Experts predict how AI could wipe us out.

 

 

Hasn't every civilization in the universe destroyed itself when knowledge reached a certain point where that becomes possible in so many different ways? No conceivable threat could have ended civilization in the so called dark middle ages. There was almost zero chance of that happening then.

They can't reproduce on their own.  We don't need an "army" of robots.

And thus the singularity Institute was born to help humanity edge towards safe and "friendly" AI.

Efforts undermined by the current "Full speed ahead and word not allowed the torpedo's!" greed-race to be the first to cash in on general AI, with fewer and fewer concerns about safety.

I can envision a time in the future after the singularity where machines have taken over and are the dominant entities on this planet.  They can do many things, but they still have to keep humans around to consult with them for their creativity, much like we consult with AI for its data processing capabilities now. 

They might even argue amongst themselves about whether human creativity is a good thing or a bad thing.  After all, human creativity might come up with some way to totally take over from the AI,  something that AI could never anticipate, making human creativity an existential crisis for the machines.

Hook

Life imitates art...

 

Future AI.

15 hours ago, Fielder said:

Hasn't every civilization in the universe destroyed itself when knowledge reached a certain point where that becomes possible in so many different ways? No conceivable threat could have ended civilization in the so called dark middle ages. There was almost zero chance of that happening then.

 

Not sure about every, but there are certainly a number of "great filters" to overcome. Perhaps explains the Fermi Paradox?

 

"No conceivable threat could have ended civilization in the so called dark middle ages."

 

Unless it did in the past, and civilization started from scratch. Natural phenomena etc. 

Edited April 28, 20251 yr by martin-w

11 hours ago, HiFlyer said:

with fewer and fewer concerns about safety.

 

We need those three laws of robotics that Asimov conjured up. Maybe more than three. 

I had a convo with Meta AI the other day, about my favorite subject, muscular hypertrophy, and it astonished me how it held a conversation with me. 

4 hours ago, LHookins said:

but they still have to keep humans around to consult with them for their creativity

 

Personally, I see no reason why they can't be creative. Once the hardware reaches an equivalent to the required biological density and the software is advanced enough, creativity will be equal to ours, perhaps better, and generated in an instant. No need to contemplate for hours, we are talking microseconds. I don't believe creativity is only possible with wet squidgy stuff. 

Martin, you are wonderful! You have illustrated my point better than I could have myself.

You read the words and understood them perfectly, but missed the underlying meaning.

This was, for all practical purposes, a very short story about how humans and AI swap places. We're afraid of AI because it can out-think us. AI is afraid of us because we can out-create it. 

I once wrote an essay about creating art from random numbers. In this case, "art" didn't mean music or pictures, it meant getting a game to react in an appropriately realistic manner.  I found that you can do 1000 tests and decide you've found a solution, but the second time you run the same code in a live situation it generates garbage. I came to the conclusion that the random number generator is perverse and hates you.

A human can generate appropriate "random numbers."  A machine cannot. This goes beyond how many neural connections there are. AI can only remix what has been done before. It can come up with new and novel combinations. But it can't create anything truly new. This code might eventually be cracked, but I'm not betting on it.

Hook

17 hours ago, spilok said:

They can't reproduce on their own. 

Yet!

Noel

Sorry, but I put more credence in "Actual Intelligence" as opposed to "Artificial Intelligence".

If you can't think for yourself... that's on you.

Not to worry, until the Storm Troopers' aim improves.

14 minutes ago, Fielder said:

Not to worry, until the Storm Troopers' aim improves.

Hey, the Storm Troopers had perfect aim! Their tactics were to use suppressive fire to force their opponents to keep their heads down (and not fire back) until the Troopers could close to arresting distance. Then their opponents could stand trial.

😄 

Hook

9 hours ago, martin-w said:

"No conceivable threat could have ended civilization in the so called dark middle ages."

I'm under the impression the Black Death of the fourteenth century wiped out something approaching a third of the population of Europe and the near East?