View Single Post
Old 12-28-2017, 10:39 AM
SamuelA is offline
Join Date: Feb 2017
Posts: 3,688
Originally Posted by LSLGuy View Post
As you almost said, the Universe is an existence proof for a way to make a universe. And biology is an existence proof for nanotech. But to jump from there to a) humans will master that tech; and further that b) we, or one of our AI tools, will do it soon is hubris. Nothing but hubris.
You realize that what you are talking about there are terms where some of them nearly infinitely times harder than others.

I don't know how the universe was made and it might require things that don't exist inside the universe itself or can be manipulated in any way. I did mention it was far fetched.

Believe it or not, most of my ideas do not require nanotech. Self replicating factories don't need it. And the AI I've talked about in my posts in the last 6 months isn't the sci fi concept of AI. We're not talking about machines that can even carry on a conversation, they are just big mathematical algorithms that explore a solution space and choose the min( possible paths ). They use other algorithms, also capable of automatic adjustment, to actually model the world and classify what is in it.

Such algorithms could control a robot to efficiently perform any repetitive factory task. But they could also scale to controlling a robot that can actually build a rube goldberg machine to solve a defined task, and later to automated designs that are as good as human engineers.

Oceans of money are flowing in this direction. The difference between now and "then" (none of these algorithms are new) is that quantity matters. It's a different world when billions of dollars and tens of thousands of people are working on AI vs millions of dollars and dozens of people in elite computer science labs are working on AI.

It may take decades to go from "we know the tech can do it, it does scale far enough" (present state) to "basically all factories, mines, farms, retail stores, warehouses, and cars and trucks can be automated, with commercial off the shelf products existing for all tasks".

And as I mention, once you hit that point, you could dip into nanotech or mass biological research and hit the problem with a million times current capabilities. Instead of a few greying PhDs like your favorite blogger, working semi-independently, you'd study the problem on a colossal scale using a lot of new tools.

So it's not going to take 5000 years. It can't. This would be like you saying that a flight around the world is going to always take 2 years in the air. Distance and time don't add up and your estimate is so far off it's silly. With that said, no I can't say exactly when, and I do not know what the social outcomes will be in a world where every job can be automated. The current system of capitalism, working exactly as it has worked for centuries, would reward only those who own robots and land and give not a penny to any of the displaced human workers. Those same rich owners can then bribe the government, since that's allowed in the United States in practice, to make sure only the most meager and grudging forms of social support are given to all those lazy "takers" who can't seem to find a job because all their skills are tasks a machine can do better.

Last edited by SamuelA; 12-28-2017 at 10:43 AM.