Latest / Humanity Working / Four Years with AI: Are We getting Smarter or Dumber?

Transcript

Four Years with AI: Are We getting Smarter or Dumber? · Humanity Working

• 0:15
  SPEAKER_02: Okay, Paul.
• 0:16
  SPEAKER_02: Four years with AI.
• 0:17
  SPEAKER_02: Are we getting smarter or dumber?
• 0:18
  SPEAKER_02: Sort of a follow-up to one of our recent podcasts and articles about this.
• 0:23
  SPEAKER_02: I found this one really, really fascinating read because it's almost like we accidentally figured out how to get better through a very how would we say brief study of prompt engineering.
• 0:41
  SPEAKER_01: This is yeah, this is a really interesting topic.
• 0:44
  SPEAKER_01: If folks want to read more about it, then they can head over to humanityworking.net and check out the article that you just mentioned.
• 0:51
  SPEAKER_01: We'll pop it in the show notes.
• 0:54
  SPEAKER_01: But yeah, the I became really, really fascinated by this based on actually some reading I was doing on X.
• 1:07
  SPEAKER_01: There was a question that came up a lot, I would say six months ago, a year ago, which is is AI making us dumber?
• 1:18
  SPEAKER_01: And I wrote about that at the time, we spoke about it at the time.
• 1:25
  SPEAKER_01: But now that question's being asked again a lot.
• 1:31
  SPEAKER_01: And so what I wanted to do is I wanted to take a look and see if the if you like if the answer is still the same, but also kind of like what we've learned in the last few few months about what it is what it is doing to us.
• 1:45
  SPEAKER_01: And so really, really interesting to come back to.
• 1:49
  SPEAKER_01: And we'll put a link to in the show notes for both the original article on this, and I think everything still holds in that original article, and we'll put a link into the latest work that we've done on this specifically.
• 2:06
  SPEAKER_02: Yeah, some some unintended consequences of of prompt engineering.
• 2:11
  SPEAKER_02: Yeah.
• 2:11
  SPEAKER_02: Uh if you started your correspondence course or your, I don't know, university study of prompt engineering, and then found yourself after a semester with no more classes to register for in that space.
• 2:24
  SPEAKER_02: Yeah, well, I think it was last week.
• 2:25
  SPEAKER_02: You've created a whole curriculum.
• 2:27
  SPEAKER_02: Yeah.
• 2:27
  SPEAKER_01: I think it was last week, wasn't it, that you uh that you were like, oh, remember when prompt engineering was a thing, right?
• 2:32
  SPEAKER_01: And it is amazing.
• 2:34
  SPEAKER_01: I mean, the whole era of AI, we're less than four years into this, you realize this generative AI thing.
• 2:40
  SPEAKER_01: And it's just bonkers to think.
• 2:44
  SPEAKER_01: Because when it first broke, that was November 30th, 2022, was when Chat GPT etched in my brain that date.
• 2:53
  SPEAKER_01: When ChatGPT 3.5 came out, which was, you know, for people to remember, that was sort of like the first version that was identifiably what we have today, but also kind of like the first version that the world got to see.
• 3:06
  SPEAKER_01: And that thing that we used then, it seemed like magic, right?
• 3:12
  SPEAKER_01: And now, when you look back on it, it feels primitive.
• 3:18
  SPEAKER_01: And what we can and now what increasingly people are doing with AI is so different that you would think any question about whether it's making us smarter or dumber, the answer to that question would be quite different.
• 3:32
  SPEAKER_01: And it turns out that it is quite different.
• 3:36
  SPEAKER_02: Yeah, I think I think the thing to be careful of is that it's easy to get sucked into it if you think that everyone is doing it or that it's mass appeal.
• 3:47
  SPEAKER_02: I think the numbers are still really tiny, though, that are using it beyond that simple chat window.
• 3:53
  SPEAKER_01: That is very true.
• 3:54
  SPEAKER_02: Other than by accident, because the Google search bar went away and is now an AI answer.
• 4:01
  SPEAKER_02: So you're you're getting answers and you're using AI more effectively almost by accident, even if you're not trying.
• 4:09
  SPEAKER_02: But as far as like going to the chat window, if somebody thinks about Chat GPT, most people it's exactly the same as it was four years ago.
• 4:18
  SPEAKER_02: It's just the models have shifted.
• 4:20
  SPEAKER_02: Maybe the answers are a little bit better.
• 4:22
  SPEAKER_02: But I think for the large majority of even somewhat daily or or casual users, they're not having these incredible experiences.
• 4:30
  SPEAKER_01: Well, well, there's a subtle argument against that though.
• 4:33
  SPEAKER_01: I would say there are differences.
• 4:35
  SPEAKER_01: There are differences when you think about what the effect it has on our brains is.
• 4:39
  SPEAKER_01: There are differences that are actually a little bit more important than we think they are.
• 4:43
  SPEAKER_01: So what you're describing, right, as is sort of like the simplest use case for interacting with AI, and I would absolutely agree that that hasn't really changed for the vast majority of people.
• 4:55
  SPEAKER_01: So what you're describing is you ask a question, you get an answer, right?
• 5:00
  SPEAKER_01: And the vast majority of people are using AI in that way.
• 5:06
  SPEAKER_01: But I think that there is a subtle shift that happens based on how good it is at doing that, right?
• 5:15
  SPEAKER_01: So back in the day when we first started using this thing and we got over the idea that it was magic, then we started to get a bit frustrated that it wasn't really answering the questions that we wanted to answer.
• 5:28
  SPEAKER_01: And so we had to get better at asking those questions.
• 5:31
  SPEAKER_01: That's the prompt engineering discipline, right?
• 5:33
  SPEAKER_01: Whether you formally did it, whether you formally did it or not, right?
• 5:37
  SPEAKER_01: And you might have gone on a course to fix that, or you might have done it through trial and error, but you'd eventually get better at asking at asking the questions.
• 5:47
  SPEAKER_01: Now it's so much better at asking at interpreting the what you want to say that you actually have to put a lot less effort into asking the right question.
• 6:02
  SPEAKER_01: So, in that sense, all of those skills that you were developing to ask the really, really good question, they're not being exercised in the same way that they were before.
• 6:12
  SPEAKER_01: On the flip side, and this came really to mind just yesterday when I was having a conversation with my wife about this, who just read the read the article.
• 6:24
  SPEAKER_01: People are now using this to enable so many different things that they would not have used it for before.
• 6:37
  SPEAKER_01: And so they're getting this sort of veneer of smarts that they would never have gone because they're not explored, they're now exploring things that they would never have explored before.
• 6:46
  SPEAKER_01: So in this case, my wife's applying for jobs right now.
• 6:50
  SPEAKER_01: She was going online to look at the information that was being sent by a potential employer as to how they used her information.
• 7:02
  SPEAKER_01: And she went, she took that information, she fed that to I think it was Copilot in this case, right?
• 7:10
  SPEAKER_01: And said, look at this and tell me what is actually going on at the employer.
• 7:16
  SPEAKER_01: Should I use AI to help create my, you know, to create my resume or not?
• 7:21
  SPEAKER_01: How are they actually viewing, based on what they're saying, how are they actually viewing my application?
• 7:28
  SPEAKER_01: What are they doing with it?
• 7:28
  SPEAKER_01: How are they using my data, etc.?
• 7:31
  SPEAKER_01: And it took her down this really targeted rabbit hole that 20 minutes later meant that she knew a whole bunch more about a hiring process than she would ever have known before.
• 7:47
  SPEAKER_01: And she sort of showed this to me.
• 7:50
  SPEAKER_01: She goes, Look, I'm learning, right?
• 7:51
  SPEAKER_01: Previously, I'd have just thrown my resume up there, but I'm actually learning.
• 7:54
  SPEAKER_01: I'm learning something.
• 7:56
  SPEAKER_01: And so there's this other interesting thing, and I think we'll get into it as we go deeper on this, whereby, even in the world of chat, the paths down which it's taking you are causing you to learn some more things than you than you would have done otherwise.
• 8:16
  SPEAKER_01: Against the backdrop of all of that other stuff that we talked about last time, against things like you know, cognitive laziness and cognitive offloading and stuff like that, and losing some of those skills.
• 8:27
  SPEAKER_01: It's doing that at the same time.
• 8:28
  SPEAKER_01: So this is where it gets really, really interesting.
• 8:31
  SPEAKER_01: And that's just in the world of chat.
• 8:33
  SPEAKER_01: So we lose stuff by virtue of the fact that we don't have to work as hard to prop to do our prompting, we gain stuff through these targeted rabbit holes that it takes us and the learning that we get from those rabbit holes.
• 8:49
  SPEAKER_02: Interesting.
• 8:50
  SPEAKER_02: We don't have to go onto this, but it strikes me a thought that that came up is that for people that are engaging, curious, yeah, actually trying to learn, I can see where it would make normal, everyday small talk even more annoying.
• 9:13
  unknown: Yeah.
• 9:14
  SPEAKER_02: Yeah.
• 9:15
  SPEAKER_02: Because you're used to a random conversation becoming deeper almost instantaneously now because you're you know, chatting with the Borg, chatting with the the all-knowing, right?
• 9:28
  SPEAKER_02: In a way.
• 9:29
  SPEAKER_02: And then you go to dinner and you're talking about the weather and you know doctor appointments, and you're just like, I would much rather be talking to ChatGPT right now.
• 9:41
  SPEAKER_01: I think that's fascinating.
• 9:43
  SPEAKER_01: Let's do a follow-up on that explicitly.
• 9:46
  SPEAKER_01: Because we've you we've obviously looked at what it's doing to soft skills and stuff in general, but I think specifically what it's doing to social skills, I think is a very, very fascinating topic of discussion.
• 9:59
  SPEAKER_02: Now you have to bring your AI into it during the conversation, right?
• 10:02
  SPEAKER_02: So everybody is chatting with each other and the AI.
• 10:05
  SPEAKER_02: It's it keeps going back to that.
• 10:07
  SPEAKER_01: Well, you've got your metaglasses whispering, whispering in your ears.
• 10:10
  SPEAKER_01: Here's how you pretend to be interested in this human.
• 10:14
  SPEAKER_02: Yeah.
• 10:14
  SPEAKER_02: Oh man, it's gonna get crazy quick.
• 10:16
  SPEAKER_02: The the interesting thing about AI learning is that we need time to absorb and learn and hone skills over time that we feel we've sort of earned.
• 10:28
  SPEAKER_02: Whereas AI doesn't really need all that time.
• 10:32
  SPEAKER_02: It has the answers immediately, it has the the full data set immediately.
• 10:36
  SPEAKER_02: Whereas we have to build that over time.
• 10:39
  SPEAKER_02: So, you know, we need it, but maybe AI doesn't need it.
• 10:44
  SPEAKER_02: And is that something that might collapse or or start to meet closer to, you know, do we start being better at it without the time?
• 10:54
  SPEAKER_01: Yeah, the timing thing, it's so, so interesting, Matt.
• 10:58
  SPEAKER_01: Because like I I think that if you spend any time with AI, one of the things you realize very quickly is AI doesn't really get time.
• 11:08
  SPEAKER_01: It's better at it than it was.
• 11:09
  SPEAKER_01: Amazingly, yeah, I've got Claude now saying, You've been working at this for a while.
• 11:13
  SPEAKER_01: Have you thought about going to bed?
• 11:14
  SPEAKER_01: I've actually had that on more than on more than one occasion now.
• 11:17
  SPEAKER_01: But it's still pretty ropey.
• 11:20
  SPEAKER_01: I mean, I think we've many of us have been in a situation where you say, I'm staying at this spot.
• 11:24
  SPEAKER_01: Where should I go visit?
• 11:25
  SPEAKER_01: And it goes, Oh, this is just a couple of minutes' walk away, and then you find out it's four and a half miles or something like that.
• 11:31
  SPEAKER_01: So it's it still struggles with these sort of like concepts around the physical world in which we live and the constraints of the physical world, and time is is one of those things because it works at the speed, you know, practically at the speed of light, and we don't.
• 11:45
  SPEAKER_01: But of course, we as humans actually aren't really good at understanding this either.
• 11:50
  SPEAKER_01: And what I mean by that is that as these shortcuts are being placed in our lives, we're undervaluing, I think, the amount of time it takes to build these capabilities, right?
• 12:06
  SPEAKER_01: To to embed things in our behaviors to develop wisdom, right?
• 12:14
  SPEAKER_01: Those things are not things that that happen overnight.
• 12:18
  SPEAKER_01: And there's stuff biologically that is happening in our brains in order for that to happen, and that stuff can't really be shortcut.
• 12:28
  SPEAKER_01: So a meaningful question for us is do we care about that?
• 12:34
  SPEAKER_01: Right?
• 12:35
  SPEAKER_01: If we don't care about that, then what we're basically saying is um human wisdom and the time that it takes to develop human wisdom is something that we can do without.
• 12:49
  SPEAKER_01: And I worry about that, right?
• 12:51
  SPEAKER_01: Because that's kind of a bull walk against some of the excesses of AI.
• 12:57
  SPEAKER_01: Interestingly, there's been a lot of stuff that has come out from senior scientists inside Open AI now that Astra has just been released.
• 13:12
  SPEAKER_01: And some of the stuff that they're talking about is based on their wisdom.
• 13:18
  SPEAKER_01: If they didn't have their wisdom, they wouldn't be talking about talking about this stuff.
• 13:22
  SPEAKER_01: And I'm not holding them up as paragons of virtue in any way, shape, or form.
• 13:25
  SPEAKER_01: But what I am saying is that some of the voices that are talking about what this could be and what we should be worried about, the things that they're talking about depend upon their depend upon their wisdom.
• 13:39
  SPEAKER_01: Can you get a facsimile of that through AI?
• 13:41
  SPEAKER_01: Sure.
• 13:42
  SPEAKER_01: I can you can you can get on you know Claude or Chat GPT today and ask it for you know things we should be worried about regarding AI's future, and it will give you some kind of like high-level facsimile of that.
• 13:55
  SPEAKER_01: But it is it it remains different.
• 13:59
  SPEAKER_01: And so wisdom can be intellectual wisdom, wisdom can be defined in sort of more aesthetics as well, you know, taste and artistic judgment and you know and things of that nature, right?
• 14:13
  SPEAKER_01: But those things take a lot of time to build up in human beings.
• 14:16
  SPEAKER_01: And if we're gonna be in any way equal partners as humans, we need that time to develop those things, and we need the opportunities to develop them.
• 14:23
  SPEAKER_01: And if we're constantly taking shortcuts with AI, we're not necessarily gonna get them.
• 14:30
  SPEAKER_02: That's the that's the worry, right?
• 14:31
  SPEAKER_02: Is that how many of the things are transferable skills that are instantaneously learned become wisdom versus things that require that time, that relearning, that you know, drilling it into us.
• 14:43
  SPEAKER_02: You know, I think about this a lot with how AI's capabilities to hold memory have increased over time.
• 14:52
  SPEAKER_02: Right.
• 14:53
  SPEAKER_02: Um where it was so frustrating to deal with AI because you're like, we've been over this, you know.
• 15:00
  SPEAKER_02: I told you this yesterday, and now I have to remind you again.
• 15:03
  SPEAKER_02: And it would come with those like ridiculous, oh, you're right.
• 15:06
  SPEAKER_02: We didn't say this yesterday.
• 15:07
  SPEAKER_02: It's like you know.
• 15:08
  SPEAKER_02: Uh well AI is always faking things.
• 15:12
  SPEAKER_01: And so occasionally you got this sort of like, uh, you know, look past the curtain to see where the faking is going on.
• 15:19
  SPEAKER_01: Yeah, I didn't actually read.
• 15:21
  SPEAKER_01: Yeah.
• 15:21
  SPEAKER_01: I know you gave me that thing and told me to read it, but I didn't.
• 15:24
  SPEAKER_01: You know, you you're right.
• 15:27
  SPEAKER_02: You're right to call me out on that.
• 15:29
  SPEAKER_02: One of my favorite bits on TV and like in in comedy right now is when people personify the AI and they have an interaction like a little skit, and the other person is pretending you're right about that.
• 15:42
  SPEAKER_02: I was thanks, thanks for calling me out.
• 15:45
  SPEAKER_02: You really should have, you know, that kind of stuff.
• 15:47
  SPEAKER_02: It's it it is so ridiculous.
• 15:49
  SPEAKER_02: But um you had a a note in the article that I th I found was really interesting, which was uh, you know, don't don't be so fast to dismiss prompt engineering because the the the whole thing is nuanced and hopeful.
• 16:04
  SPEAKER_02: And I I love that that phrasing of nuanced and hopeful because we're learning and developing these skills we didn't know we were gonna need.
• 16:14
  SPEAKER_02: We didn't realize maybe we would get good at simply because of the nature of how we're interacting with these tools, and all of them are transferable beyond interacting with the tool, which is really, really fascinating.
• 16:28
  SPEAKER_02: So maybe head into that territory a little bit.
• 16:31
  SPEAKER_01: Yeah, and I will say that this is this whole whole area, this whole area of what it's doing to our brains, is something where my personal view of it has evolved really in the last six months and has become more of a nuanced view.
• 16:51
  SPEAKER_01: I think if you'd have said to me four months ago, is AI kind of turning our brains into sludge, I would have I would have lent in that direction.
• 17:05
  SPEAKER_01: Um and I think it's because I probably undervalued the subtle ways in which it's actually uh giving actually helping improve our brains.
• 17:21
  SPEAKER_01: And it's very interesting because if you I I relate it to the to the topic of are you a are you a good driver, right?
• 17:32
  SPEAKER_01: Um so many studies have shown if you ask people, are you a better than average driver, about somewhere in the range of 80 to 90 percent of us believe that we're better than average drivers, which of course we can't all be, right?
• 17:46
  SPEAKER_01: Um and in the case of AI, if you ask the question, if you were to go on social media today and you were asking the question, does AI make people smarter or dumber?
• 18:01
  SPEAKER_01: People will separate out their answer.
• 18:03
  SPEAKER_01: They will generally believe that it's making them personally smarter and it's making everybody else dumber.
• 18:10
  SPEAKER_01: Right now, again, those two things probably can't probably can't be true at the same time.
• 18:16
  SPEAKER_01: And I was leaning towards the idea of, you know, it's it's exactly like that, that driver situation whereby maybe it is making us all dumber, but because we need to protect our own egos, we assign that to other people versus to ourselves.
• 18:33
  SPEAKER_01: But there is an alternative reading of this, right?
• 18:35
  SPEAKER_01: The alternative reading is that it's not actually making us dumber, but we're seeing certain indicators that it's making us dumber in certain ways, and we're attributing that to everybody else, but we're not feeling that we're not feeling that ourselves.
• 18:50
  SPEAKER_01: Where I've where I've come to on this map is that I think that it is making us dumber in certain ways and smarter in certain ways, but at the same time.
• 19:12
  SPEAKER_01: First thing.
• 19:12
  SPEAKER_01: Second thing, that's not new, right?
• 19:16
  SPEAKER_01: We're much worse at mental arithmetic than than than people were 50, 60 years ago because we got a thing called a calculator, right?
• 19:24
  SPEAKER_01: But we're smarter in many, many other ways, right?
• 19:26
  SPEAKER_01: We're just not as good at mental arithmetic.
• 19:29
  SPEAKER_01: So that's not new.
• 19:31
  SPEAKER_01: That's a fairly, you know, a fairly standard thing.
• 19:35
  SPEAKER_01: But then the third thing is that there is definitely a difference between people that are on the bleeding edge of how they're using AI and people that are still using it in that sort of just conversational, I'm having a chat.
• 19:53
  SPEAKER_01: And you're right to say that 98, 99% of people are using it to have a chat, right?
• 20:00
  SPEAKER_01: And about 1% of people, maybe those numbers aren't exactly right, but it's probably close, right?
• 20:05
  SPEAKER_01: About 1% of people are using it to actively build things, and those people are stretching their brains like crazy as they're using it to build things.
• 20:20
  SPEAKER_01: And the re and the reason is the same as the reason was back in 2022 when this thing launched, right?
• 20:27
  SPEAKER_01: If you're using it to build things, then you as a human are having to step in and and you know modify and you're having to iterate like crazy, and you're like, oh man, I you're building it, but you're not quite there, and you need to be able to do this.
• 20:41
  SPEAKER_01: So now all of a sudden you're building another set of skills that allow you to use it at the bleeding edge.
• 20:48
  SPEAKER_01: So that's really interesting.
• 20:51
  SPEAKER_01: You've got this group of people that are getting better in certain things, worse at certain other things, then this group of pioneers that are getting super smart by virtue of this vibe coding explosion that they're that they're living through, if you like.
• 21:08
  SPEAKER_02: Yeah, it's it's like we're you got a handful of people playing GTA 6, and there's most of the world is still playing Pong if they're games at all.
• 21:18
  SPEAKER_01: I think that's a perfect analogy, yeah.
• 21:20
  SPEAKER_02: You know, and you're you it's it's weird to to watch that occur.
• 21:26
  SPEAKER_02: And I think we we've had this.
• 21:28
  SPEAKER_02: Maybe that's always happened, maybe that's always happened with tech, but this one feels so much different than than ever before.
• 21:36
  SPEAKER_02: Maybe just again to go back to time, maybe because time is collapsing around the advancement so quickly.
• 21:42
  SPEAKER_02: Like we're not imagining that we're talking about a four year horizon here.
• 21:48
  SPEAKER_02: A four year span that's not only launched something we've never really seen before, but also advancing in a way that we've never really seen.
• 22:00
  SPEAKER_02: before uptake like we've never really seen before in an app.
• 22:04
  SPEAKER_02: ChatGPT was the biggest app download, fastest app download, right?
• 22:09
  SPEAKER_02: When it when it launched.
• 22:10
  SPEAKER_01: So there is there is a challenge here though, Matt, and we're going to be exploring this and what I'm about to say in a lot more detail in upcoming podcasts and then also on the Humanity Working website.
• 22:23
  SPEAKER_01: I think you can divide generative AI into like two eras if you like.
• 22:29
  SPEAKER_01: So the first era was just the chat era, right?
• 22:32
  SPEAKER_01: It's kind of ironic we're still calling it chat GPT because we're moving beyond chat in many, many, many different ways.
• 22:38
  SPEAKER_01: But anyway, there was the chat era and the chat era was so inclusive, I guess you could call it.
• 22:49
  SPEAKER_01: And so the main the the reason for the uptake the reason for the massive uptake was the fact that we've been struggling to use the chat interface with things like Siri and Alexa and so on.
• 23:06
  SPEAKER_01: And it was very dispiriting because they were really rubbish and you and so people used it as a kind of play thing and then they most people just largely moved beyond it.
• 23:16
  SPEAKER_01: And then suddenly this thing explodes and you could literally hand it to your grandmother and say have a chat right and there was no learning curve.
• 23:26
  SPEAKER_01: There was nothing right I mean and then you know over time we realized there's a difference between using it and using it well and that's when the prompt engineering thing comes in and all that kind of stuff.
• 23:35
  SPEAKER_01: But the reason for that massive uptake was just exactly that right you didn't have to learn Excel, you didn't have to learn you know databases or anything along those lines it was just literally chat and see what happens the next wave the wave that began really when OpenClaw became a thing and maybe a little bit before that but for many many people in our in our heads it sort of began there.
• 24:03
  SPEAKER_01: What some people are calling the agentic era but I call it the builder era if you like is is the okay I've got this thing now we're going to build things with this with this capability that is much more exclusive.
• 24:20
  SPEAKER_01: You can't hand that to your grandmother and say hey guess what you can build anything you like using this right now that's like almost like the software equivalent of I don't know the Raspberry Pi or something like that.
• 24:33
  SPEAKER_01: It is a it's a it's a different thing at least right now.
• 24:39
  SPEAKER_01: It's open to anybody who's smart and inquisitive and likes building things and so on but it's not literally sort of a no-brainer thing.
• 24:49
  SPEAKER_01: And so that to me is the interesting thing here it's like the uptake from the general public the consumer uptake of using this thing as the thing that helps you build things is likely to be much slower because it depends upon our the imagination of what we want to build, the commitment to sort of take it from A to Z.
• 25:13
  SPEAKER_01: And then it depends upon some sort of capabilities that some of us natural have had through our experiences sort of engineering principles, architecture principles and so on, and some that we don't right.
• 25:24
  SPEAKER_01: And so I think that will be interesting.
• 25:26
  SPEAKER_01: It'll be interesting to see how quick the uptake is how quickly is that 1% going to become 10% going to become 50% of people that use it.
• 25:35
  SPEAKER_02: Yeah and then I think just forecasting out from there is, you know, we're talking about building things in quotes building software tools building apps etc the physical but it's not building physical things yet.
• 25:49
  SPEAKER_02: So I think it's going to be really neat when it can start manifesting into the real world at that speed whether it's buildings physical things whatever fabrication right so you know you see the culmination of of digital printing 3D printing AI all coming together to fashion things into the physical world it starts to get really really crazy.
• 26:15
  SPEAKER_02: But yeah it's it's good.
• 26:16
  SPEAKER_02: I think I think my takeaway from that this this one is that it's not making us dumber necessarily but it's doing different things than we would have anticipated right I think of like I can't remember I don't know anyone's phone number really anymore.
• 26:34
  SPEAKER_02: And that was a that was a primary use case of of memory back in the in in the 80s for example you just needed to use your br your brain to remember phone numbers and that's just you don't have to do that just like the calculator you don't have to learn how to do that anymore if you don't want to as we discussed in that I don't feel dumber as a result.
• 26:54
  SPEAKER_01: Well that's the point I think as we discussed in the in the previous episode about uh on this topic we will lose things we are losing things there are capabilities that we're losing to me this is where the most that most important skill of metacognition thinking about thinking is is one of the most important if not the most important skill that we need to work on maintaining because there are going to be certain skills that if we lose them it doesn't matter.
• 27:31
  SPEAKER_01: The skill of being able to remember like I mean to get there's a broader skill than remembering people's phone numbers but for simplicity we'll just call it that right the skill of being able to remember other people's phone numbers is a skill that we don't need, right?
• 27:43
  SPEAKER_01: But there are going to be other things that if we lose them then we might be regretting that in five years' time or 10 years time.
• 27:54
  SPEAKER_01: And it isn't it remains I think an unanswered question as to how one goes about building wisdom in the era of AI.
• 28:05
  SPEAKER_01: I think we need it.
• 28:06
  SPEAKER_01: I think we know we need it and I think how we go about making sure that the 20 year olds that are using AI are developing wisdom in this of the developing deep wisdom that is going to be needed 10 years time, 20 years time, 30 years time is something we need to to think about really deeply and not in the young people today type way.
• 28:35
  SPEAKER_01: It's not their fault we've created this world in which they exist let's really think about solutions to that question.
• 28:46
  SPEAKER_02: If we can have a solution to how do we ensure that the next generation and the generation after that and the generation after that develop meaningful useful wisdom we're gonna be okay yeah yeah certainly hope so all right Paul another great episode look forward to learning more about this as we go and figuring out what you learn.
• 29:09
  SPEAKER_00: Just Matt Humanity Working is produced by Matt Neal.
• 29:16
  SPEAKER_00: If you liked this episode please take a moment to rate and review us on your favorite podcast platform.
• 29:22
  SPEAKER_00: It makes a huge difference in helping others find us and don't forget you can find all episodes and get our weekly newsletter at humanityworking.
• 29:32
  SPEAKER_00: See you next time