Artificial Intelligence

the demise of us all

 

If Kurzweil is so smart, why hasnt he achieved anything of any importance? I know what it sounds like to say that, and I am speaking relatively.

Because he's on the talking tours, means he's not the guy in the lab burning the midnight oil that figures out that one advancement everybody needs to move forward. Text to speech, speech recognition, if you think of all the people working on that over the last 30 years, dozens or hundreds for each company, hundreds of companies, maybe up to 100,000 "smart" people working on it in that time frame for stuff heavily dependant on processing speeds anyway....why did it take so fuckin long? and we still dont really have speech recog that really works.

The cult of celebrity, all guys like him there are at least a dozen under him doing the real work whom you dont hear about, he gets the credit for, even with those under him, spend their whole careers, have an impressive list of PhDs and yet still only manage one small advancement in their whole careers.

If he is so smart, then why cant he just sit down and nut out an AI engine all by himself? Why doesnt he own google for that matter, why couldnt he or anyone else at Google see Twitter or Facebook coming?

Its all talk

 

So, according to your strange logic, Tim Berners-Lee is a nobody because he isn't rich?

 

Possible,but it will have no feelings,and will therefore not be sentient in my opinion.Such a machine may well be able to mimic super-intelligence,but still not experience itself as a conscious entity.

 

Well, he didnt actually invent the internet.

100,000s of people did that from creating the first transistors throught to creating all the database servers

 

I just want an iBabe, haven't decided what color yet.

https://www.youtube.com/watch?v=iRGnqH7nkow"]iBabe scene from "Movie 43" - YouTube

 

You are right, he didn't invent the internet. He invented the World Wide Web, almost single-handed.


Sent from my iPhone using Tapatalk

 

How would u be able to tell externally? What if you asked the computer how it feels and it told you? Would you believe it, or would it just be pretending?

 

Even the best experts admit we are nowhere close to anything that can legitimately be called AI.
All we have now is basically a bunch of and/or switches embedded in silicone controlled by algorithms matching up the most probable answers from a data base.

 

True. However, Kurzweil thinks that we will develop non-destructive brain scanning that will transfer its information to neural networks, thus providing a huge shortcut to the software-based solutions of today.

 

I don't know, I read a quote from a neurologist the other day, "Imagine the universe the way we can see it today, everything we see with Hubble telescope, all the galaxies, stars, planets etc...the neuro-connections in the human brain is thousands of times more complex than that"

http://www.dailygalaxy.com/my_weblo...alaxy-carl-sagan-and-roger-penrose-video.html
Each cubic millimeter of tissue in the neocortex, reports Michael Chorost in World Wide Mind, contains between 860 million and 1.3 billion synapses. Estimates of the total number of synapses in the neocortex range from 164 trillion to 200 trillion. The total number of synapses in the brain as a whole is much higher than that. The neocorex has the same number of neurons as a galaxy has stars: 100 billion.

Most of the neurons in the neocortex have between 1,000 and 10,000 synaptic connections with other neurons. Elsewhere in the brain, in the cerebellum, one type of neuron has 150,000 to 200,000 synaptic connections with other neurons. Even the lowest of these numbers seems hard to believe. One tiny neuron can connect to 200,000 neurons.

 

Good post.

 

i would love to be able to download my internal visualizations and dreams directly to my 3d illustration software. this however does not require a.i., just super bandwidth and massive storage. still beyond current capabilities, but foreseeably possible.

 

Simply because it is way beyond current capabilities doesn't mean much. Moore's Law tells us that capabilities double about every 18 months. That leads to exponential growth. Compare the specs on the first Mac to the ones today.

 

Yeah but the laws of physics is the limiting factor in Moores law, and that wall has almost been reached.
Moores law isn't some esoteric magic, it was more an educated "guesstimation" based on extrapolation of the developments in manufacturing technology.
It's funny how people act like Moores law is some law of nature or something. LOL

Comparing the specs from then to today is silly considering that even with the 22 nanometer tracings that are being used now, we are still a very,very,very long way from packing the density of circuits into something as compact as a brain needed to truly achieve artificial intelligence that rivals or even comes close to a humans.

Not saying it can't be done, but Moores law will have nothing to do with it as it relates to current computing technology.
To achieve artificial intelligence mimicking a human will most likely require quantum computing technology, and that is still a ways off before it becomes a viable tech.

 

what if they started stacking transistors in a 3D configuration? Just adding one extra layer on top would easily double the density, and there would be more possible circuit paths.

 

Well, it might turn out that our brain is so interconnected, as Gong points out, that they only have to transfer a fraction of the information from a brain to a computer neural network for it to function rather well. Sort of like a hologram.

 

They already do that.
There are devices that allow people to see by tapping directly into the visual cortex, as well as hearing.
They are also making great strides with prosthetic limbs that tap directly into the nerves for the missing appendage.
I have even seen a set-up where a guy was controlling a "mouse" with just thought via a device that read the electrical impulses through his scalp.
Those are all still in the experimental trial stages, but have more than proven the concept.
So human-machine interfaces are a reality and will be seen more and more, first in medicine, then who knows.

 

true enough, but intelligence, and sapience, or even sentience, are rather separate things, and i really don't expect or believe one to automatically spawn the other.

the intelligence of non-creative humans is somewhat artificial as it is.

 

Cant do that.
I hear that heat becomes a serious issue when you do that. Which is why I feel the key to AI is wetware. Another thing they are considering is slime molds or something used for computing. Even then there are limitations. Perhaps we will advance our minds but will we become some kind of superintellegence? Hard to tell, but I wouldnt think that AI would neccessarily mean godly intellegence. Perhaps we will just become more aware of whats in our skull and how it works to the point where we can create and modify it in ways that are beneficial to us all. Such as: Curing dementia, skizophrenia(however the hell you spell that), depression, grief(imagine being able to forget that someone close to you died! you would have to forget the person of course, but what good is this memory to you at a certain point other than a source of depression), crime/aggression(yes get convicted of murder, we get to control your mind serves ya right we could even READ your mind to determine guilt rather than that silly lie detector test).
We may even come to a point were instead of enrolling in college courses for knowledge we can just download knowledge. Can you imagine how many resources could be saved??? Think about all that is wasted, all the resources we have to throw out after a student has used the models and the fake tools in class to learn how to do the job for real. Also the time and the cash. Download it to your brain instead. Just watchout for brain hackers, they could do way more than just ruin your computer.