Nobody has actually created an AI and if I ever see you use the word A.I. to describe some shitty markov chain again I'm going to punch the shit out of you
An a.i. must have beliefs, desires and intentions that impact the decisions that it makes.
OTHERWISE IT'S NOT A FUCKING A.I
I'M SICK OF THIS SHIT!
bro
Cool.
When exactly did I sub to your shit blog tho?
'AI' isn't the same as 'general AI'
AI is a large collection of different fields and techniques, including expert systems, machine learning, language processing, general AI, etc
No it isn't
AI is nothing but a marketing bullshit buzzword you faggot
This is true. Whatever AI is marketed today, it's just statistics. For example, take this huge hyped neural networks: Pictures are array of rgb and using many pictures, one can predict new array of rgb which is a new picture similar to the training pictures. Nothing like a true thinking system at all. And people are like whoa, the ai made a picture.
And btw, everytime I hear tensorflow I get triggered. They just use an array of array of rgb. Not a tensor.
You're right. We're not even close to creating AI, yet people are talking about it like it's gonna happen within 5 years. I gave up trying to explain it to people. They rather believe a flashy headline by some faggy tech blog.
Honestly Neural Networks are amazing. But they are not fucking AI. There is genuine AI research and development going on, but you will not read about it in some trendy techfagnews blog. I gave up on people a long time ago, hyping up fucking glorified decision trees as if they were the second coming of Christ and their Waifu incarnate.
The vast majority of people who have any hint of technical knowledge know human level AI will take decades if not centuries to develop.
A few people believing it's going to happen within five years are a very small minortiy.
You shouldn't use those people as proof that intelligence isn't indistinguishable from sufficiently advanced applied statistics.
BTW what's up with critics of AI focusing on Markov chains? You know you can't generate any worthwhile "array of rgb" using a Markov chain right? You know the eyes and the brain represent pictures as an array of arrays of firing rates on the neurons of the optical cortex, right?
But Turing machines are glorified decision trees. And the brain can be simulated on Turing machines.
No it can't. It could be if we managed to module selfawareness on a Turing machine. As it stands now all the hyped up "AI" is just a bunch of very limited decision trees with no selfawareness whatsoever.
Dijkstra is AI
Dijkstra smarter than u
Dijkstra come to correct conclusion where you perambulate into circular logic and come to rong conclusion
You not fit to judge AI.
Are you denying that physics can be emulated on a Turing machine, or are you saying the brain requires supernatural phenomena to function?
>ITT Anons disagree with the definition of a phrase for absolutely no reason
replace AI with communism
>An a.i. must have beliefs, desires and intentions that impact the decisions that it makes.
you are one man on the internet. Cite this idea right now and your worthless opinion will actually have value.
Oxford English Dictionary
intelligence
noun
mass noun
1. The ability to acquire and apply knowledge and skills.
Do you see anything about a belief system in the definition of intelligence you absolute retard?
>alphazero is not an AI
ur dumb bro
t. AlphaZero aka Alphabrah
Not him but likely the latter. Bio-brains are unlikely to be a fully natural phenomenon.
your brain is a fucking markov chain programmed to eat fuck and sleep
Your brain is concave OP Ai=/= sentience
>arrange pixels in a way that resembles a human face
This is just some statistics. Here's how it works:
These are some numbers
52,63,53,78,54,92,73,64,........
1) train = guess a best distribution that fits to this pattern of numbers
2) predict = generate new numbers based on the guessed distribution
Do you think it's artificial intelligence? Or just prediction? Like regression, extrapolation?
you ever talk to a normie?? a markov chain that gets stuck in a loop is 1000000x times smarter
Even by that definition there's nothing intelligent about """"""AI""""""
I was just using Op's logic against him. In reality, the confusion behind the term artificial intelligence comes from the word intelligence's subjectivity. When in reality its just machine learning and deep learning. Machine learning based training relies on a feedback loop of information with positive/negative reinforcement influencing output. Of course, there are many more approaches to it but this one is just easier to explain. It's no different than the youtube or Instagram recommendation algorithm. But the possibilities regarding how it can be applied are endless. It can replicate conversation, driving, The human voice, how photons behave, detect cancer, perform surgery, run laundromats, recognize patterns in the stock market and alter the economy, create faces. An Ai is entirely dependant on the creativity and intent behind its designer
Google's Alpha Ai is so intelligent that every game it plays humans learn techniques from it. Alpha zero literally mastered chess in 4 hours. I game we've been playing for god knows how long. And it just played starcraft a few weeks back. What the fuck are you going on about
No. You're literally reading too much sci-fi faggot.
A.I. means Artificial Intelligence, it's not that it's it's intelligence created artificially, it's artificially intelligent. Intelligence, a program that looks intelligent but actually isn't.
If you faggots can't even wrap your heads around simple shit like that, I have no idea what you're even doing on Jow Forums.
>it's a le buzzword
It's literally a term. CPU is a marketing buzzword too, yet you don't see people sperging out about that.
Calling things shit because people wanted to put cool names onto specific things is retarded. The definition of something is still the same. But we are on a board that will argue about gif vs jif, so what do you expect.
>a program that looks intelligent but actually isn't.
So not only are people misunderstanding the word intelligence but they're misunderstanding artificial as well. Wow.
Unironically you're right. You have fantasy sci-fi to thank for that.
It actually is intelligent though. There's nothing about intelligence that restricts it to biological systems created evolutionarily, as oposed to electronic systems created by men.
Modern ML goes far beyond traditional Markov chains. You have a strong instinct to reduce all progress in the field to something you are familiar with, so you can feel smug and intelligent. But you have to resist this urge. Go out and discover what makes RNNs, LSTMs etc. different from classical Markov chains.
t.professional curve fitter
t. sour grapes
How are you any less of a repository of if/then statements, pal?
lol. I make $180k a year fitting curves. Like 90% of NN crap is only useful for voice/image recognition. Let me know if you want to figure out who to deny a loan, that's more my bag.
>An a.i. must have beliefs, desires and intentions that impact the decisions that it makes.
wait wait wait, but it's "artificial"
See there you go again. First it's Markov chains, now it's curve fitting. Modern ML is a true advancement that isn't AI but isn't the same old techniques. It's something new and better than classical methods, and if you refuse to learn about it you're only hurting yourself.
1. Not the Markov Chain user
2. Curve fitting compositions of the soft max function is still curve fitting.
3. How many training examples does it take an ANN to recognize an apple because a baby can do it in 12 examples.
The entire "field" is buzzwords, misusing terminology to make it sound more than it is. The literal only useful thing to come from it is botnet.
>Machine learning based training...
LITERALLY DATA INPUT, fags wrote a program to read the data, but muh buzzwords.
You're still on the sci-fi definition of AI though. We are talking about real world usage right now and what it's defined as.
Nobody working in "AI" is trying to create a biological replication of intelligence, but usable software that seems intelligent to be applied to do specific work.
>he things Artificial anything is on par with the real thing
wew lad
See No, machine learning is not data input, it's actually much more complicated than that. It's generating algorithms based on data, by continually brute forcing and testing, then generating results and trying to find flaws in those results, using that as a feedback and looping it until you let it.
Just because people are too dumb to understand things and marketing likes to use definitions of words wrong so they become buzzwords, does not mean terms about things are wrong just because you're too stupid to understand it or someone else who is doesn't use it correctly.
You do realize "Data input" can mean almost fucking anything right? Machine Learning based training is not a buzzword. It's referring to something very specific that happens. And it's not intended to sound like its more than it is. It is WAY MORE COMPLICATED than it sounds. And the fuck do you mean the only useful thing to come out of computer science is botnet. What about the motherfucking internet and everything in it. None of it would be possible without computer science and engineering. And you have the audacity to put quotation marks over it when you call Computer Science a field. Get a degree or shut the fuck up.
>"Do you like it a lot?"
wow. you stupid gorilla nigger.
tensorflow and sklearn are alternatives, you would only need to import one or the other. But you already knew that since you're a smart guy who sees through all the bullshit and fads, right?
Greatest researcher known to man
"What is ensemble learning?"
The word can not be saved, it's gone to shitter. I am trying to replace it by saying artificial mind instead.
>t. Sammad Al-Hayadi
What you're describing is AGI you retard
baby's input data is gigantic, babys brain spent months forming in the womb, baby spent years learning language and experiencing the world. Baby learned apple very quickly bc of transfer learning from other training on much more massive dataset. Baby algorithm actually very inefficient
>I make up my own random beliefs and physically assault anyone who doesn't share them
OP confirmed SJW