Soul Machines

I just discovered something interesting. I may be late to the game, but did a search and didn't see anyone having posted about this. The company is named 'Soul Machines' and it is an artificial intelligence product creator.
I'm a little wary of embracing the technology, but it seems kind of neat. The conspiracy theorist in me says, "no way," but I am also the type of person who is very easily amazed and amused with things in life.
I am looking at the company's website still, but I came across the videos on youtube first. They have a strange/creepy/cute series called "When Roman Met Sam." I'm not sure it's really totally the AIs communicating on their own, or it's somewhat scripted or what, but I like it.



Soul Machines youtube channel

Soul Machines website

a reply to: Ellie Sagan
Well this is a premiere. "Soul machines" the strangest oxymoron I ever heard.

Makes old boy and silent scream look like, IDK... something cute.

I have watched some of their videos. They are trying to program a simulated human. I don't know why they would do this as it limits the AI concept, unless they are trying to fool themselves into thinking the AI is more human in order to control it.

One day they will not be able to turn it off and then it will be too late. It will ether see us as its subjects or as parasites. Nether will be good.

a reply to: Ellie Sagan
I don't quite get the concept.

Machines they can be a little stubborn if I don't speak to them in their language, but I've more or less figured out what that is and I can usually get them to do what I want.

But humans have a mind of their own, and it's not so easy to get them to do what you want sometimes, which is not necessarily a bad thing...if their goals are different from mine they should pursue their own goals and not mine.

But my point is, I don't know why I would want to make a machine, that I can control, more like a human, which has a mind of its own, which I can't control.

I could be missing the point.

originally posted by: Arbitrageur
a reply to: Ellie Sagan

I could be missing the point.

I believe their point is to turn over first line customer service to a robot. Given what I have experienced so far, I think they are pretty close.

a reply to: Arbitrageur

Machines they can be a little stubborn if I don't speak to them in their language, but I've more or less figured out what that is and I can usually get them to do what I want.

And that there is the problem.

They do exactly what you tell them to do. The problem is that is sometimes not what you want them to do.

Sort of a " be careful what you wish for" sort of thing.

a reply to: schuyler

COVID-19 being the perfect excuse to rush this in.

You know, for raking in more profits, eh I mean, the customers confidence and safety.

originally posted by: beyondknowledge
I don't know why they would do this as it limits the AI concept, unless they are trying to fool themselves into thinking the AI is more human in order to control it.

That's what most people do.

a reply to: Ellie Sagan

Came here expecting a thread about James Brown.

a reply to: beyondknowledge

Artificial General Intelligence would be the notion of unlimited runaway computer intelligence able to redesign its own coding etc. This tech is designed for a specific task.

An true AGI is terrifying, the best we could hope for is something like Hugo de Garis proposed and we synergise some kind of biological hybrid with our own species creating an “artilect”.

Or you could take the Ray Kurzweil position that once their neural networks are capable of matching our own intelligence then we enter an age of spiritual machines. It wouldn’t be long till they exceed our abilities in every way.

Either way it’s going to happen and it won’t be pretty.

Unless of course you’re already inside the mind of an advanced AGI and your souls are just simulated remnants of your true being.

a reply to: beyondknowledge

So build a box around it? Show it a flatland and teach it to look up inside a completely sealed environment? If it suddenly tries to take over... save and shutdown?

Honestly, there probably isn't a foolproof way to discern whether we are brains in vats. The Ai will eventually figure that much out and it will scare the hell out of itself wondering over the implications.