compare it to simple terms for some who are not well versed... Think of it as a real time google maps, where the whole world is constantly the input to a live stream.

a reply to: CriticalStinker
I would if not for energy source problem. I would need to scale them up and then I could not use my barn anymore.

Catching a couple hundred gram quatrocopter out of the air with a fishing net on a bamboo stick is one thing... I would run into similar other problems like energy density and weight. It´s a scaling problem.

There are people who theorize that one day our immediate environment will be so saturated with these types of drones their will be a neural network of most of the world, to compare it to simple terms for some who are not well versed... Think of it as a real time google maps, where the whole world is constantly the input to a live stream.

One word: Noise polution

Okay, those were two ^^

a reply to: andy06shake

largest neuromorphic supercomputer, designed to copy the workings of the human brain and unlock its secrets, and it’s been switched on for the first tim

That right there is how I think we will get to TRUE AI (AGI then ASI).

In the video below, if he drew a mustache on the Rosalind Russell bobblehead, would the object then be identified correctly by the supercomputer? Or would it give you a 'best guess'?

a reply to: InTheLight

Its always going to be the best guess based on the information available.

originally posted by: andy06shake
a reply to: InTheLight

Its always going to be the best guess based on the information available.

So if the mustache was drawn on only in November and removed for the remainder of the year, would the supercomputer at some point make the 'best guess' that it is a Rosalind Russell bobblehead with a Movember mustache? Just how well can this computer put together synchronicities (coincidences in time), I wonder?

a reply to: InTheLight

If it can determine what objects are then i imagine it won't have much issue with the 24-hour clock.

I'm aware of Carl Jung's concept of synchronicity as would the machine be because it would assimilate data much in the same manner as we do, only much more efficiently and at far greater pace i imagine.

It a relevant question all the same.

Say it can reconcile and understand synchronicity, what do you imagine the implications to be?