Do you think AI is going too far? It seems like AI is becoming more popular and advanced. You should see the things AI can do now: write poems https://youtu.be/QAnTG9u_7VQ https://youtube.com/shorts/bdsu3Q40wv8?feature=share, make "art", write essays, etc. (you may have heard of these things) It's also more advanced than before.
You're viewing a single comment thread. View all comments, or full comment thread.
Comments (31)
sorted by:
Again and again. It does not matter how you call it. What is matter is that thing unable to do anything that could be accounted as an intelligence. It does not think, it does not make decisions, it does not make conclusions it does not understand anything, it have no any properties that is thoroughly attributed to it by MSM.
ANNs is pretty limited things, they could be fine for some tasks, but they need enormous volumes of data for decent training and will always fail when input will be outside model. It could perfectly distinguish between cats and dogs when trained, but will miserably fail with racoon. And you will not be able to back-propagate racoon option with one or even dozen of images, you will need thousands if not millions (depending on number of layers and resolution) racoon images to make it work with racoons. It is simply not scalable. Same with GPT models and any other ANN stuff. It is narrow, it is not scalable, it is insanely hungry for resources at training stage. Niche technology that inflated to the enormous scale, to the level when what public believe today about it is complete lie. Just like quantum computers, meanwhile.
The only real reason for the AIhoax we observe today is an attempt of TPTB and their puppets to move any responsibility to the so called AI and hide behind it. "It is not us, it was AI decision". That's all. And the success of that farce is completely depends on total acceptance of fake AI existence. They don't care what you will think about not-an-AI, good or bad, threat or blessing, they only care about you believing that their not-an-AI could make decisions. And looks like they succed in that AIhoax much more than they succeed in coronahoax or vaccinehoax.
If you want a real AI area question to dig around - then try to find what happened with so-called "expert systems", and why they was abandoned/banned, and why only that silly ANNs get a hype. You will find pretty interesting things from the fear of new inventions, including closing ones, that expert systems should be able to produce (one of the main things they was researched and developed for) to the using need for ANN training as a reason for getting access to so-called BigData (or just "surveillance", speaking human language).
So based on your third paragraph we have the same read on the situation. Good enough for me
Nice to read that.
How do you think, is is possible to somehow revert that blind believe in AI existence? Seriously, it was easier with coronahoax/vaccinehoax to push sane person to began to ask questions.
At the current pace humanity will surrender its governance to non-existing AI and free politicians and all kinds of bureaucrats from any, even potential responsibility very soon.
Honestly I don't know. Because even people who I know at least in theory understand how these neural nets work will ask chatGPT questions as though it is some kind of oracle. But a large portion of the population wants it that way, they want to be directed and to avoid thinking and making decisions. In that way I think people need to learn to unplug and to return to tradition but I don't see that happening without everything collapsing first.
What are your thoughts?
I think we will see more and more "AI" on different support lines, bank customer services and so on. It is already there, really. Then, may be, at some point, people will be annoyed a lot by necessity to talk with a robot instead of real person, and began to boycott such companies like it was with woke ones. If that happens, may be we will have a chance.
Otherwise, "AI" will be declared time proven technology and some laws will be passed allowing "AI" for use in some specific circumstances, say, something about non-critical decisions about lost packages in state post service, few at the beginning and much more in later additions.
I think point of no return will be when some laws about AI will be passed as if AI is something real.