Potemkin wrote:Stuff gets called “AI” which actually isn’t AI
This is where I was trying to go with @ness31, by asking for details. I'm sure most of what he was alluding to that is "AI", actually isn't. Social media all by itself isn't AI (though they do deploy some AI in certain areas). The problems of social media are also created by people, not machines. (though the algorithms don't help and in fact encourage, but it's still humans torturing themselves on social media, those algorithms are a problem, but many of them cannot be called AI) The issue of social media is well documented but besides the point.
Anyway, to properly discuss AI and its impact and implications, you have to get down to very specific terms, use cases, applications, etc.. The very broad and uninformed strokes by which ness31 is throwing around his claims, simply isn't good enough. His points are always very nebulous, largely due to his lack of understanding. This is is MO (See Russia threads), makes a claim, then backs away when asked to expand on the claim.
Lots of companies will also bullshit as well. They will claim something is "AI", when it's really just a moving average filter.
Seriously, I've seen that.
To be clear again, I do think AI is going to be the end of us. However, in it's current form, it is still rather crude.
Potemkin wrote:Very hard, actually. Adaptive learning is not easy, and no deterministic coded program can do it effectively. Neural nets are much better at it, but they’re still dumb as a rock compared to the average human.
Exactly. AI is powerful, but fundamentally not really that advanced. In fact, machine learning adn inference is a VERY brute force process. What enables it today, is simply that CPUs finally have enough compute power to enable these applications.
One could make the argument that even what we call as truly AI (Machine learning and inference engines), actually isn't. I don't go that far, but you could make that case if you wanted.
I can think of 11780 reasons Trump shouldn't be president ever again.