You are mistaking a neural network with a general intelligence.
If you are not aware, "AI" other than neural networks does not exist. In the past, there was huge efforts applied to computer science and programmer startups to shut down anything connected with expert systems, self-categorisation, inventing algorithms, and other stuff that is different from neural networks and magnitudes of order closer to intelligence than stupid neural networks.
Fucking chatbot is a dumb neural network that trained on internet content that just show you compilation of most probable phrases it was trained on suitable for your request. Nothing more.
And ive met bottom workers that arent as smart as the chatbots.
Unlike chatbots they could deal somehow with something they never met.
The language parsing/analysis of the chatbots is actually more impressive than the AI. Even that can be fooled but the "intelligence" aspect is just a glorified web crawler with a couple extra features
The language parsing/analysis of the chatbots is actually more impressive than the AI
It is not impressive at all. Stupid very old IRC chatbot Elisa from early 90s written in perl does not differ a lot. And it was not even neural network, just a small script with very simple algorithms who could, unlike chatbots, "self-educate" during conversation and hold a sensible dialogue with user. And despite it was heavily English language dependent, I was able to get sensible dialogues in Russian too. Kind of, because order of words in Russian sentence does not really matter, so that dim the picture a lot.
There is nothing impressive when you feed terabytes of questions and anwsers pairs from real internet conversations, and see something similar as a result. That is exactly now neural networks work. No any AI magic at all. Chatbot does not understand you. It does not even learn from you. It just calculate most probable reply from known replies for the question that is closest to yours.
I disagree, if you ask it classification questions it can figure things out based on pulling other definitions. The results are again that of a glorified search engine but it is able to meaningfully parse the question and the datasets in order to draw a conclusion. The conclusion drawn is very surface level because it's not actually intelligent but the language parsing may be better than you're giving it credit for
I am aware, you, however, are expecting neural networks to be fantabulous replacements to workers just because they use the term "AI".
Some dogs are very intelligent - but that doesnt mean i think they are human. I also dont go around and say that humans are the only type of intelligence or that the bird intelligence that humans have created aren't ready to replace humans.
And you should meet the bottoms before you overestimate their intelligence, again. They are already using them to replace workers at fast food places. A chat bot will take your order at the drive through. Where's that bottom worker going to go?
Oh, right. As I said, they'll be useful idiots for the great reset when they get promised their desires to tear down the world around themselves.
are expecting neural networks to be fantabulous replacements to workers just because they use the term "AI".
Everything that is named by them "AI" is ANNs. There no any single other thing that could be named "AI". So, always, when you see they use term "AI", in is 100% ANN.
All other research on something that could be eventually resemble any intelligence is heavily curtailed long ago. They have specific reasons for that, but it is another rabbit hole.
A chat bot will take your order at the drive through.
Until something unexpected happen. Oversimplyfing that problem - ANN have no any means to understand that normal customer will never order 1000 hamburgers. So, owner of such drive through will have to hire somebody with apropriate knowledge of matter to look after "AI". There still be same workers with same skills to be competent enough to judge "AI" "decisions", just doing another job.
We had fully automatic cafes in USSR times. Vending machines with cakes, donut cooking machines with vending frontend serving hot donuts, coffe/tea machines, machines that wash crockery and so on. Everything worked fine witout any "AI". There was absolutely no any need for any "AI" for that task at all. But there still was same amount of workers to load machines, to look for them, to service them, to resolve customer problems and so on. And no "AI" could resolve that and replace that workers even now. IDK, it is insane how quickly people forgot something that goes out of their range of view. Automatic cafes is a distant past. They are not something new, and they do not free workforce. Moreover, automatic cafes demand more advanced tech and more skilled personnel than regular ones. In USSR it was not a problem, but in the modern world it is. Personnel, apart knowing regular cafe workflow, had to be familiar with advanced tech and automatisation and have an ability to make more complex decisions.
That nowdays "AI" shit have nothing to do with what you talk about. It is purely a way to hide behind non-existing "AI". They train ANN to produce results they want, and then "You see, it was not us, it was AI decision! We have nothing to do with that! AI knows better!". Simple as that.
If you are not aware, "AI" other than neural networks does not exist. In the past, there was huge efforts applied to computer science and programmer startups to shut down anything connected with expert systems, self-categorisation, inventing algorithms, and other stuff that is different from neural networks and magnitudes of order closer to intelligence than stupid neural networks.
Fucking chatbot is a dumb neural network that trained on internet content that just show you compilation of most probable phrases it was trained on suitable for your request. Nothing more.
Unlike chatbots they could deal somehow with something they never met.
The language parsing/analysis of the chatbots is actually more impressive than the AI. Even that can be fooled but the "intelligence" aspect is just a glorified web crawler with a couple extra features
It is not impressive at all. Stupid very old IRC chatbot Elisa from early 90s written in perl does not differ a lot. And it was not even neural network, just a small script with very simple algorithms who could, unlike chatbots, "self-educate" during conversation and hold a sensible dialogue with user. And despite it was heavily English language dependent, I was able to get sensible dialogues in Russian too. Kind of, because order of words in Russian sentence does not really matter, so that dim the picture a lot.
There is nothing impressive when you feed terabytes of questions and anwsers pairs from real internet conversations, and see something similar as a result. That is exactly now neural networks work. No any AI magic at all. Chatbot does not understand you. It does not even learn from you. It just calculate most probable reply from known replies for the question that is closest to yours.
I disagree, if you ask it classification questions it can figure things out based on pulling other definitions. The results are again that of a glorified search engine but it is able to meaningfully parse the question and the datasets in order to draw a conclusion. The conclusion drawn is very surface level because it's not actually intelligent but the language parsing may be better than you're giving it credit for
Just try to talk with this toy using non-existing words.
I am aware, you, however, are expecting neural networks to be fantabulous replacements to workers just because they use the term "AI".
Some dogs are very intelligent - but that doesnt mean i think they are human. I also dont go around and say that humans are the only type of intelligence or that the bird intelligence that humans have created aren't ready to replace humans.
And you should meet the bottoms before you overestimate their intelligence, again. They are already using them to replace workers at fast food places. A chat bot will take your order at the drive through. Where's that bottom worker going to go?
Oh, right. As I said, they'll be useful idiots for the great reset when they get promised their desires to tear down the world around themselves.
Everything that is named by them "AI" is ANNs. There no any single other thing that could be named "AI". So, always, when you see they use term "AI", in is 100% ANN.
All other research on something that could be eventually resemble any intelligence is heavily curtailed long ago. They have specific reasons for that, but it is another rabbit hole.
Until something unexpected happen. Oversimplyfing that problem - ANN have no any means to understand that normal customer will never order 1000 hamburgers. So, owner of such drive through will have to hire somebody with apropriate knowledge of matter to look after "AI". There still be same workers with same skills to be competent enough to judge "AI" "decisions", just doing another job.
We had fully automatic cafes in USSR times. Vending machines with cakes, donut cooking machines with vending frontend serving hot donuts, coffe/tea machines, machines that wash crockery and so on. Everything worked fine witout any "AI". There was absolutely no any need for any "AI" for that task at all. But there still was same amount of workers to load machines, to look for them, to service them, to resolve customer problems and so on. And no "AI" could resolve that and replace that workers even now. IDK, it is insane how quickly people forgot something that goes out of their range of view. Automatic cafes is a distant past. They are not something new, and they do not free workforce. Moreover, automatic cafes demand more advanced tech and more skilled personnel than regular ones. In USSR it was not a problem, but in the modern world it is. Personnel, apart knowing regular cafe workflow, had to be familiar with advanced tech and automatisation and have an ability to make more complex decisions.
That nowdays "AI" shit have nothing to do with what you talk about. It is purely a way to hide behind non-existing "AI". They train ANN to produce results they want, and then "You see, it was not us, it was AI decision! We have nothing to do with that! AI knows better!". Simple as that.