larryepage wrote:
The problem, as I see it, is that this technology is derivative. It only works by assembling bits and pieces of what already exists. It is not truly creative.
Our society right now has a very low threshold of what constitutes creative or innovative work. We can readily see that all around us, including in schools and on this site. Against this lowered threshold, the alleged capabilities of AI seem impressive. Measured against more meaningful standards, they pale pretty quickly.
I read an article that suggested that AI is within a few years of passing a singularity...becoming more capable than the human intelligence that created it. I believe (and hope) that is just idiotic nonsense. Of course, my belief does not guarantee that we will not just refuse to pull the plug and thus cede our existence to "the machine."
The problem, as I see it, is that this technology... (
show quote)
Answer to first sentence:
Wrong
Imaging AI's can and do produce absolutely new creative images which were'n produced by any human being before it.
The do so by studying millions of tagged images, that means photos, oil paintings, whatever that a person tagged with "boat", "plane", "dog", whatever.
It is *exactly* how an artist learns... o a child.
Btw: "derivative technology" is a legal term, and therefore a) does not need to have any relationship with the physical world and b) there are enough AI patents to support the notion that AI is not "derivative technology"
/**************************************************/
Answer to second sentence:
Wrong again.
There is no known limit on how complex or "good" work (by whatever standard) an AI can produce.
If there is a limitation, it is time bound. That means it will be lifted shortly. Just remembers Moore's Law.
You seem to be making uniformed statements, you can start learning about AI here:
https://arxiv.org/pdf/1802.05800.pdfhttps://arxiv.org/pdf/2207.02696.pdf/**************************************************/
Answer to third sentence:
"Mein guter Freund und Kupferstecher" as my grandpa used to say each time anybody spoke something really naïve...
Believe me: we will no pull the plug. Period
Because Moore's Law and our own very nature, AI will advance boundless. At first by us, and then by itself. Did you know that AI is used to design the lastes ... ahh, AI chips?
On the other hand, genetic engineering is advancing at breakneck speeds... some say that in 30 to 50 years the human genome will be known as well as any other programming language. An we (or the AI) will be able to make any kind of modifications or creations from scratch.
At some point both lineages will met: carbon humans and silicone AI. And then and nobody will be able to tell which is which.
I have no proof, but I'm pretty sure that this happened millions of times before in the observable Universe, and that this is the real solution to Fermi's Paradox.