rated R for big space battles
7 Likes
Time to go to sleep… nighties lovelies!
7 Likes
7 Likes
Many people worry that they won’t have a lover for Valentine’s Day. How come nobody is worried that they’ll still remain uneducated by May 24th?!
7 Likes
Pretty nice example how piece of glass could hold piece of mind.
I wonder if we will make in future some famous persons minds that will be held in glass and consulted even after death.
Like crystal skulls, but actually real and not some fakes.
7 Likes
As with AI, the neural network does nothing like thinking. It literally scatters light through different paths until some of those paths end in the same area and that area is the assigned target for the light… maybe. Because, maybe they’re feeding the light to the network and see where do all the “2” exit the network, and call that area “2”, rather than actually designing a path so the inputs of “2” finish in the intended area for “2”. Like drawing a bullseye around the place where the light beams ended…
Anyway, i’m off to bed, with little light in my room so I sleep rather than think. Nighties lovelies!
PS: did you ever thought that in a way, we sleep at night so we don’t accidentally stumble upon a predator who sees us without us seeing it…? 
7 Likes
I wonder what A.I. sees when it looks at the Rorschach test?
Timezone.
6 Likes
Eventually many things can be seen. Same as with humans, if A.I. is finally similar to humans and can live their own lifes it would be indistinguishable result.
What is experienced by human are many sensory experiences filtered by our system of valuing, system that is effect of previous experiences. With A.I. we feed it visuals mainly now, but with as many experienes as humans have and while independently gathering and valuing, it can be truly personalized, like with humans. The environment makes or breakes, and part of that environment is also the mind itself with its valuing.
6 Likes
5 Likes
Well, obviously, when an AI sees a Rorschach test, it says “this is a Rorschach test”. 
They don’t think. They learn to analyze statistical data obtained from, say, images, and then compare numbers from new image to those of old images and picks the highest probability. But they don’t think and don’t know what they’re seeing neither what it means to our minds. And this is why I’d bet that sooner than later some AI company will run ito seriosu legal trouble after pedos train their AI to generate child p**n. Just tag the iamges with a random tag and feed them to the AI, until it finds the association between the tag and the images being provided and begins exploring the target area, generating images which should look similar to the seeder ones and then human say “right” or “wrong”. With long and careful training, eventually the AI would generate images similar to the original ones, on demand, associated to a tag nobody would search nor type randomly.
FAI, with time, determination and a network of accomplices, you could train an AI to generate red corvettes from 1967 under the tag “ds8h342bsdfjnkd7sdf78rwe46fsdmfsd987dfsiwerofufsdHREIpko”, and each time you asked for “ds8h342bsdfjnkd7sdf78rwe46fsdmfsd987dfsiwerofufsdHREIpko” the AI would gladly generate something similar to a red corvette from 1967. Or whatever immoral monstrosity you had teached it to generate.
And the best part? AIs don’t have a “unlearn” feature.
So out there, it’s a sure bet: someone in the big world is feeding child p**n to some AI and that AI is generating whatever it’s been trained to do… and sooner or later Police somewhere will find out.
Because AIs don’t think, can’t think, and never will think any more than the horse who “knew to count” by reacting to the expression of people when it stomped the hoof the right number of times.
And now i’m of to bed. Nighties lovelies!
5 Likes
Never say never. 
Who knows what is possible with more research.
5 Likes
No amount of research and improvement on canvas, ropes and timbers will get you a flying sailboat…
5 Likes
But nobody is doing A.I. out of canvas, ropes and timbers. We will wait to see where this is going. For now the thing can be more in jeopardy because of the ethics and law tho.
I dont know how viable is fighting progress instead of adaptation, but people will try anyway.
5 Likes
“The future is now, teacher”
5 Likes
Well, they’re pretending to make thinking machines with boolean logic, which is even moee primitive than canvas and wood are to an Avatar titan. We don’t even know what it is to think, but AI smoke-and-mirrors pretends to imitate thinking without thinking, by number crunching reality bites turned to numbers. It’s all an imitation game… and in the case of art AI it’s based on the massive theft of intellectual property.
5 Likes
Just two days ago we had a similar news piece running in local TV, and some people were evaluating the chances to use these tools to improve learning and teach teachers how to use them with that purpose… merrily ignoring the fact that students are using the AI to pretend that they learn rather than actually learn. Because, you know, there’s nothing that could go wrong with trusting your money, safety or health to someone who faked their way into pretending they know what they’re doing…
Trust me, I pretended to be an engineer!
6 Likes
Its yet for that court to decide if that was theft, or allowed usage. Article is clear that it is disputed and unclear, and law is not quick enough with adaptation to faster changing reality. Even A.I. lawyers could start to play a role.
Obviously a learned A.I. will have to be put to the work instead of all those engineers.
I already use A.I. in work. Many people are. The results are everything is being done faster with less people. Especially when there are qualified workers shortages because of society getting older.
Its a process that cant be stopped I think.
6 Likes
But AIs don’t think. They might imitate something someone thought before, but they can’t imagine new stuff. They can’t do problem solving, let alone the kind of problem solving that it’s hard for humans because it does not have a known solution, or the solution has only been tested in a different situation. AIs at best woudl try and do something someone did before and might succeed or don’t based on whatever information they have -but can’t do it when there is information they don’t have; they can’t get that information, not even know that they need that information. AIs don’t think. they imitate some easy stuff we do with hinking, but that’s not actual thinking.
And now I’m off to bed, without thinking. Nighties lovelies!
Also: making-of, kinda
Still a very expensive solution to some problem nobody really knows…
6 Likes
