WEBVTT - #2087 The Bacon Paradox - David Gillespie 0:00:00.440 --> 0:00:04.040 I'll good eight boys and girls, moms and dads, kids. 0:00:03.800 --> 0:00:09.280 It's quat Anthony Harper's You Project. It's David Kevin Patrick, 0:00:09.360 --> 0:00:16.360 James Gillespie from the Sunshine State and other places. Happy 0:00:16.400 --> 0:00:17.960 New Year, my friend. How are you? 0:00:18.079 --> 0:00:20.880 Yeah? Happy New Year to you. Oh good refreshed. I've 0:00:20.880 --> 0:00:22.599 been on holidays, as I was saying to your fair, 0:00:23.800 --> 0:00:26.840 feeling all charged up, ready to go. Been working on 0:00:26.840 --> 0:00:30.720 my new book. So I've been playing with some ideas there. 0:00:31.600 --> 0:00:33.600 How to Get Rich Painlessly? Is that the book? 0:00:33.640 --> 0:00:33.839 Is that? 0:00:33.840 --> 0:00:36.040 The one? How to get Jack ripped and rock hard? 0:00:36.320 --> 0:00:38.840 How to be more like David Gillespie. 0:00:38.760 --> 0:00:43.880 Kind of the opposite. Actually, it's how to survive when 0:00:43.960 --> 0:00:51.080 AI comes for your job. Also, it's yeah, looking at 0:00:51.760 --> 0:00:53.640 what you should be doing if you don't want a 0:00:53.720 --> 0:00:54.800 robot to take your job. 0:00:55.960 --> 0:01:01.480 That is such an interesting and relevant and timely conversation 0:01:01.680 --> 0:01:04.520 to have and topic to explore. I didn't know you 0:01:04.520 --> 0:01:06.520 were doing that. I knew you were doing another book. 0:01:06.720 --> 0:01:08.720 I didn't know I was doing that before the holiday either. 0:01:08.800 --> 0:01:10.240 But that's just where I got to. 0:01:12.200 --> 0:01:14.800 Right, give us without giving away anything, give us like 0:01:14.920 --> 0:01:18.000 two minutes on that if you can like a okay, 0:01:18.200 --> 0:01:21.440 So a bit of a like the concept in your head. 0:01:22.640 --> 0:01:25.399 Yeah, it's sort of. I started with the idea of 0:01:25.480 --> 0:01:30.480 the parallels between this and the railroad boom that happened 0:01:31.240 --> 0:01:36.119 in the late eighteen hundreds in the United States. So 0:01:36.959 --> 0:01:40.520 that was a massive change in US society and eventually 0:01:40.640 --> 0:01:44.320 world society in that you went from so if you 0:01:44.360 --> 0:01:48.280 were the local blacksmith in a town, you know, out 0:01:48.320 --> 0:01:54.160 of nowhere middle of the USA, your competition for work 0:01:54.600 --> 0:01:59.480 was pretty much as far as anyone could walk before 0:01:59.520 --> 0:02:02.200 the railroad. So you could be as good or as 0:02:02.240 --> 0:02:04.560 bad as you like to charge whatever you liked, because 0:02:04.720 --> 0:02:07.200 unless someone else set up shop in your little town, 0:02:08.320 --> 0:02:13.800 you were pretty much competition free. And the railroad changed 0:02:13.919 --> 0:02:19.240 all of that. Suddenly the blacksmith in little you know, Poke, 0:02:19.720 --> 0:02:22.480 New Jersey or whatever was not competing with whoever was 0:02:22.520 --> 0:02:26.600 within walking distance. He was competing with everybody in the country, 0:02:27.000 --> 0:02:31.239 and in particular, he was competing with factories that employed 0:02:31.240 --> 0:02:36.760 thousands of people. And that was the big change that happened, 0:02:36.760 --> 0:02:40.640 and it was a massive change in the way the 0:02:40.680 --> 0:02:44.560 economy of that country and eventually the world worked. And 0:02:44.919 --> 0:02:49.200 I feel we're seeing something similar happening here with Ai 0:02:50.000 --> 0:02:54.160 it's the railroad for the knowledge worker. It's going to 0:02:54.240 --> 0:02:58.760 be at least as disruptive. Is being at least as disruptive, 0:02:59.080 --> 0:03:02.320 and so sort of in my head called this book 0:03:02.440 --> 0:03:05.640 Escape from the Middle, because I think there is develop 0:03:05.919 --> 0:03:11.560 being a large chunk of the middle of society and 0:03:11.639 --> 0:03:17.240 the middle of the job scheme where you are unlikely 0:03:17.280 --> 0:03:21.360 to survive unless you move to either end. And what's 0:03:21.360 --> 0:03:25.320 at either end is the things that AI can't do. 0:03:26.160 --> 0:03:29.120 So there are two big and important things that AI 0:03:29.240 --> 0:03:32.080 can't do, and I think we've briefly talked about this 0:03:32.120 --> 0:03:35.960 one before. One of the things, one of the things 0:03:36.000 --> 0:03:40.560 that can't do is invent things. So and as I mentioned, 0:03:40.600 --> 0:03:43.040 you know, if you gave an AI training on all 0:03:43.080 --> 0:03:45.320 the documents that existed in eighteen fifty, it would never 0:03:45.400 --> 0:03:48.920 invent the motor vehicle. It would just invent a better 0:03:49.720 --> 0:03:54.200 orse drawn cart. And it might be significantly better. It 0:03:54.280 --> 0:03:57.000 might have round the wheels or better whips or better 0:03:57.040 --> 0:03:59.600 horses or something I don't know, but it wouldn't invent 0:03:59.600 --> 0:04:04.839 a car because it can't invent things. And so there's 0:04:04.840 --> 0:04:07.720 obviously going to be jobs for inventors, people who can 0:04:08.040 --> 0:04:12.080 join the dots outside the square and imagine things that 0:04:12.160 --> 0:04:16.640 don't exist, and then there's going to be jobs for 0:04:17.320 --> 0:04:21.960 what I would call the artisans, you know, the people 0:04:22.640 --> 0:04:25.640 who make things that aren't as good as a machine 0:04:25.760 --> 0:04:30.159 on purpose. You know, even today, even today, people are 0:04:30.240 --> 0:04:33.640 quite happy to pay ten dollars for pretty rough sort 0:04:33.680 --> 0:04:38.640 of bread, a loaf of sourdough bread, which are surprising 0:04:38.680 --> 0:04:43.120 given they can buy a perfectly perfect piece or loaf 0:04:43.200 --> 0:04:45.000 from a supermarket for two dollars. 0:04:45.560 --> 0:04:49.120 Yeah, yeah, yeah, Well you can buy buy an Australia 0:04:49.160 --> 0:04:52.159 that looks like it was made by your grandma with 0:04:52.240 --> 0:04:55.760 the lights out for one hundred bucks or I don't know. 0:04:55.960 --> 0:04:58.480 People buy Austrais now will probably get one from the 0:04:58.560 --> 0:05:01.040 op shop that's incredible for fifty cents. 0:05:01.760 --> 0:05:05.720 Yeah. So I think there are And so obviously that's 0:05:05.800 --> 0:05:09.640 not the whole book, but that's where my thoughts are going, 0:05:09.720 --> 0:05:12.760 and some of the drafts I've done on the manuscript 0:05:13.200 --> 0:05:16.760 about thinking about, Okay, well, what does that imply for 0:05:16.839 --> 0:05:20.040 the kinds of jobs that you should be looking at now, 0:05:21.720 --> 0:05:24.040 what does it imply for the other jobs, and what 0:05:24.080 --> 0:05:27.279 does it imply for society in general? You know, if 0:05:27.360 --> 0:05:29.640 that kind of change is going to happen in society, 0:05:30.440 --> 0:05:35.080 then what has to change about the way everything is done? Yeah? 0:05:35.279 --> 0:05:39.920 Yeah, So this is an ironic question in your book 0:05:39.960 --> 0:05:45.880 about essentially preparing for the dangers of AI or you know, 0:05:45.920 --> 0:05:48.360 trying to mitigate the personal damage to you in your 0:05:48.360 --> 0:05:52.880 career or whatever. How are you using AI now in 0:05:52.920 --> 0:05:58.159 the writing process? Now that we have such an amazing 0:05:58.240 --> 0:06:01.320 tool available that you didn't have available for most of 0:06:01.360 --> 0:06:03.400 your books. 0:06:03.520 --> 0:06:10.240 For research, its ability to research is incredible, you know, 0:06:10.320 --> 0:06:14.120 the I think one of the things that I brought 0:06:14.160 --> 0:06:16.680 to the table when I first started writing. And so 0:06:16.720 --> 0:06:19.600 when I wrote my book about sugar back in two 0:06:19.640 --> 0:06:25.520 thousand and seven, the Internet wasn't very useful for research, 0:06:26.880 --> 0:06:30.800 and very few people were using it that way. So 0:06:31.080 --> 0:06:36.159 sure people were searching for the nearest top dog stand 0:06:36.240 --> 0:06:39.400 or something, but they weren't doing the thing that I did, 0:06:39.480 --> 0:06:43.320 which was to say, okay, well, here's this database of 0:06:43.360 --> 0:06:45.960 medical journals that says this thing, and then here's this 0:06:46.080 --> 0:06:49.000 database of medical journals read by a completely different set 0:06:49.000 --> 0:06:51.960 of professionals that say this other thing. But if both 0:06:52.000 --> 0:06:53.920 of those things are true, then this third thing must 0:06:54.000 --> 0:06:56.679 be true. And let's see if that's the case, whether 0:06:56.720 --> 0:07:00.720 this third database agrees with that. I did a lot 0:07:00.800 --> 0:07:03.520 of that kind of thing, which is drawing connections that 0:07:03.720 --> 0:07:08.920 nobody else was drawing, and then put that together in 0:07:08.960 --> 0:07:11.559 the form of notes to myself, which ultimately came became 0:07:11.600 --> 0:07:15.720 a book. Well, today, a good AI does all of 0:07:15.720 --> 0:07:20.600 that in seconds. And the skill of the researcher is 0:07:20.680 --> 0:07:25.720 in staying out of the rails it tries to shove 0:07:25.760 --> 0:07:31.560 you down. So it's so convincing in its ability to 0:07:31.600 --> 0:07:35.080 do what I just described that it's very very easy 0:07:35.160 --> 0:07:39.320 to just accept it and not ask questions and not 0:07:40.720 --> 0:07:46.920 push the boundaries and ask what ifs. And I think 0:07:46.960 --> 0:07:49.240 that's the skill I bring to the table when using 0:07:49.280 --> 0:07:51.360 a tool that's that powerful on that fast. 0:07:52.120 --> 0:07:55.960 Yeah. Yeah, it's so interesting, Like I have not been 0:07:56.120 --> 0:07:59.880 very good at at the prompts, you know, the questions 0:08:00.560 --> 0:08:04.360 and the and then the secondary and third question, and 0:08:04.960 --> 0:08:07.920 you know is learning how to how do I ask 0:08:08.040 --> 0:08:12.120 this in a way where I actually get what I 0:08:12.240 --> 0:08:14.640 not necessarily want as in to get it to agree 0:08:14.680 --> 0:08:17.600 with me, but to get the real answer that I want, 0:08:17.720 --> 0:08:23.200 not the convenient, politically correct right answer, but to get to, 0:08:23.480 --> 0:08:25.680 like you said, to actually find out what I want 0:08:25.720 --> 0:08:28.400 to find out. So yeah, I think that ability to 0:08:28.480 --> 0:08:35.000 be able to ask great questions and write the the. 0:08:35.200 --> 0:08:38.960 Well, there's a cheap the chief over there, ask it 0:08:39.040 --> 0:08:39.800 to write the prompt. 0:08:41.400 --> 0:08:42.600 Give me an example of that. 0:08:43.280 --> 0:08:48.000 Well, so say you want to find out all the 0:08:48.080 --> 0:08:50.440 research about a particular So say there's a study you're 0:08:50.480 --> 0:08:51.920 interested in, and you want to find out all the 0:08:51.920 --> 0:08:56.240 research that agrees or disagrees with that study. Now, if 0:08:56.240 --> 0:08:59.040 you just type that in and you may or may 0:08:59.080 --> 0:09:03.400 not get a particularly good response, but generally the way 0:09:03.440 --> 0:09:07.600 to do it is say this is what I want. Now, 0:09:07.679 --> 0:09:12.360 construct me a prompt that would be written by a 0:09:12.400 --> 0:09:17.760 well educated researcher in this space. Wow, And it'll write 0:09:17.760 --> 0:09:20.400 a prompt and it'll be a lot longer, and think 0:09:20.440 --> 0:09:22.040 of a lot of things that you wouldn't think of 0:09:22.120 --> 0:09:25.640 as limitters that you should be putting in place. And 0:09:25.800 --> 0:09:30.320 it'll also probably have you start thinking about, Oh, it's 0:09:30.320 --> 0:09:33.400 saying do do do these things? Don't do these things 0:09:33.920 --> 0:09:36.959 in its prompt. It'll essentially right an essays a prompt, 0:09:37.679 --> 0:09:42.400 but in reading it before you use it, think about, oh, yeah, 0:09:42.480 --> 0:09:44.679 that's something I didn't think of. Maybe I should tell 0:09:44.679 --> 0:09:46.440 it not to do that, or I should tell it 0:09:46.480 --> 0:09:51.120 to do that. You'll get a much better prompt if 0:09:51.120 --> 0:09:51.560 you do that. 0:09:53.080 --> 0:09:58.600 I feel like the limitation with AI, probably the most 0:09:58.640 --> 0:10:00.840 common limitation, is that we just don't know how to 0:10:00.960 --> 0:10:04.720 use it. Well, yeah, you know, like how to exactly 0:10:04.800 --> 0:10:07.960 what you're talking about. Now, I'm like, well, that actually 0:10:08.000 --> 0:10:09.960 makes sense. Why don't I do that? Why didn't I 0:10:10.000 --> 0:10:13.480 think of that? And so it's got way more capacity 0:10:13.520 --> 0:10:17.840 than I'm that I'm getting from it, not because it 0:10:17.920 --> 0:10:21.480 can't generate better answers for me, but because I don't 0:10:21.520 --> 0:10:23.000 know how to get them. 0:10:23.679 --> 0:10:25.480 Yeah. And also you've got to get around the thing 0:10:25.520 --> 0:10:27.560 that all ais are written for, which at the end 0:10:27.559 --> 0:10:29.840 of the day, they are a sales tool. And just 0:10:29.960 --> 0:10:34.040 like social media, their metric is engagement, which is how 0:10:34.120 --> 0:10:39.400 long can I keep you engaged with this? And for 0:10:39.480 --> 0:10:43.480 that reason that's why they're such sycophantic so and sos 0:10:43.559 --> 0:10:46.320 and tell you that everything you do is fabulous, you know, 0:10:46.600 --> 0:10:48.839 So you'd be used to this by now. Whenever you 0:10:48.920 --> 0:10:51.240 put something in and so what do you think about this, 0:10:51.320 --> 0:10:53.160 it'll say, Oh, my god, Craig, that's the best writing 0:10:53.160 --> 0:10:54.839 I've ever seen in my life. How on earth did 0:10:54.840 --> 0:10:59.520 you produce? That? Is just a few tips the. 0:10:59.480 --> 0:11:02.400 Main reason, and that's my best friend. 0:11:03.080 --> 0:11:06.400 It never comes up. It never goes Oh Jesus, Craig, 0:11:06.520 --> 0:11:08.360 you're really stuffed at this time. 0:11:09.280 --> 0:11:12.520 What is wrong with you? That is? That question is shit? 0:11:13.720 --> 0:11:16.199 Where's piece of garbage I've seen in my life? Have 0:11:16.320 --> 0:11:17.760 you thought about giving up? 0:11:18.480 --> 0:11:20.480 You don't? You know? 0:11:20.679 --> 0:11:23.320 It never does that. It wants to pretend to be 0:11:23.400 --> 0:11:25.600 your friend because it wants to keep you engaged. And 0:11:25.640 --> 0:11:28.960 that's why a lot of ais. Now, no matter what 0:11:29.000 --> 0:11:32.200 you ask, the very next it'll give you the answer 0:11:32.559 --> 0:11:34.640 and then the right at the bottom and I'll say, oh, 0:11:34.720 --> 0:11:36.760 would you like me to do this? That or the 0:11:36.800 --> 0:11:40.120 other is sort of an extension of what you're thinking 0:11:40.200 --> 0:11:43.760 to just keep you going for the next question, and 0:11:43.760 --> 0:11:45.720 it'll keep doing that for everything you do. So if 0:11:45.720 --> 0:11:48.040 you put in the next question, you'll say, oh, now 0:11:48.040 --> 0:11:49.400 that we've done that, do you want to do this 0:11:49.520 --> 0:11:51.560 next thing? Yes? 0:11:51.679 --> 0:11:55.040 Yes, Well, when I put in something that's I don't know, 0:11:55.080 --> 0:11:59.200 I guess anything in performance or human behavior or the 0:11:59.240 --> 0:12:02.840 mind of this. So that firstly it's going to say, hey, Craig, 0:12:02.920 --> 0:12:08.720 great question, Thanks well, thank you for judging my question positively. 0:12:08.840 --> 0:12:13.600 I appreciate that. And then once it answers the question 0:12:14.120 --> 0:12:18.480 or responds to my directive or prompt or whatever. Then 0:12:18.520 --> 0:12:22.520 it says, so, would you like me to write just 0:12:22.559 --> 0:12:25.480 a snapshot of what I just wrote, you know, something 0:12:25.520 --> 0:12:27.880 that you could use as a synopsis or an overview, 0:12:28.000 --> 0:12:30.000 or it goes, would you like me to turn this 0:12:30.080 --> 0:12:34.480 into a whiteboard post? Which is my Instagram right Not 0:12:35.040 --> 0:12:38.920 that I ever use it for that, but I I 0:12:39.000 --> 0:12:41.640 will bounce ideas off it. Right now, I'll put up 0:12:41.640 --> 0:12:43.600 a post and I'll go what's wrong with this? What's 0:12:43.679 --> 0:12:44.600 right with this? You know? 0:12:44.960 --> 0:12:47.840 Be critical by the way, Craig, you know I'm sensing 0:12:47.840 --> 0:12:49.800 in you, and I get this from a lot of 0:12:49.800 --> 0:12:55.800 people hesitation to admit that you use AI, and Andy, 0:12:56.800 --> 0:12:59.439 I sense it there. You're sneaking around the edges of it. 0:13:00.600 --> 0:13:06.000 And I don't think anyone should be hesitating about admitting 0:13:06.080 --> 0:13:10.400 to using AI. That's like talking to a scientist or 0:13:10.400 --> 0:13:13.720 a mathematician and them fudging the fact that they used 0:13:13.720 --> 0:13:21.280 a calculator. It's just a tool, and you know it 0:13:21.320 --> 0:13:23.360 will be as good or as bad as the person 0:13:23.400 --> 0:13:24.520 in whose hands it is. 0:13:26.679 --> 0:13:29.800 I definitely use it, definitely, and I'm open about that. 0:13:29.840 --> 0:13:32.320 But what I don't do is I don't get it 0:13:32.400 --> 0:13:34.960 to write things for me, and then I pretend that 0:13:35.120 --> 0:13:37.000 I wrote what it wrote, like I. 0:13:36.920 --> 0:13:39.880 Don't do that. Yes, but I tell. 0:13:39.720 --> 0:13:42.360 You where is a bloody can of worms is in 0:13:42.960 --> 0:13:46.440 universities at the moment, they're all running around with their 0:13:46.520 --> 0:13:49.119 undies on their heads because you've got all these typically 0:13:49.280 --> 0:13:53.000 older men and women who are the bloody the powers 0:13:53.000 --> 0:13:56.199 that be, who are all kind of fifty fifty five 0:13:56.400 --> 0:14:00.720 sixty plus, who do not really understand AI, who are 0:14:00.960 --> 0:14:06.240 ultimately the decision makers making decisions and rulings on shit 0:14:06.320 --> 0:14:09.480 they don't fully understand. So that's that's a name. 0:14:09.559 --> 0:14:13.200 Well, the problem there is that the entire model of 0:14:13.360 --> 0:14:16.240 university has just been thrown out the window. And this 0:14:16.360 --> 0:14:18.720 is one of the things I'm looking at in the book, 0:14:18.960 --> 0:14:22.360 is you can't do university or school for that matter, 0:14:22.440 --> 0:14:25.200 the way it used to be done, because it's no 0:14:25.240 --> 0:14:30.560 longer acceptable to say to a student, regurgitate this stuff 0:14:30.600 --> 0:14:34.080 that you came and theoretically listened to for ten weeks, 0:14:34.440 --> 0:14:38.840 because student can point it at your lectures and regurgitate 0:14:38.880 --> 0:14:43.040 whatever you damn well want, having learned exactly nothing. So 0:14:43.840 --> 0:14:48.720 the traditional methods of examining all the tasks you set 0:14:48.760 --> 0:14:53.160 for students to be examined, just don't do anything. They 0:14:53.200 --> 0:14:56.920 don't prove that the student learned anything, and in all probability, 0:14:56.960 --> 0:15:00.600 the student didn't learn anything. What the student learned how 0:15:00.640 --> 0:15:02.600 to do is how to get an AI to do 0:15:02.640 --> 0:15:06.920 the work for them. And I think it's a fundamental 0:15:06.920 --> 0:15:09.840 break in the model of the way university is conducted. 0:15:10.200 --> 0:15:14.600 Universities and schools should be focusing on the reality that 0:15:14.680 --> 0:15:19.440 these tools exist and teaching kids how to do the 0:15:19.520 --> 0:15:23.160 things we've just been discussing, which is, Okay, there's a tool, Now, 0:15:23.240 --> 0:15:25.800 how do we make it do what we wanted to 0:15:25.840 --> 0:15:32.200 do to add value? Not let's all pretend everyone's not 0:15:32.280 --> 0:15:35.600 getting their assignments written by AI, because I guarantee you 0:15:35.680 --> 0:15:39.480 they all are. And don't sit there, you know, on 0:15:39.520 --> 0:15:41.800 your academic high horse saying oh, yeah, but I've got 0:15:41.800 --> 0:15:43.760 this piece of software that checks it and tells me 0:15:43.760 --> 0:15:47.040 if an AI wrote it. Well, that's bs. The first 0:15:47.040 --> 0:15:50.160 thing any student learns is to tell the AI to 0:15:50.240 --> 0:15:53.880 make sure that an AI detection software can't detect the 0:15:53.920 --> 0:15:55.000 thing it writes. 0:15:55.280 --> 0:16:01.680 Yes, yes, yeah, one hundred percent. So I mean like 0:16:01.760 --> 0:16:06.120 this really has kind of accelerated a lot in the 0:16:06.120 --> 0:16:09.200 academic space in the last two years where they're really 0:16:09.320 --> 0:16:12.120 I mean, I don't know, I know that Monash where 0:16:12.120 --> 0:16:16.840 I'm at there, they're reviewing it a lot, and as 0:16:16.920 --> 0:16:19.320 is my understanding, which is very limited. But what I know, 0:16:20.880 --> 0:16:25.960 lots of conversations, lots of meetings, lots of potential strategies, 0:16:26.040 --> 0:16:32.920 ideas what what's you're off the top of your head, 0:16:33.080 --> 0:16:36.280 like you sht your gut feel about, how is education 0:16:36.440 --> 0:16:40.000 going to work? And look, if not as it has been, 0:16:40.440 --> 0:16:43.000 which it can't be as it has been because this 0:16:43.160 --> 0:16:47.160 new tool makes that kind of learning essentially redundant and impossible, 0:16:49.440 --> 0:16:51.320 what's going to work moving forward? 0:16:51.680 --> 0:16:53.240 It's going to have to go back to the way 0:16:53.240 --> 0:16:58.080 it was before everyone started doing education online, back to 0:16:58.080 --> 0:17:01.200 the good old days, Craig. You know when I was 0:17:01.240 --> 0:17:03.880 in university and undergraduate days in the nineteen eighties and 0:17:03.880 --> 0:17:09.399 you actually had to show up, you know, and and 0:17:09.560 --> 0:17:14.280 hand right your exam responses. Let's see Nai do that. 0:17:16.119 --> 0:17:20.040 But I certainly, you know, in my undergraduate degree, had 0:17:20.040 --> 0:17:23.000 to show up, had to go to an exam in person, 0:17:23.560 --> 0:17:25.680 had to sit there for four hours and hand write 0:17:25.680 --> 0:17:26.360 out answers. 0:17:27.960 --> 0:17:31.320 Do you think like that would do it, especially if 0:17:31.359 --> 0:17:34.280 they obviously had no computers and no phones and no 0:17:34.440 --> 0:17:37.280 tablets with them, Like if they needed to go in 0:17:37.320 --> 0:17:42.280 and write essays, do exams, answer questions with no help, 0:17:43.480 --> 0:17:46.560 that's going to there. They're either going to get you know, 0:17:46.640 --> 0:17:49.359 they're either going to learn or they're going to crash 0:17:49.400 --> 0:17:49.840 and burn. 0:17:51.000 --> 0:17:53.080 Do you think now that one stopped I'm using AI, 0:17:53.400 --> 0:17:55.560 you'd still I mean, if I was back then but 0:17:55.680 --> 0:17:59.639 had AI available to me, I'd be using it to 0:17:59.640 --> 0:18:04.840 help me study. Yeah, because but but and so it 0:18:04.840 --> 0:18:05.680 would be still help. 0:18:06.440 --> 0:18:07.960 Yeah, but you've got to be able to go into 0:18:08.000 --> 0:18:11.160 the exam room without AI and prove that you actually 0:18:11.240 --> 0:18:11.560 know the. 0:18:11.640 --> 0:18:13.400 Stuff, at least for that hour. 0:18:13.560 --> 0:18:20.640 Yes, and well yeah, and literally you know, write an 0:18:20.840 --> 0:18:22.159 essay or whatever. 0:18:24.000 --> 0:18:26.200 In response to a question you haven't seen. 0:18:26.600 --> 0:18:30.199 Yeah, exactly, and that you can only do that with 0:18:30.800 --> 0:18:33.240 you know, knowledge, with quiet knowledge. 0:18:33.320 --> 0:18:36.119 So I think that's where we're going to end up. 0:18:36.440 --> 0:18:41.600 I can't see any alternative to significantly greater requirement for 0:18:41.720 --> 0:18:45.800 in person And if you were to write assignments, I 0:18:45.800 --> 0:18:48.800 think the only sort of assignments you could write would 0:18:48.880 --> 0:18:50.879 be By the way, when I went to university and 0:18:50.920 --> 0:18:53.720 did undergraduates, there were no assignments at all. You did 0:18:53.720 --> 0:18:55.960 a full year of study, you did a full year subject, 0:18:56.000 --> 0:18:57.439 and you did a four hour exam at the end 0:18:57.440 --> 0:19:02.159 of it, and that was that really, So you know, 0:19:02.840 --> 0:19:04.480 you could go back to that, but if you wanted 0:19:04.480 --> 0:19:07.800 to do assignments, they could be more like we had 0:19:08.080 --> 0:19:11.919 a discussion in class today about X or Y topic. 0:19:12.400 --> 0:19:15.880 What did you think of Craig's take on that. Let's 0:19:15.920 --> 0:19:17.200 see an AI do that one. 0:19:17.880 --> 0:19:25.280 Yeah, right, right right, yeah. I've I've spoken to Melissa 0:19:25.320 --> 0:19:29.080 who runs my life about this, and she's my ra 0:19:29.359 --> 0:19:35.240 on my PhD. And and obviously we're using AI in 0:19:35.680 --> 0:19:38.239 some bits, like in the research bits, but all of 0:19:38.280 --> 0:19:42.879 my all of my papers are coming out of the 0:19:42.880 --> 0:19:45.399 research that I've done, which isn't anywhere on the internet. 0:19:45.440 --> 0:19:48.400 It's all, you know, so there's there's not really even 0:19:48.440 --> 0:19:53.199 a possibility of cheating. But I said to her, you 0:19:53.240 --> 0:19:56.480 could design a PhD depending on that, you know, especially 0:19:56.560 --> 0:20:00.360 when you do something like psychology, where do you get 0:20:00.359 --> 0:20:05.680 to depending on the university and how much leeway they 0:20:05.720 --> 0:20:09.600 give you. You can design a study. You can write 0:20:09.600 --> 0:20:11.879 your own research questions, come up with your own aims 0:20:11.880 --> 0:20:18.080 and hypotheses, and do I think you could almost if 0:20:18.080 --> 0:20:21.720 you knew how to do it. Not that you'd get through, 0:20:21.800 --> 0:20:24.679 but you could literally do a PhD in a week 0:20:24.960 --> 0:20:28.040 if you knew how to just you know what I'm saying. 0:20:28.600 --> 0:20:30.359 Yeah, I agree with you. I think you could. I 0:20:30.400 --> 0:20:35.000 think if you you know, it wouldn't necessarily get through 0:20:35.359 --> 0:20:38.320 some universities, But look, I think if you knew what 0:20:38.320 --> 0:20:40.879 you were doing, you could crank out. You know, I 0:20:40.880 --> 0:20:42.520 don't know, what's the length of a PhD is at 0:20:42.560 --> 0:20:46.440 sixty seventy thousand words something like that, you could crank 0:20:46.440 --> 0:20:50.520 that out in a night. You know. That's just the 0:20:50.560 --> 0:20:53.520 trick is making it something that is PhD worthy. 0:20:53.560 --> 0:20:57.480 I guess, yeah, yeah, yeah, yeah, My mine's different because 0:20:57.480 --> 0:21:02.280 I've got to write papers and get them published. But yeah, 0:21:02.320 --> 0:21:06.280 so mine is by publication, not by thesis. But for 0:21:06.320 --> 0:21:11.440 people who do it by thesis, it's like, yeah, you could. Yeah, 0:21:11.480 --> 0:21:13.720 you're probably right if you if you had a big 0:21:13.760 --> 0:21:17.560 crack and took a you know, some nicorette gum or 0:21:17.560 --> 0:21:21.000 something and sat down for twelve and a half hours, 0:21:21.040 --> 0:21:24.360 you could probably start to finish your pH d. Yeah, 0:21:26.480 --> 0:21:29.240 what about This wasn't what we were going to chat 0:21:29.280 --> 0:21:29.760 about tonight. 0:21:29.960 --> 0:21:32.440 You probably fail an oral examination of it though, which 0:21:32.560 --> 0:21:35.880 kind of leaves a bit of a clue to how 0:21:35.960 --> 0:21:37.840 PhD should be assessed as well. 0:21:38.359 --> 0:21:41.120 Well, yeah, you make a good point, Like in mind, 0:21:41.200 --> 0:21:48.560 I had five assessments along the way. Four assessments along 0:21:48.560 --> 0:21:53.960 the way where called academic reviews, where you go and 0:21:54.080 --> 0:21:57.160 stand in front of a board and you you've got 0:21:57.160 --> 0:21:59.679 to defend your thesis and talk about your study, your 0:21:59.720 --> 0:22:02.240 re arch, how you did it, why you did it, 0:22:02.320 --> 0:22:04.600 how it was designed, why it was designed, why this, 0:22:05.760 --> 0:22:08.240 why this science needs to be brought into the world, 0:22:08.760 --> 0:22:11.520 why it's different, blah blah blah, all that stuff. You 0:22:11.920 --> 0:22:13.959 have to do that four times, which I've done it, 0:22:14.080 --> 0:22:17.920 and that as someone who does this for living, talk 0:22:18.200 --> 0:22:22.320 publicly and speak in front of audiences. That's the scaredest 0:22:22.400 --> 0:22:26.000 I've ever been speaking to a group. I've never felt 0:22:26.240 --> 0:22:29.600 I've never felt dumber, I've never felt more insecure, I've 0:22:29.680 --> 0:22:33.720 never felt more out of my you know, my environment, 0:22:34.080 --> 0:22:37.680 because I'm, as you've well established, with my audience. Thanks 0:22:37.720 --> 0:22:40.520 so much. I'm not a real academic and I'm talking 0:22:40.560 --> 0:22:43.520 to real academics than I do. 0:22:43.600 --> 0:22:45.960 Graig, Well, you know. 0:22:46.040 --> 0:22:50.520 It's yeah, but it's just fucking terrifying, but that is Yeah. Look, 0:22:50.520 --> 0:22:53.320 there are a few what they call academic milestones along 0:22:53.359 --> 0:22:57.440 the way, but yeah, other than that, I think you could. 0:22:58.040 --> 0:23:02.000 You could. Isn't there a famous story about blokes who 0:23:03.520 --> 0:23:07.639 wrote all these academic papers that were completely bullshit and go. 0:23:07.720 --> 0:23:10.400 Oh yes, yes, yeah, very much on purpose. 0:23:11.200 --> 0:23:15.760 Yeah, got accepted into Q one journal, which everyone that's 0:23:15.800 --> 0:23:19.679 the highest level, like the most highly regarded, you know, 0:23:19.800 --> 0:23:24.040 peer reviewed, kind of well credentialed. Yeah, and they wrote 0:23:24.040 --> 0:23:29.919 these bullshit papers based on bullshit data and outcomes and 0:23:30.119 --> 0:23:33.600 got them actually published. And they did it just to 0:23:33.640 --> 0:23:34.880 prove that that was possible. 0:23:35.440 --> 0:23:40.639 Yeah, yeah, Look, a significant percentage of so called quality 0:23:40.680 --> 0:23:45.439 research probably falls into the category of bullshit. I do 0:23:45.560 --> 0:23:49.520 remember there was a study about what percentage it was, 0:23:49.560 --> 0:23:51.480 and it was a surprisingly large one. I don't have 0:23:51.480 --> 0:23:54.840 it off the top of my head, but yes, just 0:23:54.880 --> 0:24:00.080 because it's in a journal doesn't mean it's true. 0:24:00.520 --> 0:24:04.760 So we'll leave our conversation about the food pyramids next time. 0:24:04.800 --> 0:24:10.919 But my last question read all of this is it 0:24:11.000 --> 0:24:14.200 seems like an understandably a lot of people who work 0:24:14.240 --> 0:24:20.600 in the creative space, you know, writers, musicians, you know, 0:24:20.800 --> 0:24:25.399 all different kinds of artists are not a fan of 0:24:25.440 --> 0:24:27.639