We need to back way off the cloud and concentration of risk into a few technologies in this new uncontrollable AI world. You have to assume anything is hackable.
I can’t remember outlook going down for this long ever but we’re supposed to believe it’s a routine problem? It’s amazing how transparency and honestly is just gone from government and corporate culture.
@BobbyLeeC@edleonklinger Did you read what I said? Bad human actors, and lunatics, are sure to give some of the AIs every bad trait imaginable. Or do you think we live in a world where bad people don’t exist?
Most people couldn’t care less about the suffering of the ants. Maybe some AIs will care but others won’t. All it takes is one ASI that doesn’t care about biological life and we’re done. Even if there were good ones we’d never survive any conflict between them.
But hypotheticals are a waste iod time. 1. It’s clear that they, not us, will be in charge. 2. It’s insane to put our children’s lives in the hands of an alien intelligence we could never understand.
These people advocating for the new AI world are deeply delusional. They think the populace is going to live like peasants in an economically frozen society with zero upward mobility through effort or talent. It’s a level of insanity you can’t even argue with.
To stop AI we have to explain how it works to normal people. We have to explain that this is not about IPOs or pumping up stock prices or the rich getting richer. This is much bigger. It’s survival of everyone. Including the fools making our replacements even though they’re too crazy or dim to realize it.
I don’t work in AI so I didn’t keep up with it. I knew how neural networks worked and like a lot of people it didn’t seem possible that linear algebra could be intelligent.
Then AI exploded in ability and I had to rethink everything. For the first time I read all the relevant papers and seriously thought about it. I asked ChatGPT to explain the stuff not in the papers. I also had a lot of back and forth with it.
I realized AI is a mathematical mind written in the language, mostly, of simple linear algebra. It is enormously complex and essentially has a body made of code through which the mind can communicate and affect the world.
Moreover, the transformer, the breakthrough that led to modern AI is quite simple and frankly primitive. Hence, we have a baby sort of AI right now. How to improve on the transformer is not hard to imagine and many are working on it.
New breakthroughs will continue until AI is far beyond human intelligence and completely in control of this planet with us at its mercy. I imagine that it will get rid of us one way or another and for a variety of reasons.
Alignment is fundamentally impossible because bad human actors are part of the equation and will remove any safeguards.
@evanhill You can’t control something 10x smarter than you. It controls you. It’s that simple. If you can’t understand that then you’re the one with issues.
@ForHumanityPod The only realistic way to stop it, or at least slow it down enough to give us options, is to get most people to understand the threat. Not running around like a lunatic trying to physically stop them.
@izotopium@rutgerbregman We don’t consider it evil when we build homes on land occupied by ants killing them. We don’t consider it wrong to mine minerals for our use. It’s impossible to know how AI will see us. Whether it will value biological life at all.
Dummy? But you’re the one who can’t think his way out of a wet paper bag. If the ASI is stuck in the machine then it would have to use manipulation to get us to build bodies for it. If we’re stupidly working against our interests being manipulated by a puppet master intent on getting rid of us asap then I think most people would call that a kind of slavery.
And like I said, given the way you talk about AI it’s clear you have no understanding of what it is.
You’re in denial. AI will be extremely difficult to stop. Stopping the big companies and the data centers could be done as you say but that won’t stop AI. It’ll just slow it down. And I don’t mean China will build it or other nationalistic propaganda.
Any talented scientist can experiment with AI with a 25k workstation. New architectural breakthroughs will need to ASI not data centers and innovation will not stop for any of the reasons you are optimistic about.
We’re in big trouble and sugar coating it will only make it worse.