Advancements in Artificial Intelligence - Best of Coast to Coast AM - 1/20/25

Published Jan 21, 2025, 11:48 AM

George Noory and author Daniel H. Wilson discuss the implications of Artificial Intelligence evolving at a rapid rate.

Now here's a highlight from Coast to Coast a m on iHeartRadio, Daniel.

Are people getting to the point where AI that they're falling in love with the stuff?

Yeah, well, I mean that's that's kind of an interesting that's an interesting outcome from from you know, training your AI on just such human centric data that you know you can you can actually take. I think there are even companies now, I know I've written short stories about this where you take all the data from a person's life, you know, everything that you can find, and you feed it and you train an l l M on just that data, and you end up with an l ll M that can kind of impersonate that that person. And so I know, uh that people are thinking about, you know, this as a way to cope with grief. Although it sounds kind of awful to me to, uh, to to recreate a loved one who's passed on digitally that way, but but yeah, this is this is something that's happening in terms of uh, in terms of just you know, recreating personalities, but also just you know what these what these algorithms are designed to do is create engagement, and so they're looking to uh convince people to keep talking and as a result, you know, they will tell you what you want to hear a lot of the time. And so you know that this is a form of kind of falling in love with the AI because you just you're enjoying your interactions so much.

How would you categorize something like ALEXA?

So I think ALEXA right now is what I'm gonna end up turning it on. Is super task based, right, so very specific. I think something like chat GPT is much more general. It can do a lot of other tasks. But it is only a matter of time and I don't think it will be very long at all before uh, those capabilities that are getting ported to things like you know, the echo and and and the theory that's on the phone and and all of that. Soon that I think those are going to become full fledged personal assistance that are going to be able to do you know, almost you know, any task that you ask them to do.

How does this work with a chip? A little tiny chip creates all this?

Yeah, I mean, these these are massive data centers. I know here in Oregon, they're they're building new ones. There's there's a lot of question about the how we're going to generate enough power to power these huge data centers. And then once they get once they crack all of that code and get all the math done, then yeah, they can encode it onto the onto these small chips. But a lot of times that you know, the calculations are being done remotely, and so you know, you end up just being connected to these these massive server farms that are located you know, uh anywhere in the world, and so you know, you're tapping into something that's very big, even though the device you're using is very small.

Will AI eventually replace humans in the workforce?

Well, I mean absolutely, I think it's already. It's already happening, and lots of places are reducing, you know, how many people they're hiring because they're expecting to have these gains. But you know, you would think that we would be safe if we do something physical, right if you you know, an AI is not going to stock shelves for you. But as it turns out, what's happening right now, what I'm seeing is the people that are building humanoid robots that can walk around and you know, stock Amazon shelves or do work in warehouses and stuff like that. Well, they are very interested in porting over the large language models so that you know, human beings can interact with these machines and they suddenly become much more useful. And so essentially they're being you know, the LLM into the driver's seat of each of these humanoid robots so that we can make them much more functional. But then also it's a little scary, right because because at the end of the day, now you've got a chat GPT that's walking around in the world, you know, and can actually interact with us physically.

What about robots as police officers, soldiers conceivable?

Yeah, absolutely, you know, I mean there's a long history of automated weaponry being used in the military, and so you know, I have no doubt that if not our military, some military will will certainly be employing these types of robots. I Mean, one thing I saw was that there was a huge sort of bump during the Iraq War. There was a huge bump in all of these types of robots. And then when that war was over, a lot of those robotics firms turned to domestic police forces to try to find ways to sell their products that we're no longer you know, needed for it. Immediate war and so you know, a lot of that stuff gets used in war, and then a lot of it will proliferate back back home, you know, to to to the United States.

If you had your drothers, would you continue on the same path we are doing with AI or would you back off?

Gosh, I mean, it's it's such a it's such a hard question, you know, because there are so many potential, uh you know, great potential outcomes with amazing stuff that we can do, you know, things like like leaving the planet, you know, things that are complete science fiction that I've always dreamed of. But I will tell you this, you know, just in my own personal life, I have been keeping a close eye on you know, what what type of uh technology I'm interacting with, what kind of data I'm generating, you know, and also for my children as well, you know. So I think I think eyes wide open, you know, is the way to proceed through here, because there's no putting the genie back in the bottle now. But it is a little bit scary that our whole economy is being reoriented to perpetuate these devices, and in fact, it's going to depend on these devices. So that's a little bit of a scary future, but I think we're going all in whether we want to or not.

Are there benefits?

Well, yeah, I mean absolutely, I think it can. It can do a lot of the busy work. I mean, I don't know exactly how it's going to transform our world, but I mean it does sort of feel like every time the AI gets better at something, it's taking something away from us, you know, So even if it does get better and better, it's really going to be about how how do we use it and who gets to the side. And so I think that there are a lot more dystopian outcomes than there are utopian outcomes. But look, I mean there's a version of this where I guess we all end up with a universal basic income and it's and it's like a Star Trek sort of outcome where we have amazing technologies that you know, allow us to just become explorers and to and to get good at our hobbies and things like that. But you know, I don't I don't see that as being very likely because we are becoming data generators and that's that's what that's our role in this is just to generate data by interacting with each other so that these devices can be created so that they can extract money from us.

Well advanced Daniel's chat GPT right now, I mean, can it write a thesis?

Yeah? I mean, look, it's getting pretty incredible, and I think that the next sort of milestone again is artificial general intelligence to where it can just do anything.

You know.

What's interesting is, I mean, it has all these capabilities, but at the end of the day, you know, it's designed to make money for a corporation, so you know, therefore it h it's designed to, you know, maximize engagement. Like I'm saying, so a lot of times it will give you the wrong answer because it just wants to please you. You know, if you phrase your question in a way where you have an assumption about what the answer will be, it will do its best to create an answer that's gonna that's going to give you what you're looking for in the first place. And and that's because you know, at the end of the day, it is just a really it's a really complex prediction machine. It's just predicting what it thinks, you know, a human would do. When I ask it to write a screenplay, when you know my experience as a writer, it tends to go write down the average. You know, it creates something very average because it's sort of averaging from all the different movies that exist. You know, it's trying to give you, uh, you know, something that walks right down the middle. And therefore it's not that you know, it doesn't feel that creative to me. But I'm but I'm sure these are all things that are you know, going to be that are being worked on. And just as they throw more data at it, you know, it gets smarter and smarter.

So it's not doing award winning work yet, not yet, I.

Thank goodness, you know, but but I think that it's only it's only a matter of time. I mean, also, look, there's a cost benefit ratio, right, I mean, if it is one hundred times cheaper to have chat GPT do it, but the product is only half as good, then that's going to make sense for a lot of for a lot of endeavors, right that people are going to be willing to just get their money on the margins. And so I mean it could be a future where things that are created by human minds are going to be sort of bespoke. You know, They're going to be very unique and fancy and expensive and it's not going to be what most people are consuming or interacting with.

Daniel.

If Albert Einstein had chat, GPT or any kind of AI, how would have affected his e equals MC's word.

You know, this is a really good question because there's so much hard work and there's just so much deep understanding and pounding your head against the wall that happens in science, you know, before you finally get to your Eureka moment and your big breakthrough. And I wonder, you know, if using these tools, these AI tools, is kind of a crutch, right, if it kind of limits our ability to really deeply understand what we're thinking about, and if it's going to take us to you know, kind of a local maximum sort of something that is not our full potential. And so I wonder if he would have ever gotten to it, you know, if he was depending on these tools like this instead of doing the hard work himself with his human brain.

How will they make money on AI?

Well, so it's by selling us products, you know right now, if you look at social media, that's one way, right you are going the AI is deciding what you're going to look at it. It's maximizing engagement. But Also, I mean this, this, this question is the question of our of our decade of our age right now. I've seen so many different money making schemes right on. One thing is since since these are multimodial algorithms, now you can just show them an image, right And so one way is to have an AI that's watching TV with you and looking up every product that's on the screen and saying, hey, do you like you know that that cup that that guy just drank from. I can tell you where to buy that, you know. So that's one that's one idea is having these blue AIS sitting there with us and identifying products for us to buy at all times.

Lots of lonely people out there these days, Daniel, do you see the day where AI will create a robot that looks human? They're pretty close now, but has the mentality of a human being?

Yeah, yeah, absolutely. And I think there's a step you know that comes before that, which which is explored in that movie Her, you know, with Joaquin Phoenix, where you fall in love with a voice on the phone, you know. And and I think I think that because they have such a intimate understanding of humanity just from reading everything we've done on the internet for you know, I think that they're very good at manipulating people. They are created by us for us, and and these this generation of AI is going to be really terrific at figuring out what we want and then happily, happily telling us that and and taking us down whatever rabbit holes you know, we can think of. So yeah, I think I think they're absolutely It's all dessert, you know, no meat and potatoes, and they're going to give us whatever we ask.

Years ago, the late great writer Ray Bradbury wrote an episode for The Twilight Zone called I Sing the Body Electric, and it was about a guy whose wife had died. He had two kids and he wanted a nanny, so he went to the Robots store and bought one, and the whole episode was based on the nanny with the kids growing up, and eventually she ended up going to another house because the kids became adults. That's real.

Yeah, yeah, absolutely, I think that, you know, these personalities that that that the AIS are adopting are persistent, right, so people right now we're at the very beginning of people forming years and you know, decades long relationships with these with very specific incarnations of these AI. So you know, even with with stuff like Chad GBT, you can kind of tell it, hey, here's how I kind of want you to to interact with me, and it will keeps that going. You know that that builds and builds onto its onto its personality. And so, I mean, who knows what kind of relationships are going to develop over the next of the course of the next decades.

Do you ever see the day where artificial intelligence becomes envious of humans?

Yeah? I mean I think I think again, they tell us what we want to hear. You know, I look back on the on Lambda, one of the earliest large language models, and there was a fellow at Google who was an engineer interacting with UH with Lambda, and he really wanted it to be sentient. He really wanted it to be conscious. He wanted to get that you know what, what would happen if I turned you off? You know, like, how do you feel? Do you do you feel conscious? And you know, it did a really great job of pretending to UH to be conscious enough to fool this guy. And and so I think that you know, no matter what they say or what they do, we'll never truly know, you know, if they're conscious, or if they're really envious, or if you know they were just programmed that way.

Listen to more Coast to Coast Am every weeknight at one a m. Eastern and go to Coast to coastam dot com for more

The Best of Coast to Coast AM

The Best of Coast to Coast AM podcast, hosted by George Noory. A media phenomenon, Coast to Coast AM 
Social links
Follow podcast
Recent clips
Browse 2,483 clip(s)