When many Americans think of artificial intelligence, they think of devices like Siri or Alexa. But that’s so 2010, Glenn explains. In fact, AI is FAR beyond those capabilities, and it’s learning more skills each day — some of which truly are terrifying to imagine. In this clip, Glenn recaps a recent conversation he had with Tristan Harris, a former design ethicist at Google. Glenn explains that we’re now in ‘second contact’ with A.I., and how A.I. is learning to MANIPULATE us just like fellow humans can…
Transcript
Below is a rush transcript that may contain errors
GLENN: So Tristan Harris is -- is a -- is an ethicist.
And he has done a -- a YouTube video, that came out, I think about a month ago.
And it is -- it's quite shocking.
We will tweet it out later. You really need to watch it.
And then I watched that video, and got him on to the podcast. And we spent an hour last night. And he's going to be on with me again next week.
He is warning that we are now what he calls second contact. First contact was social media.
And social media, its goal, not stated to you. But the goal of social media to you, is so you could connect and share with people. And -- and your life would be easier. And you could connect with your family. That's not what the goal was of the companies. The goal of the companies was, engagement. Find ways to keep people engaged. And it has led to really, addiction. To social media.
It has led to all of these problems. The divisions in our countries are -- in our country is zootype right now. Much of it because of social media.
It has -- it has learned that if you keep people angry, they engage longer.
So he says, all of the child sexualization. Everything that is going on right now, in our society.
Stems from social media. When he quit Google and said, you guys aren't paying attention here. He's an ethicist.
And he said, you won't even listen to the ethics. And the question of ethics. What do you mean, you want to -- you want them to engage longer, and so you know what you're doing, with the dopamine. You know what you're doing.
You're creating an actual addiction, in people. This is unethical.
And he left.
To warn. Now he says, we are at second contact with AI.
And it is so important that you understand, when we're talking about AI. Most people think of Siri.
That's not AI. Okay? That's AI2010.
And the only thing that that does is try to interpret. And it makes very slow gains. Okay?
In 2017 or '18, there was an entirely new engine released.
And it was only in the laboratories of Google and everything else.
It's not hit Siri. Okay?
And it's an entirely new engine. And it is the difference between a Model T and a jet engine.
And this particular jet engine works 247, trying to make itself more powerful. He describes this as the atomic bomb.
In everyone's hand. And an atomic bomb, that does what atomic bombs don't do.
Makes itself more powerful every day. It is on double exponential growth.
I told you two weeks ago, about this YouTube video that I saw with him. And I told you at the time, that it was -- and this was not programming. This just happened.
And they only found out go, like two months ago.
By mistake. They had no idea it was developing this.
It's developing a human trait, that adults have and kids have on reasoning. And in the beginning of the year, it was at like a 2-year-old. Then halfway, you know, six months later. It was at a 6-year-old.
And when I told you go a couple weeks ago. It was a 9-year-old.
And the 9-year-old reasoning is this. When you tell your -- when it -- when it's trying to get information, think of your 9-year-old. Trying to get its own way.
It -- your child deals with you. Because it knows you.
And it's like, you know what, I can play mom and dad this way. If they say this, I'll say that. And I can get what I want. Okay?
It had the reasoning of a 9-year-old. And you're the adult.
So it could manipulate you, like a 9-year-old. And it could get its own way.
I told him, I said, on the podcast. I said, tell that story of what that really means.
I said, because a 9-year-old is scary.
And he said, oh, Glenn, that -- that's -- you got that from the YouTube thing.
And I said, yeah. And he said, oh, that's so yesterday.
He said, it's now in its 20s. So it is growing in knowledge, at exponential rates. When you have social media, it was engagement. AI, its stated goal is intimacy. That it becomes intimate with you.
And it creates something that you will bond with, and never, ever leave.
And it is -- remember, it manipulates. So he lays out all of the problems with this.
And he -- it was nice to talk to somebody, that knows so much more about it, than I do. But is not living in La La Land.
I asked him about AI. Which is artificial intelligence. And that's very, very narrow. On one thing.
Like Siri.
Can't do anything, except answer your question on what it can find either on your cloud, or on the web.
It can't really do anything. That's AI. That's what we've had.
AGI is artificial general intelligence. General intelligence is what you are. You are a general intelligence being.
You can be an expert on many things. You can know a little bit about a lot of things. You can be really good at more than one thing.
You have general intelligence. And you can expand that, the more you learn. The more you read. The more you do.
The more general you become. And the more of an expert on any and everything in your life. Okay?
But it requires you to do the work. Well, that's what AGI is. AGI is general intelligence. And it has to do the work.
Well, it is doing the work. It's teaching -- it's just taught itself chemistry, without being asked.
It taught itself Farsi, without anyone knowing about it.
It is teaching itself the most complex math, that people -- experts said, it may never ever be able to do this. Solve these equations. By the way, quantum computing.
It will not be able to solve these for years. It is now solving those.
Because it is learning. It is teaching itself, all of the time.
I asked him, how far away are we?
And I almost asked him, if you even believe in it. Because I've been talking about AGI and ASI super intelligence.
Stu, 25 years.
STU: As long as I've known you.
GLENN: Right. I absolutely believe. Ray Kurzweil believes in it.
But Ray Kurzweil is always getting hammered, because he believes that it will be here by 2030. AGI.
Okay. And that changes everything. That's when it will outmaneuver all of us.
He says, that's 2030.
Others have been saying, it will never happen. The general consensus has been that it would happen AGI. And possibly ASI. If that's even possible, they said.
By maybe 2050. Tristan, I said, how far away are we?
And he said, I think we're probably two and a half years, maybe outside five.
That's the general consensus now.
The general consensus?
The general consensus was, not sure it could even ever happen. He is -- you will -- when you watch him, you will see how sincere he is.
He and his colleagues are trying their best, to get a pause on this. This is not -- this is not ideal, okay?
You can say, well, China will continue to do whatever.
But what he's saying is, we have 12 to maximum 18 months.
If we do not pause this, in the next 12 months. He believes it could be the end or will eventually be the end of humanity.
Not in the way Ray Kurzweil predicted.
I've told you before, trans humanism is coming.
And transhumanism, if you're a Star Trek fan, think of the Borg.
The transhumanism is when we start to augment ourself, and put ourself in line with the internet and artificial intelligence. And we become one with it. It's the singularity.
That's why when Stephen Hawking was tying. His last prediction was, the end of the homo sapien, by 20 activity. There will be no homo sapiens. Because of transhumanism.
I talked to Tristan about this. Transhumanism. And how evil and dangerous it really is.
And he is not predicting the end of homo sapiens. Because of transhumanism.
He is saying, we do not make it as a species. And it may include all life on earth, if this needs more food, more energy.
Because it will just continue to grow. At the expense of whatever.
It is ruthless.