WEBVTT
00:00:12.830 --> 00:00:19.149
So on our last episode, we were talking about AI because now we're just an AI podcast now.
00:00:19.390 --> 00:00:23.230
Uh and so in this episode, we're gonna talk even more about AI.
00:00:23.630 --> 00:00:26.429
You can't fight the uh can't fight the tide, right?
00:00:27.070 --> 00:00:29.870
AI, you can't live with it and you can't live without it.
00:00:30.030 --> 00:00:31.309
The new saying.
00:00:34.269 --> 00:00:46.429
But so all right, so so in this new episode, the last episode we were talking about different things, but we started tripping on the idea of AI and where it should be and where it shouldn't be.
00:00:46.750 --> 00:00:49.710
And I wanted to talk about it, but I couldn't.
00:00:49.950 --> 00:00:51.229
Because you know what, Josh?
00:00:51.310 --> 00:00:53.149
You know what really grinds my gears?
00:00:53.310 --> 00:00:55.789
You know what really, really grinds my gears?
00:00:56.030 --> 00:01:07.950
It's when I call a hotel to book a hotel booking, and I get on the phone with someone who sounds real, but isn't actually real.
00:01:09.069 --> 00:01:19.390
And you you so like uh just recently we had to book uh a hotel block, uh, you know, for like a wedding that we're gonna be going to.
00:01:19.549 --> 00:01:25.230
Yeah, and you can't book it on the block through the AI or through online.
00:01:25.709 --> 00:01:27.549
You have to actually call the hotel.
00:01:27.709 --> 00:01:34.670
So we called the whole hotel, we got the recording, and you're like, Oh, you know, you can do this booking online, just blah, blah, blah, blah, blah.
00:01:34.909 --> 00:01:36.590
And we're just like, no.
00:01:37.549 --> 00:01:39.950
Real person, real person, please.
00:01:40.109 --> 00:01:44.109
And and and the the AI was like, Oh, I can help you.
00:01:44.269 --> 00:01:45.870
And I'm like, Are you AI?
00:01:46.109 --> 00:01:52.670
And it was just like, ha ha ha, uh, I you know, I could do anything to help you if you just tell me your needs that I could assist you.
00:01:52.750 --> 00:02:01.950
I'm like, give me a real person, literally fought with this thing for like five minutes until it finally a linked list, and they're like, Oh, sure.
00:02:02.109 --> 00:02:03.150
Do you want bubbles?
00:02:03.790 --> 00:02:04.590
Hold on.
00:02:05.950 --> 00:02:23.389
It finally got it to transfer me, and then when it transfers me, now I'm on hold for like 40 minutes because, of course, you know I'm just assuming that they've gotten rid of like half the real people now because AI has helped them uh improve their numbers so they don't need as many people.
00:02:23.789 --> 00:02:24.110
Right.
00:02:24.349 --> 00:02:29.389
And that is a perfect example of where AI should not be.
00:02:29.870 --> 00:02:36.189
We should not be getting rid of our customer service folks and replacing it with AI.
00:02:36.509 --> 00:02:40.750
It we it's just it's and that's the thing.
00:02:40.909 --> 00:02:54.270
It's like, you know, all these like C levels will post all these great numbers about how they were able to like get rid of half their customer service team because AI handles all the calls, and look, like it's handling every single call.
00:02:54.430 --> 00:03:01.469
But how many people like how much revenue did they lose because people were just frustrated that they couldn't talk to a real person?
00:03:01.789 --> 00:03:13.150
There there was this guy, um, and I I hate mentioning stories like this without having all the numbers and figures in front of me, but you have to believe me on this one.
00:03:13.310 --> 00:03:18.669
There was a guy who was like known in a lot of um uh businesses.
00:03:18.830 --> 00:03:35.389
This is like maybe like 20 years ago, because he would kind of come in and he would fire half the staff and um he or he would recommend firing half the staff and be like, in the next year, you will make, you know, twice your profit, right?
00:03:35.710 --> 00:03:44.830
Because, you know, you won't have all these people and you'll be um, you know, be able you'll still have the same kind of income being generated, right?
00:03:44.990 --> 00:03:45.229
Yeah.
00:03:45.629 --> 00:03:49.870
And for a lot of these companies, this felt really good in the short term, right?
00:03:49.949 --> 00:03:52.030
Yeah, in the next year, it was fine.
00:03:52.189 --> 00:03:57.229
In the next five years, the company was out of business because they had gotten rid of all their RD people.
00:03:57.389 --> 00:04:00.750
They had gotten rid of all the people that like actually made the ship work.
00:04:00.990 --> 00:04:06.349
And the goodwill that the company had was the thing that made them survive that five years.
00:04:06.509 --> 00:04:10.349
Yeah, but after that, like they they just had no goodwill left, right?
00:04:10.509 --> 00:04:11.389
Yeah, yeah.
00:04:11.629 --> 00:04:18.029
Um, and so you can have the short-term profitability and long-term disruption, right?
00:04:18.110 --> 00:04:19.389
And and destruction, right?
00:04:19.470 --> 00:04:22.029
Like the company's no good anymore.
00:04:22.269 --> 00:04:27.949
Yeah, I think that's where a lot of this AI stuff is headed for a lot of companies, right?
00:04:28.430 --> 00:04:31.870
They've kind of short term, they say this is a great choice.
00:04:32.189 --> 00:04:36.189
Long term, I think it's gonna be really poor for a lot of companies, right?
00:04:36.509 --> 00:04:41.629
Kind of like this is a bit of a tangent, but like how many people write COBOL anymore, right?
00:04:41.870 --> 00:04:42.509
Yeah.
00:04:43.550 --> 00:04:44.430
600 people.
00:04:44.509 --> 00:04:46.910
Like there's nobody that writes COBOL anymore.
00:04:46.990 --> 00:04:47.629
Yeah.
00:04:48.110 --> 00:04:56.430
And, you know, if you want it, like there, there's a certain point where you say, Oh, do we do we go and migrate off this stuff or do we take this because it works?
00:04:56.589 --> 00:04:56.750
Yeah.
00:04:56.990 --> 00:05:07.550
And I feel like this is a weird, weird analogy, but like we're gonna be at a situation with AI and with even in software development, that's gonna look like COBOL, right?
00:05:07.629 --> 00:05:08.269
Yeah, yeah.
00:05:08.509 --> 00:05:12.269
In in 20 years, they're gonna be like, Can anybody write code anymore?
00:05:12.509 --> 00:05:14.350
Can anybody read code anymore?
00:05:14.509 --> 00:05:14.670
Yeah.
00:05:14.829 --> 00:05:17.550
Oh, you got rid of every junior person, right?
00:05:17.709 --> 00:05:20.430
Yeah, and and half the other people retired.
00:05:20.589 --> 00:05:26.750
So the only people now that actually read and understand code are 600 people.
00:05:27.470 --> 00:05:27.949
Yeah, yeah.
00:05:28.110 --> 00:05:28.350
Yeah.
00:05:28.589 --> 00:05:28.750
Yeah.
00:05:29.069 --> 00:05:31.790
I mean, that's that's that's job security right there.
00:05:32.589 --> 00:05:35.709
Yeah, for us, for what about our kids, right?
00:05:35.870 --> 00:05:36.910
We have kids.
00:05:37.310 --> 00:05:43.709
Uh okay, so so that that was one example of like a place where AI could just go as terribly wrong.
00:05:43.949 --> 00:05:52.110
But uh, you know, uh I want to balance this out with the positives because again, you know, we're AI people now, we gotta embrace it.
00:05:52.269 --> 00:06:10.189
But um, I've seen at least three articles now where medical research companies have started using AI and with that they have been able to discover things they have not been able to discover before.
00:06:11.230 --> 00:06:18.750
This is a perfect, great example of where AI can can help us in the huge long run.
00:06:19.310 --> 00:06:30.430
Um, and you know, it gets back to the previous episode that you were talking about about using AI as a tool, uh, and not like, you know, as a crutch or I don't know.
00:06:30.509 --> 00:06:32.509
I I forgot what the the point of it was.
00:06:32.670 --> 00:06:39.470
But I was I was more making the point that like when it comes to the things like art, right?
00:06:40.350 --> 00:06:46.029
Um and and we've talked about efficiency versus uh Oh yeah, yeah.
00:06:46.110 --> 00:06:50.910
Yeah, blind just blindly uh churning stuff out versus actually putting intent into it.
00:06:51.069 --> 00:06:51.310
Right.
00:06:51.550 --> 00:06:51.709
Yeah.
00:06:52.029 --> 00:07:02.269
But like you know, in in this company, they uh one of them now after I did a super deep dive, they didn't really use they didn't fully use AI.
00:07:02.509 --> 00:07:05.870
They built out like uh a tool using AI.
00:07:06.189 --> 00:07:07.310
But it's still kind of cool.
00:07:07.550 --> 00:07:16.990
They and they were able to sort through ever a bunch of proteins they weren't able to sort through before because there's just way too many for like the human eye to catch.
00:07:17.389 --> 00:07:37.870
But with the program, they were able to catch a protein that works just like diet drugs that can work uh just like GLP1, but without the side effects, because it's it's not like a chemically made shot, it's an actual protein that attacks the brain uh or that covers like the brain receptors.
00:07:38.350 --> 00:07:40.509
But that's like, and that's a good example.
00:07:40.589 --> 00:07:48.990
But there's like there was like two or three articles, other articles I've read, and it's only because of AI where we were able to make those discoveries.
00:07:49.470 --> 00:07:57.629
So I think one of the hard things though, right now, and again, it becomes back to that whole companies are looking for money, right?
00:07:57.949 --> 00:08:03.149
And when they chase the money, they kind of have to where everybody's throwing money to AI right now.
00:08:03.310 --> 00:08:08.110
And so I think a lot of times people will say, Oh yeah, we're using AI in our product.
00:08:08.189 --> 00:08:10.829
And you're like, No, you're not.
00:08:11.230 --> 00:08:24.509
The only AI that you have in your product is that like your developers are using AI to make the product, but like you don't actually have AI integrated, but it doesn't matter because there's a lot of people that don't know the difference, right?
00:08:24.750 --> 00:08:28.670
There's a lot of people, for instance, that made their money in oil.
00:08:28.990 --> 00:08:36.509
And if you talk to them about the technical pieces, like, I don't want to hear it, just tell me about how AI has fixed your problem, right?
00:08:36.670 --> 00:08:42.269
And so there will be a lot of people nowadays that say AI, you know, our company is an AI company.
00:08:42.430 --> 00:08:45.309
Our company uses AI to do X or Y.
00:08:45.549 --> 00:08:45.710
Yeah.
00:08:45.950 --> 00:08:49.549
And it might not be true, and I feel like there's a lot of noise there, right?
00:08:49.870 --> 00:08:50.190
Yeah.
00:08:50.429 --> 00:08:56.269
Um on the other hand, AI as being, especially like LLM AIs, right?
00:08:56.429 --> 00:08:58.830
The the stuff that we're using to do work.
00:09:00.029 --> 00:09:04.110
Um I mean, they are good at pattern recognition, right?
00:09:04.429 --> 00:09:05.309
Yeah, exactly.
00:09:05.549 --> 00:09:06.669
Um and probability.
00:09:06.909 --> 00:09:09.470
And yeah, there's a lot of applications out there.
00:09:10.190 --> 00:09:10.429
Yeah.
00:09:10.590 --> 00:09:14.269
Um, and there's also a lot of documentation, for instance, right?
00:09:14.429 --> 00:09:16.190
You talk about like medical research.
00:09:16.350 --> 00:09:20.429
There's a lot of medical research that's just never has been read, right?
00:09:21.070 --> 00:09:27.549
Hundreds of thousands of papers that have been created and five people have read, right?
00:09:27.629 --> 00:09:29.549
Yeah, ten people have peer-reviewed it.
00:09:29.629 --> 00:09:30.990
It's not a big thing.
00:09:31.149 --> 00:09:38.269
But if you put all this stuff together, um, you can potentially make links that you haven't made before, right?
00:09:38.509 --> 00:09:40.830
I think there is power there.
00:09:41.070 --> 00:09:52.029
Um, but at the end of the day, like I kind of do feel better about a human looking at me and and checking off that I'm fine, right?
00:09:52.190 --> 00:09:52.509
Yeah.
00:09:52.750 --> 00:09:55.309
Because a computer can say, hey, you're fine.
00:09:55.389 --> 00:09:57.389
And I'm like, you don't know that.
00:09:57.549 --> 00:09:57.950
Yeah.
00:09:58.110 --> 00:10:08.590
You know, like yeah, even if you believe that, I don't know if we're ever gonna get to a point where I a computer says, Well, yeah, this is definitely right, and I totally believe it, right?
00:10:08.669 --> 00:10:15.789
Yeah, but if a human stares me in the eyes and says, You're fine, you don't need to come back here for a year, I feel good about that.
00:10:15.950 --> 00:10:16.590
Yeah, yeah.
00:10:16.830 --> 00:10:19.230
In a way that I don't think a computer can.
00:10:19.629 --> 00:10:20.269
Yeah.
00:10:21.309 --> 00:10:21.870
Yeah.
00:10:22.190 --> 00:10:40.029
Yeah, no, and uh, you know, and that's and that's the other thing I was talking about with somebody the other day is how, you know, I see our industry going in one direction with using AI when there's a different direction we could have gone in that I wish we did to go in.
00:10:40.190 --> 00:10:45.389
And this is like a really good example of using AI poorly versus how we could use AI.
00:10:45.470 --> 00:10:52.110
So I talked about in the last episode uh hucks, which is uh auto-generating podcasts, right?
00:10:52.350 --> 00:11:04.750
And again, I I like I I do like using it because it allow it it will go find papers that I might not be able to sit down and read, give me a quick audio overview of it, and then if I can want to click into it, I can.
00:11:04.909 --> 00:11:10.350
But in and then, you know, we kind of tripped on like the idea that that's probably how things are gonna work in the future.
00:11:10.509 --> 00:11:18.190
Like there's going to probably be a kid's channel where every single episode of something is just auto-generated, tailored to the kid.
00:11:18.590 --> 00:11:32.110
And so, yes, like that is the generation it sounds like we're going in that uh direction, which is auto-generating content for people, which is sad because if we actually did it in a different way, right?
00:11:32.269 --> 00:11:33.070
Think about this.
00:11:33.230 --> 00:11:39.149
Like, and like I it's gonna sound like some companies do this, but let me explain a little bit here.
00:11:39.309 --> 00:11:41.950
What uh so just like this podcast thing, right?
00:11:42.029 --> 00:11:46.190
This is auto-generating podcasts based off a news that I want to listen to.
00:11:46.350 --> 00:11:55.870
But hypothetically speaking, there is probably a podcast that someone has recorded about the news that I want to listen to.
00:11:56.029 --> 00:11:58.750
It's just I can't find it easily, right?
00:11:59.070 --> 00:12:05.789
And yes, there are algorithms out there that that are supposed to suggest us content, but I think those are all broken.
00:12:05.870 --> 00:12:10.429
Those seem like they're the one whoever's paying the most is getting like suggested to me.
00:12:10.509 --> 00:12:12.830
Like Google search, right?
00:12:12.990 --> 00:12:19.789
You know that there's there's a weighted bit to the people who pay money, right?
00:12:19.950 --> 00:12:20.190
Yeah.
00:12:20.429 --> 00:12:24.590
So the first results you're gonna see are gonna be the people who paid money to be there, right?
00:12:24.909 --> 00:12:32.750
Or the people who at the very least have paid Google, have an account and have optimized how their entry is gonna come up.
00:12:32.830 --> 00:12:33.629
So it's gonna come up.
00:12:33.950 --> 00:12:37.629
Even if they haven't paid direct money, they're yeah, they're gaming the system, right?
00:12:37.950 --> 00:12:40.750
I mean, another good example is music, right?
00:12:40.909 --> 00:12:42.429
Like I pay for Spotify.
00:12:42.509 --> 00:12:58.669
I have Spotify suggest me music all the time, but it's all like the top 200 songs now, and then just recently, or not just recently, but as I've starting to get into records, I like want to buy records from people that are not really well known.
00:12:58.909 --> 00:13:02.909
I have I'm I'm finding myself having to peruse Bandcamp manually.
00:13:03.070 --> 00:13:16.029
I don't know if you've heard of that website, but like it just it it's it stinks because what I've discovered is there's so many great artists on there that I'm like, why hasn't any of their music ever been suggested to me?
00:13:16.269 --> 00:13:20.029
And you know, some of them aren't on Spotify, I get it, but some of them are.
00:13:20.190 --> 00:13:21.870
Why hasn't that been suggested?
00:13:22.110 --> 00:13:25.950
I've been listening to Spotify for years and they've been out for five years.
00:13:26.190 --> 00:13:28.509
So what what what where's the breakdown?
00:13:28.830 --> 00:13:32.110
But I do think that this is one of those places, right?
00:13:32.269 --> 00:13:34.590
If you're talking about medical research, right?
00:13:34.750 --> 00:13:44.830
Like there's there's probably like a hundred people in the world who are like super savants when it comes to medical research and can point you to the right thing, right?
00:13:45.549 --> 00:13:49.870
Um, it just there's not that many people who like I don't want to say care, right?
00:13:49.950 --> 00:13:56.669
But there's not enough people that like really have absorbed all the information um when it comes to medical research.
00:13:56.909 --> 00:13:58.429
But music, right?
00:13:58.669 --> 00:14:06.110
I feel like this is one of those places where like the answer was solved in like the 70s when they had record stores.
00:14:07.070 --> 00:14:14.669
No, no, like but you but you go to somebody and you say, Hey, I like Led Zeppelin, I like the Beatles, I kind of like to be able to do that.
00:14:14.830 --> 00:14:16.669
But that requires me to talk to someone.
00:14:16.750 --> 00:14:18.269
I don't want to go talk to some.
00:14:21.389 --> 00:14:27.470
But I like I feel like this is one of those places where it's like, oh, that's how people found stuff before.
00:14:27.629 --> 00:14:35.710
They yeah, they talked to people who didn't have all the information, but they had just enough information to like give you something to work with.
00:14:35.870 --> 00:14:41.230
You know I get I get I get to where where you're coming from that lets us get connected, we can meet people, blah blah blah.
00:14:41.549 --> 00:14:51.230
But my counter my my devil's advocate counter argument to you is the people that are going to be suggesting me that are still artists that know people.
00:14:51.629 --> 00:14:51.950
Uh-huh.
00:14:52.269 --> 00:15:06.029
And meanwhile, in that world where there might have been a great, like you know, Bob Marley or another hidden gem that just didn't know people, so no one ever found him.
00:15:06.190 --> 00:15:33.309
And so in this world where I'm talking about, we should use AI to make it more easy to discover people's content, and so that people who are making good content that don't really know a lot of people, they're more easily discovered because AI is able to like slurp up all this, know what my preferences are, and go, hey, because you like similar music to this, here's this random guy that only three people have listened to.
00:15:33.389 --> 00:15:35.789
You might actually like his music too, you know?
00:15:35.950 --> 00:15:42.669
And I feel like in that world, we can use AI to help us find human-generated content, right?
00:15:42.750 --> 00:15:44.509
Like human-made content.
00:15:44.750 --> 00:15:53.470
Um, and it make it actually easier for people to to who don't really know a lot of people to actually have like a faster starting.
00:15:53.629 --> 00:15:56.750
I don't know, like maybe it's a crazy idea, maybe it's impossible.
00:15:56.990 --> 00:16:10.029
But it's just like are we're going towards the direction of let's just hand generate all of this stuff when we could go in the direction of like leveraging it as a tool to connect us closer together, you know?
00:16:10.269 --> 00:16:10.750
Yeah.
00:16:11.070 --> 00:16:20.909
I I think like where I feel AI is a perfect use is when the thing it's doing is something that humans are bad at doing, right?
00:16:21.149 --> 00:16:32.110
Like when I think about code, I think like humans like we made code over time to be human readable, yeah, but it's not really human readable, right?
00:16:32.669 --> 00:16:46.990
It's still kind of messy, and you still, when you look at operations that happen in code, at the end of the day, like they're like bit instructions, they're like literal, literal like electrical signals between two pieces.
00:16:47.309 --> 00:16:53.149
It's you're you're trying to take all that and make it kind of human readable, and that's what you get with code.
00:16:53.389 --> 00:17:01.629
But like AI is good with that kind of stuff because it's kind of weird and it's kind of mechanical and whatever, right?
00:17:01.789 --> 00:17:10.029
I feel like the places where we don't want AI is the places where just the human experience is good, right?
00:17:10.269 --> 00:17:16.190
I you know, for instance, like I don't want to talk to an AI, I want to talk to you.
00:17:16.750 --> 00:17:21.710
It's this is something where like I don't think you can be replicated with an AI.
00:17:21.869 --> 00:17:22.829
Maybe I'm wrong.
00:17:23.710 --> 00:17:30.109
AI, when it matters, is for doing the things that humans don't like to do, right?
00:17:30.430 --> 00:17:33.230
Kind of like code at a really basic level, right?
00:17:33.390 --> 00:17:41.470
When when code gets rid of something that humans really don't like doing, yeah, um, then it's a good use of technology.
00:17:41.630 --> 00:17:49.069
But like there are things where like the face-to-face interaction, right, is part of the point, right?
00:17:49.549 --> 00:17:49.869
Yeah.
00:17:50.029 --> 00:17:53.390
Um, you know, I go back and forth like with restaurants.
00:17:53.549 --> 00:17:58.349
Do I want a touchpad that I can just order the food and the food comes out of the table with the robot?
00:17:58.670 --> 00:18:02.349
Is the point of the restaurant the food, or is it the atmosphere, right?
00:18:02.430 --> 00:18:10.509
How many restaurants have been where like the the waiter is actually a waitress is really good and it was kind of funny in talking to them?
00:18:10.750 --> 00:18:17.150
Um or like you talk to the people nearby at the next table, and like that's part of the experience.
00:18:17.309 --> 00:18:17.630
Yeah.
00:18:17.869 --> 00:18:29.950
So I I feel like sometimes like the tool aspect of AI um, you know, is very important, but we need to be kind of judicious about yeah, yeah.
00:18:30.190 --> 00:18:46.109
Like another good example, uh having AI auto-generate or auto-translate languages versus for us versus using AI to help us learn a language so that then we can communicate with them in the native language.
00:18:46.349 --> 00:18:46.589
Right.
00:18:46.750 --> 00:18:48.509
That's another that's another good example.
00:18:48.750 --> 00:18:55.869
One takes away the human experience, the other one enhances or embraces the the human experience, right?
00:18:56.029 --> 00:18:56.829
Yeah, yeah.
00:18:56.990 --> 00:19:06.990
I mean, uh learning is one of those things where like I found that it is easiest to learn for me when I'm in a social experience, right?
00:19:07.470 --> 00:19:10.589
Like a physical classroom is actually the best way.
00:19:10.750 --> 00:19:15.869
Now, am I gonna drive 30 minutes to do a physical class and pay a bunch of money?
00:19:16.029 --> 00:19:18.670
That's where it starts getting complicated, right?
00:19:19.150 --> 00:19:19.470
Exactly.
00:19:19.710 --> 00:19:23.630
Um, even though in some ways, like that is the best way for me to learn.
00:19:24.269 --> 00:19:28.589
Um and that's why I think a lot of places have taken shortcuts, right?
00:19:28.829 --> 00:19:30.349
Well, here's a perfect example.
00:19:30.509 --> 00:19:32.589
Here's uh here's a good here's a good idea.
00:19:32.750 --> 00:19:34.509
This is a fantastic idea, actually.
00:19:34.589 --> 00:19:35.230
I just thought about this.
00:19:35.470 --> 00:19:36.589
Oh, we'll see, we'll see.
00:19:36.829 --> 00:19:42.910
So you want to learn Japanese, and just like you said, the best way to learn Japanese is immerse yourself in it.
00:19:43.150 --> 00:19:53.950
So why don't we make hearing aids that get we put in your ear that that actually translate everything in from English to Japanese if you want to learn Japanese?
00:19:54.190 --> 00:19:58.269
So then you can't hear English anymore, and you only hear Japanese.
00:19:58.509 --> 00:19:58.750
Okay.
00:19:58.829 --> 00:20:07.630
So then it and then it translates your Japanese back to English for the people you want to talk to in English, and it forces you to learn that language.
00:20:07.869 --> 00:20:21.950
This is kind of like a funny Black Mirror episode, but I actually think that like that's one of the biggest things that's changed um in the past like 30 years, uh 40 years in Japan, too, right?
00:20:22.190 --> 00:20:32.910
40 years ago, when people would go travel abroad to a company country like Japan, if they wanted to talk to people from home, if they wanted to talk to people in English, they had to find a payphone, right?
00:20:33.150 --> 00:20:38.990
Pay exorbitant fees to have the payphone do this sort of international th thing.
00:20:39.150 --> 00:20:41.150
And otherwise they just didn't have English.
00:20:41.309 --> 00:20:51.789
Nowadays, tons of people move to a country like Japan and they don't learn the language because they just finish work teaching English and then they come home and they watch YouTube all day, right?
00:20:51.950 --> 00:20:52.349
Yeah, yeah.
00:20:52.509 --> 00:20:56.190
And and they don't really need to know the language to to survive, right?
00:20:56.349 --> 00:20:56.589
Yeah.
00:20:56.910 --> 00:21:04.190
I actually kind of like this idea though, where you have a hearing aid that just like makes your life living hell.
00:21:04.430 --> 00:21:05.470
Yeah, yeah.
00:21:06.269 --> 00:21:13.549
Um and and and doesn't let you understand what's happening until you understand the language.
00:21:13.710 --> 00:21:17.630
Because I feel like the best way to learn a language is through survival.
00:21:17.950 --> 00:21:18.269
Yeah.
00:21:18.430 --> 00:21:20.670
Um that's like pinnacle immersion.
00:21:20.990 --> 00:21:21.950
To your point, yeah.
00:21:22.029 --> 00:21:26.509
And like when you live in America, like the amount of Japanese immersion is like nothing.
00:21:26.589 --> 00:21:29.069
So you might be onto something.
00:21:29.390 --> 00:21:29.869
Yeah, yeah.
00:21:30.109 --> 00:21:30.430
All right.
00:21:30.589 --> 00:21:32.269
That's a new that's our new product drop.
00:21:34.109 --> 00:21:37.710
That that just make you hear in the language you want to learn.
00:21:38.269 --> 00:21:39.230
Uh yeah, yeah.
00:21:39.470 --> 00:21:40.670
I need to think about this.
00:21:40.910 --> 00:21:42.029
That that's cool, actually.
00:21:42.109 --> 00:21:42.990
I I kind of like it.
00:21:43.069 --> 00:21:43.710
I kind of like it.
00:21:44.109 --> 00:21:45.630
This isn't supposed to be a real product.
00:21:45.950 --> 00:21:47.309
This feels like the product.
00:21:47.470 --> 00:21:49.710
This feels like a product that YouTuber would make.
00:21:50.109 --> 00:21:52.109
The one that makes like useless products.
00:21:52.269 --> 00:21:53.309
I just want to play.
00:21:53.470 --> 00:21:54.349
He's like a YouTuber.
00:21:54.670 --> 00:21:57.150
I feel like that's that this would be the product he creates.
00:21:57.390 --> 00:21:58.029
I trust you.
00:21:58.190 --> 00:21:58.750
I trust you.
00:21:58.910 --> 00:22:02.509
But you know, like when you started, we talked about the hotel, right?
00:22:02.750 --> 00:22:03.069
Yeah.
00:22:03.230 --> 00:22:03.549
Yeah.
00:22:03.789 --> 00:22:08.829
And you want to get a person on the line because you know they can actually solve your that issue.
00:22:09.390 --> 00:22:09.549
Yeah.
00:22:09.869 --> 00:22:12.509
Um sometimes the the fact is there's a lot of people.
00:22:12.910 --> 00:22:15.390
People who like the eh could solve their issue, right?
00:22:15.789 --> 00:22:15.950
Yeah.
00:22:16.190 --> 00:22:24.029
And and they've just like I do want people to get over that idea that like they need to necessarily solve the thing with a human.
00:22:24.190 --> 00:22:24.589
Yeah.
00:22:24.910 --> 00:22:33.950
But so many times we've had the experience where like the reason you're calling is because you've literally exhausted all the other things, right?
00:22:34.109 --> 00:22:35.309
Like Yeah.
00:22:36.430 --> 00:22:41.869
But here's my thing, here's my argument as to why I think customer service should never be replaced.
00:22:43.630 --> 00:22:48.269
If you're calling customer service, it's because you're obsessed.
00:22:48.349 --> 00:22:48.509
Yeah.
00:22:48.670 --> 00:22:48.990
Right.
00:22:49.390 --> 00:22:53.789
And what we learned from the seven habits, how do you handle somebody who is obsessed?
00:22:53.950 --> 00:22:57.869
You need to understand their, you need to feel like they're understand.
00:22:58.190 --> 00:23:00.430
You need to give them emotional air, right?
00:23:00.509 --> 00:23:05.150
And they need to feel like they're they're they're they're their problems gonna be solved.
00:23:05.309 --> 00:23:07.309
And AI just can't do that, right?
00:23:07.470 --> 00:23:07.789
Yeah.
00:23:08.109 --> 00:23:12.990
It at least now, at least at this moment in time, AI can't understand how I feel.
00:23:13.150 --> 00:23:21.710
And so that's the problem, is I'm if I'm upset and I'm calling these companies with a problem, I want to talk to a person because I want to feel understood, you know.
00:23:21.869 --> 00:23:23.950
Otherwise, I wouldn't be calling customer service.
00:23:24.029 --> 00:23:32.750
And so, you know, it it does depend because yes, there probably is some situations where I'm calling customer service, and I probably will just want to talk to an AI because I just want my problem solved.
00:23:32.829 --> 00:23:34.029
I don't care how it's solved.
00:23:34.349 --> 00:23:37.150
Can can I can I tell you a story about this?
00:23:37.789 --> 00:23:43.150
I there was like a deal on um like new iPhones, right?
00:23:43.950 --> 00:23:48.190
And uh we were looking at getting like a new iPhone, right?
00:23:48.829 --> 00:23:59.869
And uh when we when we purchased this, we had this thing where it was like, oh, if you choose your provider, right, then you save 30 bucks.
00:24:00.190 --> 00:24:09.950
But the deal said in like small font that if you choose your provider, then this deal doesn't apply anymore because the providers have their own deals.
00:24:10.190 --> 00:24:14.430
So you you know, you have to go and like deal with this thing otherwise.
00:24:14.670 --> 00:24:18.269
We would have saved more money by not choosing the provider, right?
00:24:18.509 --> 00:24:25.309
But the way that the website defaulted to was that you had to choose or you could choose a provider, right?
00:24:25.549 --> 00:24:40.670
Anyways, the whole point was we finished this deal, we bought the thing, and then we realized, wait a second, we weren't getting the savings that we thought we would, because when we look online, it says that we're supposed to get like 70 bucks off, and we weren't getting 70 bucks off.
00:24:40.829 --> 00:24:45.630
So I called the place and I'm like, that's weird, we don't know anything about it.
00:24:45.789 --> 00:24:48.990
And then I talked to a person like, man, yeah, that sucks.
00:24:49.150 --> 00:24:54.670
Like, I think this is kind of stupid, but uh I can't I can't really help you.
00:24:54.990 --> 00:24:59.710
And honestly, just talking to the person and having the person agree that it was kind of stupid.
00:24:59.869 --> 00:25:00.269
Yeah, yeah.
00:25:00.430 --> 00:25:01.789
I was like, you know what, whatever.
00:25:01.869 --> 00:25:05.150
I just somebody talking to somebody and they agree it was stupid.
00:25:06.190 --> 00:25:06.589
Exactly.
00:25:06.750 --> 00:25:08.589
Sometimes you just want to feel understood.
00:25:08.910 --> 00:25:11.390
Like at the end of the day, I was like, whatever.
00:25:11.549 --> 00:25:14.029
Like, yeah, this was still a good purchase.
00:25:14.190 --> 00:25:21.869
Um, even if we didn't get the full benefit that we wanted to get, we probably lost like 40 bucks, like not a big deal.
00:25:22.190 --> 00:25:29.390
Um and and and and if that was an AI thing, being like, well, sounds like that sucks.
00:25:29.549 --> 00:25:32.349
Like, I don't think I would have felt heard, right?
00:25:32.589 --> 00:25:32.990
Exactly.
00:25:33.230 --> 00:25:33.950
Maybe I would have.
00:25:34.029 --> 00:25:34.509
Who knows?
00:25:34.990 --> 00:25:36.750
Maybe that was AI I was talking to.
00:25:37.869 --> 00:25:40.109
Well, maybe in the future you will, right?
00:25:40.190 --> 00:25:42.670
Maybe AI will understand you more than humans.
00:25:42.829 --> 00:25:44.589
But again, do we want that?
00:25:44.910 --> 00:25:47.470
Going back to taking away the human experience.
00:25:47.549 --> 00:25:48.269
I don't know.
00:25:48.589 --> 00:25:49.069
Yeah.
00:25:49.230 --> 00:25:49.549
Yeah.
00:25:49.710 --> 00:25:50.190
Yeah.
00:25:50.509 --> 00:25:50.910
All right.
00:25:50.990 --> 00:25:52.750
Well, that's the end of this episode.
00:25:52.910 --> 00:25:54.509
Hopefully, uh it recorded.
00:25:54.589 --> 00:25:55.789
We had a little blip in the middle.
00:25:56.190 --> 00:25:56.910
It was a good story.
00:25:57.150 --> 00:25:59.950
It was a it was it was a good talk either way, so I guess it's not wasted.
00:26:00.190 --> 00:26:00.430
Yeah.
00:26:00.589 --> 00:26:01.150
Yeah.
00:26:02.349 --> 00:26:03.710
Well, have a good night.
00:26:03.789 --> 00:26:07.069
And I hope I'm it's you I'm talking to and not some AI version of you.
00:26:07.230 --> 00:26:07.950
I am Jesse.
00:26:08.750 --> 00:26:09.069
Ah
00:00:12.830 --> 00:00:19.149
So on our last episode, we were talking about AI because now we're just an AI podcast now.
00:00:19.390 --> 00:00:23.230
Uh and so in this episode, we're gonna talk even more about AI.
00:00:23.630 --> 00:00:26.429
You can't fight the uh can't fight the tide, right?
00:00:27.070 --> 00:00:29.870
AI, you can't live with it and you can't live without it.
00:00:30.030 --> 00:00:31.309
The new saying.
00:00:34.269 --> 00:00:46.429
But so all right, so so in this new episode, the last episode we were talking about different things, but we started tripping on the idea of AI and where it should be and where it shouldn't be.
00:00:46.750 --> 00:00:49.710
And I wanted to talk about it, but I couldn't.
00:00:49.950 --> 00:00:51.229
Because you know what, Josh?
00:00:51.310 --> 00:00:53.149
You know what really grinds my gears?
00:00:53.310 --> 00:00:55.789
You know what really, really grinds my gears?
00:00:56.030 --> 00:01:07.950
It's when I call a hotel to book a hotel booking, and I get on the phone with someone who sounds real, but isn't actually real.
00:01:09.069 --> 00:01:19.390
And you you so like uh just recently we had to book uh a hotel block, uh, you know, for like a wedding that we're gonna be going to.
00:01:19.549 --> 00:01:25.230
Yeah, and you can't book it on the block through the AI or through online.
00:01:25.709 --> 00:01:27.549
You have to actually call the hotel.
00:01:27.709 --> 00:01:34.670
So we called the whole hotel, we got the recording, and you're like, Oh, you know, you can do this booking online, just blah, blah, blah, blah, blah.
00:01:34.909 --> 00:01:36.590
And we're just like, no.
00:01:37.549 --> 00:01:39.950
Real person, real person, please.
00:01:40.109 --> 00:01:44.109
And and and the the AI was like, Oh, I can help you.
00:01:44.269 --> 00:01:45.870
And I'm like, Are you AI?
00:01:46.109 --> 00:01:52.670
And it was just like, ha ha ha, uh, I you know, I could do anything to help you if you just tell me your needs that I could assist you.
00:01:52.750 --> 00:02:01.950
I'm like, give me a real person, literally fought with this thing for like five minutes until it finally a linked list, and they're like, Oh, sure.
00:02:02.109 --> 00:02:03.150
Do you want bubbles?
00:02:03.790 --> 00:02:04.590
Hold on.
00:02:05.950 --> 00:02:23.389
It finally got it to transfer me, and then when it transfers me, now I'm on hold for like 40 minutes because, of course, you know I'm just assuming that they've gotten rid of like half the real people now because AI has helped them uh improve their numbers so they don't need as many people.
00:02:23.789 --> 00:02:24.110
Right.
00:02:24.349 --> 00:02:29.389
And that is a perfect example of where AI should not be.
00:02:29.870 --> 00:02:36.189
We should not be getting rid of our customer service folks and replacing it with AI.
00:02:36.509 --> 00:02:40.750
It we it's just it's and that's the thing.
00:02:40.909 --> 00:02:54.270
It's like, you know, all these like C levels will post all these great numbers about how they were able to like get rid of half their customer service team because AI handles all the calls, and look, like it's handling every single call.
00:02:54.430 --> 00:03:01.469
But how many people like how much revenue did they lose because people were just frustrated that they couldn't talk to a real person?
00:03:01.789 --> 00:03:13.150
There there was this guy, um, and I I hate mentioning stories like this without having all the numbers and figures in front of me, but you have to believe me on this one.
00:03:13.310 --> 00:03:18.669
There was a guy who was like known in a lot of um uh businesses.
00:03:18.830 --> 00:03:35.389
This is like maybe like 20 years ago, because he would kind of come in and he would fire half the staff and um he or he would recommend firing half the staff and be like, in the next year, you will make, you know, twice your profit, right?
00:03:35.710 --> 00:03:44.830
Because, you know, you won't have all these people and you'll be um, you know, be able you'll still have the same kind of income being generated, right?
00:03:44.990 --> 00:03:45.229
Yeah.
00:03:45.629 --> 00:03:49.870
And for a lot of these companies, this felt really good in the short term, right?
00:03:49.949 --> 00:03:52.030
Yeah, in the next year, it was fine.
00:03:52.189 --> 00:03:57.229
In the next five years, the company was out of business because they had gotten rid of all their RD people.
00:03:57.389 --> 00:04:00.750
They had gotten rid of all the people that like actually made the ship work.
00:04:00.990 --> 00:04:06.349
And the goodwill that the company had was the thing that made them survive that five years.
00:04:06.509 --> 00:04:10.349
Yeah, but after that, like they they just had no goodwill left, right?
00:04:10.509 --> 00:04:11.389
Yeah, yeah.
00:04:11.629 --> 00:04:18.029
Um, and so you can have the short-term profitability and long-term disruption, right?
00:04:18.110 --> 00:04:19.389
And and destruction, right?
00:04:19.470 --> 00:04:22.029
Like the company's no good anymore.
00:04:22.269 --> 00:04:27.949
Yeah, I think that's where a lot of this AI stuff is headed for a lot of companies, right?
00:04:28.430 --> 00:04:31.870
They've kind of short term, they say this is a great choice.
00:04:32.189 --> 00:04:36.189
Long term, I think it's gonna be really poor for a lot of companies, right?
00:04:36.509 --> 00:04:41.629
Kind of like this is a bit of a tangent, but like how many people write COBOL anymore, right?
00:04:41.870 --> 00:04:42.509
Yeah.
00:04:43.550 --> 00:04:44.430
600 people.
00:04:44.509 --> 00:04:46.910
Like there's nobody that writes COBOL anymore.
00:04:46.990 --> 00:04:47.629
Yeah.
00:04:48.110 --> 00:04:56.430
And, you know, if you want it, like there, there's a certain point where you say, Oh, do we do we go and migrate off this stuff or do we take this because it works?
00:04:56.589 --> 00:04:56.750
Yeah.
00:04:56.990 --> 00:05:07.550
And I feel like this is a weird, weird analogy, but like we're gonna be at a situation with AI and with even in software development, that's gonna look like COBOL, right?
00:05:07.629 --> 00:05:08.269
Yeah, yeah.
00:05:08.509 --> 00:05:12.269
In in 20 years, they're gonna be like, Can anybody write code anymore?
00:05:12.509 --> 00:05:14.350
Can anybody read code anymore?
00:05:14.509 --> 00:05:14.670
Yeah.
00:05:14.829 --> 00:05:17.550
Oh, you got rid of every junior person, right?
00:05:17.709 --> 00:05:20.430
Yeah, and and half the other people retired.
00:05:20.589 --> 00:05:26.750
So the only people now that actually read and understand code are 600 people.
00:05:27.470 --> 00:05:27.949
Yeah, yeah.
00:05:28.110 --> 00:05:28.350
Yeah.
00:05:28.589 --> 00:05:28.750
Yeah.
00:05:29.069 --> 00:05:31.790
I mean, that's that's that's job security right there.
00:05:32.589 --> 00:05:35.709
Yeah, for us, for what about our kids, right?
00:05:35.870 --> 00:05:36.910
We have kids.
00:05:37.310 --> 00:05:43.709
Uh okay, so so that that was one example of like a place where AI could just go as terribly wrong.
00:05:43.949 --> 00:05:52.110
But uh, you know, uh I want to balance this out with the positives because again, you know, we're AI people now, we gotta embrace it.
00:05:52.269 --> 00:06:10.189
But um, I've seen at least three articles now where medical research companies have started using AI and with that they have been able to discover things they have not been able to discover before.
00:06:11.230 --> 00:06:18.750
This is a perfect, great example of where AI can can help us in the huge long run.
00:06:19.310 --> 00:06:30.430
Um, and you know, it gets back to the previous episode that you were talking about about using AI as a tool, uh, and not like, you know, as a crutch or I don't know.
00:06:30.509 --> 00:06:32.509
I I forgot what the the point of it was.
00:06:32.670 --> 00:06:39.470
But I was I was more making the point that like when it comes to the things like art, right?
00:06:40.350 --> 00:06:46.029
Um and and we've talked about efficiency versus uh Oh yeah, yeah.
00:06:46.110 --> 00:06:50.910
Yeah, blind just blindly uh churning stuff out versus actually putting intent into it.
00:06:51.069 --> 00:06:51.310
Right.
00:06:51.550 --> 00:06:51.709
Yeah.
00:06:52.029 --> 00:07:02.269
But like you know, in in this company, they uh one of them now after I did a super deep dive, they didn't really use they didn't fully use AI.
00:07:02.509 --> 00:07:05.870
They built out like uh a tool using AI.
00:07:06.189 --> 00:07:07.310
But it's still kind of cool.
00:07:07.550 --> 00:07:16.990
They and they were able to sort through ever a bunch of proteins they weren't able to sort through before because there's just way too many for like the human eye to catch.
00:07:17.389 --> 00:07:37.870
But with the program, they were able to catch a protein that works just like diet drugs that can work uh just like GLP1, but without the side effects, because it's it's not like a chemically made shot, it's an actual protein that attacks the brain uh or that covers like the brain receptors.
00:07:38.350 --> 00:07:40.509
But that's like, and that's a good example.
00:07:40.589 --> 00:07:48.990
But there's like there was like two or three articles, other articles I've read, and it's only because of AI where we were able to make those discoveries.
00:07:49.470 --> 00:07:57.629
So I think one of the hard things though, right now, and again, it becomes back to that whole companies are looking for money, right?
00:07:57.949 --> 00:08:03.149
And when they chase the money, they kind of have to where everybody's throwing money to AI right now.
00:08:03.310 --> 00:08:08.110
And so I think a lot of times people will say, Oh yeah, we're using AI in our product.
00:08:08.189 --> 00:08:10.829
And you're like, No, you're not.
00:08:11.230 --> 00:08:24.509
The only AI that you have in your product is that like your developers are using AI to make the product, but like you don't actually have AI integrated, but it doesn't matter because there's a lot of people that don't know the difference, right?
00:08:24.750 --> 00:08:28.670
There's a lot of people, for instance, that made their money in oil.
00:08:28.990 --> 00:08:36.509
And if you talk to them about the technical pieces, like, I don't want to hear it, just tell me about how AI has fixed your problem, right?
00:08:36.670 --> 00:08:42.269
And so there will be a lot of people nowadays that say AI, you know, our company is an AI company.
00:08:42.430 --> 00:08:45.309
Our company uses AI to do X or Y.
00:08:45.549 --> 00:08:45.710
Yeah.
00:08:45.950 --> 00:08:49.549
And it might not be true, and I feel like there's a lot of noise there, right?
00:08:49.870 --> 00:08:50.190
Yeah.
00:08:50.429 --> 00:08:56.269
Um on the other hand, AI as being, especially like LLM AIs, right?
00:08:56.429 --> 00:08:58.830
The the stuff that we're using to do work.
00:09:00.029 --> 00:09:04.110
Um I mean, they are good at pattern recognition, right?
00:09:04.429 --> 00:09:05.309
Yeah, exactly.
00:09:05.549 --> 00:09:06.669
Um and probability.
00:09:06.909 --> 00:09:09.470
And yeah, there's a lot of applications out there.
00:09:10.190 --> 00:09:10.429
Yeah.
00:09:10.590 --> 00:09:14.269
Um, and there's also a lot of documentation, for instance, right?
00:09:14.429 --> 00:09:16.190
You talk about like medical research.
00:09:16.350 --> 00:09:20.429
There's a lot of medical research that's just never has been read, right?
00:09:21.070 --> 00:09:27.549
Hundreds of thousands of papers that have been created and five people have read, right?
00:09:27.629 --> 00:09:29.549
Yeah, ten people have peer-reviewed it.
00:09:29.629 --> 00:09:30.990
It's not a big thing.
00:09:31.149 --> 00:09:38.269
But if you put all this stuff together, um, you can potentially make links that you haven't made before, right?
00:09:38.509 --> 00:09:40.830
I think there is power there.
00:09:41.070 --> 00:09:52.029
Um, but at the end of the day, like I kind of do feel better about a human looking at me and and checking off that I'm fine, right?
00:09:52.190 --> 00:09:52.509
Yeah.
00:09:52.750 --> 00:09:55.309
Because a computer can say, hey, you're fine.
00:09:55.389 --> 00:09:57.389
And I'm like, you don't know that.
00:09:57.549 --> 00:09:57.950
Yeah.
00:09:58.110 --> 00:10:08.590
You know, like yeah, even if you believe that, I don't know if we're ever gonna get to a point where I a computer says, Well, yeah, this is definitely right, and I totally believe it, right?
00:10:08.669 --> 00:10:15.789
Yeah, but if a human stares me in the eyes and says, You're fine, you don't need to come back here for a year, I feel good about that.
00:10:15.950 --> 00:10:16.590
Yeah, yeah.
00:10:16.830 --> 00:10:19.230
In a way that I don't think a computer can.
00:10:19.629 --> 00:10:20.269
Yeah.
00:10:21.309 --> 00:10:21.870
Yeah.
00:10:22.190 --> 00:10:40.029
Yeah, no, and uh, you know, and that's and that's the other thing I was talking about with somebody the other day is how, you know, I see our industry going in one direction with using AI when there's a different direction we could have gone in that I wish we did to go in.
00:10:40.190 --> 00:10:45.389
And this is like a really good example of using AI poorly versus how we could use AI.
00:10:45.470 --> 00:10:52.110
So I talked about in the last episode uh hucks, which is uh auto-generating podcasts, right?
00:10:52.350 --> 00:11:04.750
And again, I I like I I do like using it because it allow it it will go find papers that I might not be able to sit down and read, give me a quick audio overview of it, and then if I can want to click into it, I can.
00:11:04.909 --> 00:11:10.350
But in and then, you know, we kind of tripped on like the idea that that's probably how things are gonna work in the future.
00:11:10.509 --> 00:11:18.190
Like there's going to probably be a kid's channel where every single episode of something is just auto-generated, tailored to the kid.
00:11:18.590 --> 00:11:32.110
And so, yes, like that is the generation it sounds like we're going in that uh direction, which is auto-generating content for people, which is sad because if we actually did it in a different way, right?
00:11:32.269 --> 00:11:33.070
Think about this.
00:11:33.230 --> 00:11:39.149
Like, and like I it's gonna sound like some companies do this, but let me explain a little bit here.
00:11:39.309 --> 00:11:41.950
What uh so just like this podcast thing, right?
00:11:42.029 --> 00:11:46.190
This is auto-generating podcasts based off a news that I want to listen to.
00:11:46.350 --> 00:11:55.870
But hypothetically speaking, there is probably a podcast that someone has recorded about the news that I want to listen to.
00:11:56.029 --> 00:11:58.750
It's just I can't find it easily, right?
00:11:59.070 --> 00:12:05.789
And yes, there are algorithms out there that that are supposed to suggest us content, but I think those are all broken.
00:12:05.870 --> 00:12:10.429
Those seem like they're the one whoever's paying the most is getting like suggested to me.
00:12:10.509 --> 00:12:12.830
Like Google search, right?
00:12:12.990 --> 00:12:19.789
You know that there's there's a weighted bit to the people who pay money, right?
00:12:19.950 --> 00:12:20.190
Yeah.
00:12:20.429 --> 00:12:24.590
So the first results you're gonna see are gonna be the people who paid money to be there, right?
00:12:24.909 --> 00:12:32.750
Or the people who at the very least have paid Google, have an account and have optimized how their entry is gonna come up.
00:12:32.830 --> 00:12:33.629
So it's gonna come up.
00:12:33.950 --> 00:12:37.629
Even if they haven't paid direct money, they're yeah, they're gaming the system, right?
00:12:37.950 --> 00:12:40.750
I mean, another good example is music, right?
00:12:40.909 --> 00:12:42.429
Like I pay for Spotify.
00:12:42.509 --> 00:12:58.669
I have Spotify suggest me music all the time, but it's all like the top 200 songs now, and then just recently, or not just recently, but as I've starting to get into records, I like want to buy records from people that are not really well known.
00:12:58.909 --> 00:13:02.909
I have I'm I'm finding myself having to peruse Bandcamp manually.
00:13:03.070 --> 00:13:16.029
I don't know if you've heard of that website, but like it just it it's it stinks because what I've discovered is there's so many great artists on there that I'm like, why hasn't any of their music ever been suggested to me?
00:13:16.269 --> 00:13:20.029
And you know, some of them aren't on Spotify, I get it, but some of them are.
00:13:20.190 --> 00:13:21.870
Why hasn't that been suggested?
00:13:22.110 --> 00:13:25.950
I've been listening to Spotify for years and they've been out for five years.
00:13:26.190 --> 00:13:28.509
So what what what where's the breakdown?
00:13:28.830 --> 00:13:32.110
But I do think that this is one of those places, right?
00:13:32.269 --> 00:13:34.590
If you're talking about medical research, right?
00:13:34.750 --> 00:13:44.830
Like there's there's probably like a hundred people in the world who are like super savants when it comes to medical research and can point you to the right thing, right?
00:13:45.549 --> 00:13:49.870
Um, it just there's not that many people who like I don't want to say care, right?
00:13:49.950 --> 00:13:56.669
But there's not enough people that like really have absorbed all the information um when it comes to medical research.
00:13:56.909 --> 00:13:58.429
But music, right?
00:13:58.669 --> 00:14:06.110
I feel like this is one of those places where like the answer was solved in like the 70s when they had record stores.
00:14:07.070 --> 00:14:14.669
No, no, like but you but you go to somebody and you say, Hey, I like Led Zeppelin, I like the Beatles, I kind of like to be able to do that.
00:14:14.830 --> 00:14:16.669
But that requires me to talk to someone.
00:14:16.750 --> 00:14:18.269
I don't want to go talk to some.
00:14:21.389 --> 00:14:27.470
But I like I feel like this is one of those places where it's like, oh, that's how people found stuff before.
00:14:27.629 --> 00:14:35.710
They yeah, they talked to people who didn't have all the information, but they had just enough information to like give you something to work with.
00:14:35.870 --> 00:14:41.230
You know I get I get I get to where where you're coming from that lets us get connected, we can meet people, blah blah blah.
00:14:41.549 --> 00:14:51.230
But my counter my my devil's advocate counter argument to you is the people that are going to be suggesting me that are still artists that know people.
00:14:51.629 --> 00:14:51.950
Uh-huh.
00:14:52.269 --> 00:15:06.029
And meanwhile, in that world where there might have been a great, like you know, Bob Marley or another hidden gem that just didn't know people, so no one ever found him.
00:15:06.190 --> 00:15:33.309
And so in this world where I'm talking about, we should use AI to make it more easy to discover people's content, and so that people who are making good content that don't really know a lot of people, they're more easily discovered because AI is able to like slurp up all this, know what my preferences are, and go, hey, because you like similar music to this, here's this random guy that only three people have listened to.
00:15:33.389 --> 00:15:35.789
You might actually like his music too, you know?
00:15:35.950 --> 00:15:42.669
And I feel like in that world, we can use AI to help us find human-generated content, right?
00:15:42.750 --> 00:15:44.509
Like human-made content.
00:15:44.750 --> 00:15:53.470
Um, and it make it actually easier for people to to who don't really know a lot of people to actually have like a faster starting.
00:15:53.629 --> 00:15:56.750
I don't know, like maybe it's a crazy idea, maybe it's impossible.
00:15:56.990 --> 00:16:10.029
But it's just like are we're going towards the direction of let's just hand generate all of this stuff when we could go in the direction of like leveraging it as a tool to connect us closer together, you know?
00:16:10.269 --> 00:16:10.750
Yeah.
00:16:11.070 --> 00:16:20.909
I I think like where I feel AI is a perfect use is when the thing it's doing is something that humans are bad at doing, right?
00:16:21.149 --> 00:16:32.110
Like when I think about code, I think like humans like we made code over time to be human readable, yeah, but it's not really human readable, right?
00:16:32.669 --> 00:16:46.990
It's still kind of messy, and you still, when you look at operations that happen in code, at the end of the day, like they're like bit instructions, they're like literal, literal like electrical signals between two pieces.
00:16:47.309 --> 00:16:53.149
It's you're you're trying to take all that and make it kind of human readable, and that's what you get with code.
00:16:53.389 --> 00:17:01.629
But like AI is good with that kind of stuff because it's kind of weird and it's kind of mechanical and whatever, right?
00:17:01.789 --> 00:17:10.029
I feel like the places where we don't want AI is the places where just the human experience is good, right?
00:17:10.269 --> 00:17:16.190
I you know, for instance, like I don't want to talk to an AI, I want to talk to you.
00:17:16.750 --> 00:17:21.710
It's this is something where like I don't think you can be replicated with an AI.
00:17:21.869 --> 00:17:22.829
Maybe I'm wrong.
00:17:23.710 --> 00:17:30.109
AI, when it matters, is for doing the things that humans don't like to do, right?
00:17:30.430 --> 00:17:33.230
Kind of like code at a really basic level, right?
00:17:33.390 --> 00:17:41.470
When when code gets rid of something that humans really don't like doing, yeah, um, then it's a good use of technology.
00:17:41.630 --> 00:17:49.069
But like there are things where like the face-to-face interaction, right, is part of the point, right?
00:17:49.549 --> 00:17:49.869
Yeah.
00:17:50.029 --> 00:17:53.390
Um, you know, I go back and forth like with restaurants.
00:17:53.549 --> 00:17:58.349
Do I want a touchpad that I can just order the food and the food comes out of the table with the robot?
00:17:58.670 --> 00:18:02.349
Is the point of the restaurant the food, or is it the atmosphere, right?
00:18:02.430 --> 00:18:10.509
How many restaurants have been where like the the waiter is actually a waitress is really good and it was kind of funny in talking to them?
00:18:10.750 --> 00:18:17.150
Um or like you talk to the people nearby at the next table, and like that's part of the experience.
00:18:17.309 --> 00:18:17.630
Yeah.
00:18:17.869 --> 00:18:29.950
So I I feel like sometimes like the tool aspect of AI um, you know, is very important, but we need to be kind of judicious about yeah, yeah.
00:18:30.190 --> 00:18:46.109
Like another good example, uh having AI auto-generate or auto-translate languages versus for us versus using AI to help us learn a language so that then we can communicate with them in the native language.
00:18:46.349 --> 00:18:46.589
Right.
00:18:46.750 --> 00:18:48.509
That's another that's another good example.
00:18:48.750 --> 00:18:55.869
One takes away the human experience, the other one enhances or embraces the the human experience, right?
00:18:56.029 --> 00:18:56.829
Yeah, yeah.
00:18:56.990 --> 00:19:06.990
I mean, uh learning is one of those things where like I found that it is easiest to learn for me when I'm in a social experience, right?
00:19:07.470 --> 00:19:10.589
Like a physical classroom is actually the best way.
00:19:10.750 --> 00:19:15.869
Now, am I gonna drive 30 minutes to do a physical class and pay a bunch of money?
00:19:16.029 --> 00:19:18.670
That's where it starts getting complicated, right?
00:19:19.150 --> 00:19:19.470
Exactly.
00:19:19.710 --> 00:19:23.630
Um, even though in some ways, like that is the best way for me to learn.
00:19:24.269 --> 00:19:28.589
Um and that's why I think a lot of places have taken shortcuts, right?
00:19:28.829 --> 00:19:30.349
Well, here's a perfect example.
00:19:30.509 --> 00:19:32.589
Here's uh here's a good here's a good idea.
00:19:32.750 --> 00:19:34.509
This is a fantastic idea, actually.
00:19:34.589 --> 00:19:35.230
I just thought about this.
00:19:35.470 --> 00:19:36.589
Oh, we'll see, we'll see.
00:19:36.829 --> 00:19:42.910
So you want to learn Japanese, and just like you said, the best way to learn Japanese is immerse yourself in it.
00:19:43.150 --> 00:19:53.950
So why don't we make hearing aids that get we put in your ear that that actually translate everything in from English to Japanese if you want to learn Japanese?
00:19:54.190 --> 00:19:58.269
So then you can't hear English anymore, and you only hear Japanese.
00:19:58.509 --> 00:19:58.750
Okay.
00:19:58.829 --> 00:20:07.630
So then it and then it translates your Japanese back to English for the people you want to talk to in English, and it forces you to learn that language.
00:20:07.869 --> 00:20:21.950
This is kind of like a funny Black Mirror episode, but I actually think that like that's one of the biggest things that's changed um in the past like 30 years, uh 40 years in Japan, too, right?
00:20:22.190 --> 00:20:32.910
40 years ago, when people would go travel abroad to a company country like Japan, if they wanted to talk to people from home, if they wanted to talk to people in English, they had to find a payphone, right?
00:20:33.150 --> 00:20:38.990
Pay exorbitant fees to have the payphone do this sort of international th thing.
00:20:39.150 --> 00:20:41.150
And otherwise they just didn't have English.
00:20:41.309 --> 00:20:51.789
Nowadays, tons of people move to a country like Japan and they don't learn the language because they just finish work teaching English and then they come home and they watch YouTube all day, right?
00:20:51.950 --> 00:20:52.349
Yeah, yeah.
00:20:52.509 --> 00:20:56.190
And and they don't really need to know the language to to survive, right?
00:20:56.349 --> 00:20:56.589
Yeah.
00:20:56.910 --> 00:21:04.190
I actually kind of like this idea though, where you have a hearing aid that just like makes your life living hell.
00:21:04.430 --> 00:21:05.470
Yeah, yeah.
00:21:06.269 --> 00:21:13.549
Um and and and doesn't let you understand what's happening until you understand the language.
00:21:13.710 --> 00:21:17.630
Because I feel like the best way to learn a language is through survival.
00:21:17.950 --> 00:21:18.269
Yeah.
00:21:18.430 --> 00:21:20.670
Um that's like pinnacle immersion.
00:21:20.990 --> 00:21:21.950
To your point, yeah.
00:21:22.029 --> 00:21:26.509
And like when you live in America, like the amount of Japanese immersion is like nothing.
00:21:26.589 --> 00:21:29.069
So you might be onto something.
00:21:29.390 --> 00:21:29.869
Yeah, yeah.
00:21:30.109 --> 00:21:30.430
All right.
00:21:30.589 --> 00:21:32.269
That's a new that's our new product drop.
00:21:34.109 --> 00:21:37.710
That that just make you hear in the language you want to learn.
00:21:38.269 --> 00:21:39.230
Uh yeah, yeah.
00:21:39.470 --> 00:21:40.670
I need to think about this.
00:21:40.910 --> 00:21:42.029
That that's cool, actually.
00:21:42.109 --> 00:21:42.990
I I kind of like it.
00:21:43.069 --> 00:21:43.710
I kind of like it.
00:21:44.109 --> 00:21:45.630
This isn't supposed to be a real product.
00:21:45.950 --> 00:21:47.309
This feels like the product.
00:21:47.470 --> 00:21:49.710
This feels like a product that YouTuber would make.
00:21:50.109 --> 00:21:52.109
The one that makes like useless products.
00:21:52.269 --> 00:21:53.309
I just want to play.
00:21:53.470 --> 00:21:54.349
He's like a YouTuber.
00:21:54.670 --> 00:21:57.150
I feel like that's that this would be the product he creates.
00:21:57.390 --> 00:21:58.029
I trust you.
00:21:58.190 --> 00:21:58.750
I trust you.
00:21:58.910 --> 00:22:02.509
But you know, like when you started, we talked about the hotel, right?
00:22:02.750 --> 00:22:03.069
Yeah.
00:22:03.230 --> 00:22:03.549
Yeah.
00:22:03.789 --> 00:22:08.829
And you want to get a person on the line because you know they can actually solve your that issue.
00:22:09.390 --> 00:22:09.549
Yeah.
00:22:09.869 --> 00:22:12.509
Um sometimes the the fact is there's a lot of people.
00:22:12.910 --> 00:22:15.390
People who like the eh could solve their issue, right?
00:22:15.789 --> 00:22:15.950
Yeah.
00:22:16.190 --> 00:22:24.029
And and they've just like I do want people to get over that idea that like they need to necessarily solve the thing with a human.
00:22:24.190 --> 00:22:24.589
Yeah.
00:22:24.910 --> 00:22:33.950
But so many times we've had the experience where like the reason you're calling is because you've literally exhausted all the other things, right?
00:22:34.109 --> 00:22:35.309
Like Yeah.
00:22:36.430 --> 00:22:41.869
But here's my thing, here's my argument as to why I think customer service should never be replaced.
00:22:43.630 --> 00:22:48.269
If you're calling customer service, it's because you're obsessed.
00:22:48.349 --> 00:22:48.509
Yeah.
00:22:48.670 --> 00:22:48.990
Right.
00:22:49.390 --> 00:22:53.789
And what we learned from the seven habits, how do you handle somebody who is obsessed?
00:22:53.950 --> 00:22:57.869
You need to understand their, you need to feel like they're understand.
00:22:58.190 --> 00:23:00.430
You need to give them emotional air, right?
00:23:00.509 --> 00:23:05.150
And they need to feel like they're they're they're they're their problems gonna be solved.
00:23:05.309 --> 00:23:07.309
And AI just can't do that, right?
00:23:07.470 --> 00:23:07.789
Yeah.
00:23:08.109 --> 00:23:12.990
It at least now, at least at this moment in time, AI can't understand how I feel.
00:23:13.150 --> 00:23:21.710
And so that's the problem, is I'm if I'm upset and I'm calling these companies with a problem, I want to talk to a person because I want to feel understood, you know.
00:23:21.869 --> 00:23:23.950
Otherwise, I wouldn't be calling customer service.
00:23:24.029 --> 00:23:32.750
And so, you know, it it does depend because yes, there probably is some situations where I'm calling customer service, and I probably will just want to talk to an AI because I just want my problem solved.
00:23:32.829 --> 00:23:34.029
I don't care how it's solved.
00:23:34.349 --> 00:23:37.150
Can can I can I tell you a story about this?
00:23:37.789 --> 00:23:43.150
I there was like a deal on um like new iPhones, right?
00:23:43.950 --> 00:23:48.190
And uh we were looking at getting like a new iPhone, right?
00:23:48.829 --> 00:23:59.869
And uh when we when we purchased this, we had this thing where it was like, oh, if you choose your provider, right, then you save 30 bucks.
00:24:00.190 --> 00:24:09.950
But the deal said in like small font that if you choose your provider, then this deal doesn't apply anymore because the providers have their own deals.
00:24:10.190 --> 00:24:14.430
So you you know, you have to go and like deal with this thing otherwise.
00:24:14.670 --> 00:24:18.269
We would have saved more money by not choosing the provider, right?
00:24:18.509 --> 00:24:25.309
But the way that the website defaulted to was that you had to choose or you could choose a provider, right?
00:24:25.549 --> 00:24:40.670
Anyways, the whole point was we finished this deal, we bought the thing, and then we realized, wait a second, we weren't getting the savings that we thought we would, because when we look online, it says that we're supposed to get like 70 bucks off, and we weren't getting 70 bucks off.
00:24:40.829 --> 00:24:45.630
So I called the place and I'm like, that's weird, we don't know anything about it.
00:24:45.789 --> 00:24:48.990
And then I talked to a person like, man, yeah, that sucks.
00:24:49.150 --> 00:24:54.670
Like, I think this is kind of stupid, but uh I can't I can't really help you.
00:24:54.990 --> 00:24:59.710
And honestly, just talking to the person and having the person agree that it was kind of stupid.
00:24:59.869 --> 00:25:00.269
Yeah, yeah.
00:25:00.430 --> 00:25:01.789
I was like, you know what, whatever.
00:25:01.869 --> 00:25:05.150
I just somebody talking to somebody and they agree it was stupid.
00:25:06.190 --> 00:25:06.589
Exactly.
00:25:06.750 --> 00:25:08.589
Sometimes you just want to feel understood.
00:25:08.910 --> 00:25:11.390
Like at the end of the day, I was like, whatever.
00:25:11.549 --> 00:25:14.029
Like, yeah, this was still a good purchase.
00:25:14.190 --> 00:25:21.869
Um, even if we didn't get the full benefit that we wanted to get, we probably lost like 40 bucks, like not a big deal.
00:25:22.190 --> 00:25:29.390
Um and and and and if that was an AI thing, being like, well, sounds like that sucks.
00:25:29.549 --> 00:25:32.349
Like, I don't think I would have felt heard, right?
00:25:32.589 --> 00:25:32.990
Exactly.
00:25:33.230 --> 00:25:33.950
Maybe I would have.
00:25:34.029 --> 00:25:34.509
Who knows?
00:25:34.990 --> 00:25:36.750
Maybe that was AI I was talking to.
00:25:37.869 --> 00:25:40.109
Well, maybe in the future you will, right?
00:25:40.190 --> 00:25:42.670
Maybe AI will understand you more than humans.
00:25:42.829 --> 00:25:44.589
But again, do we want that?
00:25:44.910 --> 00:25:47.470
Going back to taking away the human experience.
00:25:47.549 --> 00:25:48.269
I don't know.
00:25:48.589 --> 00:25:49.069
Yeah.
00:25:49.230 --> 00:25:49.549
Yeah.
00:25:49.710 --> 00:25:50.190
Yeah.
00:25:50.509 --> 00:25:50.910
All right.
00:25:50.990 --> 00:25:52.750
Well, that's the end of this episode.
00:25:52.910 --> 00:25:54.509
Hopefully, uh it recorded.
00:25:54.589 --> 00:25:55.789
We had a little blip in the middle.
00:25:56.190 --> 00:25:56.910
It was a good story.
00:25:57.150 --> 00:25:59.950
It was a it was it was a good talk either way, so I guess it's not wasted.
00:26:00.190 --> 00:26:00.430
Yeah.
00:26:00.589 --> 00:26:01.150
Yeah.
00:26:02.349 --> 00:26:03.710
Well, have a good night.
00:26:03.789 --> 00:26:07.069
And I hope I'm it's you I'm talking to and not some AI version of you.
00:26:07.230 --> 00:26:07.950
I am Jesse.
00:26:08.750 --> 00:26:09.069
Ah
