YouSaid · the spoken record

Scott Young

lines on the record: 121
first: 2020-11-16
most recent: 2020-11-16
sittings or episodes: 1
sources: podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

“Well, I think the more you understand things, the better you get transfer. So this kind of goes back to this intuition chunking idea, but like the way transfer works, and I mean, I'm using a model. Like I'm not saying that this is like a scientifically proven thing, but this is just sort of how I visualize it abstractly based on what I've read. But the idea is if you think of this kind of hierarchical model where you have kind of like at the very low level, there's sort of very superficial features of the things that you're encountering. And then there's increasingly abstract features about the problem. And chunking is this kind of finding abstract and more abstract patterns, right? So you're, oh, this is a problem of this type is an example of like, that's not obvious unless you have this, all these base patterns to notice that that's what's going on. And so the idea here is that understanding is this process. Understanding is this process of kind of building hierarchies of information that goes.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“That I can apply. But I mean, economics also has a lot of those patterns, and I do a degree in that, but I feel like I get a lot of value out of those two.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“I want to get the entry level computer programming job that is like, you know, would be the path to a career there. I would have done it a little differently. Like, I mean, the MIT education kind of primes you more for working on not the routine stuff most programmers are working on. It would be more like priming you for working on kind of harder programming problems. So it would have been like a good jumping off point if I wanted to get into certain branches of programming where it's sort of like, okay, this is harder to do. Like the average programmer just doesn't know how to do it because it's kind of complicated and the math is hard and things like that. But I mean, I think it would have not been a bad way to get a foothold into a job the same way that getting an actual CS degree is a good foothold into a job, even though the skills are not exactly the same. I think one of the benefits of having done it though is just that after you see the mental models and the ideas, you kind of have, oh, these are these abstracts.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Kind of projects, and I'm sure there's people I don't know of that did things that were similar, but certainly not the massive waves that I had predicted. And so in that sense, it's a little bit like, okay, didn't quite work out how I was wanting it to work out, but I think at the kind of the more object level, like not my kind of implications of the project, but like what the project actually was, I think it was pretty successful, you know, and I did learn a lot of programming. I did do programming after. Don't get me wrong, but I think if my goal had been”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“I had already learned some programming before, so I did know some programming, but I didn't know academic CS. And I was thinking about going back to school to study it. Like I was thinking about doing back as a mature student and doing another undergrad in computer science. And so this just seemed like the perfect opportunity. And because I had a blog and because I was kind of commenting on student and educational issues, it just felt like, you know, this is just, am I crazy? Why has no one done this before? Like, why am I like, and if I'm going to be the first person doing this, like presumably there's going to be, you know, millions to follow, right? So it seemed very exciting. It's kind of like, this is going to be the road to how everyone's going to do or how a substantial amount of people are going to do education in the future. Now, nothing has changed. I know of very few people who have kind of done anything like the MIT challenge to that scope. I know a handful of people have done their own similar.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“I was just kind of like this is instead of spending hundreds of thousands of dollars, you can do it for free, like that they're giving this away and you could get this knowledge and you could learn it. And I was interested in learning computer science”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“I feel like a lot of the assumptions that I went in going to the MIT challenge did not bear out. So there's a weird sort of bittersweet quality to that like the thing that people fixate on that like I did it is the part that I think was pretty successful. I mean you could look through my exams and deem that my sort of scores weren't high enough for your standards or something but like in the general sense I think I accomplished that but my idea going into the project was that I just finished university. I'm the kind of, you know, like yourself, I was a student up until that point and school has been my life, right? Like when you're in university, all you know is school. And it seemed like this was a huge part of my life and it seemed like this was a huge opportunity that like, why aren't more people doing this? Why aren't more people, you know, this was pre-reading about like kind of all the signaling stories and kind of my general level of cynicism about university in general?”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Anti David Epstein range kind of thesis, but that's not really my point. My point isn't that you should only learn really narrow specialized things and like that's the only thing that matters, but simply that once you decide on what a goal is, it always makes more sense to approach it directly, you know. The kind of having”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“I learned lots of different subjects in the off chance that I'll be able to get one of those deep insights. Now, I think if we're learning for this sort of concrete purpose, though, it's really useful to think of it in that way. It's very useful to think about it, okay, what is it that allows me to perform in the way that I want to be able to perform? And so as soon as I'm learning for a task, I adjust how I'm approaching things for that task, right? So if I'm doing a research paper, I know what I'm trying to get out of things and how I'm trying to read stuff. And I'm not doing something that's totally unrelated. And so I think the directness idea just sort of goes to the fact that, yes, it is good to learn lots of things. And I think that if you do learn lots of things just through sheer volume, you're going to get some of this nice transfer effects. And I'm not even really against learning broadly. Like I think you could kind of close reading of ultralearning could be seen as like a kind of...”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Again, this specificity with the brain kind of comes down to this idea that yes, there is this kind of ability for trans for you to be able to do this, but that's kind of like the magical bonus that happens sometimes when you're lucky. But the base reality is that when you learn most skills, they adhere quite specifically to the context you learn them in and when you're doing it. So if you're budgeting for a learning project where you're trying to accomplish a particular goal and you're sort of like flow chart of how you're going to do it is do this thing that's unrelated to what I want to be good at then the magic transfer happens and then at the end I'm able to perform this skill. That's like a really bad like flowchart for doing this. What you should do is be like, okay, how do I line these up? And so I think learning broadly, learning for the ability to have these sort of mental models or concepts or deeper understandings of things. Like I'm a big supporter of that. I do a lot of that.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, I mean, we have tons of studies showing that people who attend classes don't transfer a lot of what they learn to the outside world. So again, it goes back to my Einstein point where I was making it. Like if you see data, which says that X doesn't work for almost all people, you should be skeptical of that as a general method for being successful at X, right? Like, I mean, it just definitely raises the question of whether it's successful. It's not just that this only impacts kind of like, well, universities are full of like dull, uninspired students, but if like you're the bright student, then you'll do well. Like, I mean, one of the studies that I quote in there was that people who were honors level physics students could not solve problems that differed superficially from the ones that they had studied.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“That, but to me, that's not the how do I program stuff in JavaScript, that is an understanding this kind of abstract theory level, which I mean, we were talking about low ability for transfer, but it's more just that you have to kind of get these abstract ideas in both domains for it to work. Most people don't do that.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Thing, right? That's a graph, or this, like you understand that, and I think similarly, I think that I was just doing a lot of research on motivation, for instance, and there's all this stuff in neuroscience about the motor loop, which is this sort of basic kind of circuit motif that runs through the basal ganglia. And it's kind of how you're like you have all this stuff going on in your brain at the same time. So it's like, how does your brain decide to do one and only one thing? Because you've got like a billion neurons and they're all firing all at the same time. And so it's this pattern that basically just don't lock everything except the one thing that you want to do. And so yes, this is a kind of neuroscience idea, but you can really visualize it in terms of like, oh, like if you had to design a circuit to do this, this is very similar to electrical engineering where it's sort of like we need to amplify one pattern and like, you know, we have to run it through a filter and there's like this center surround pattern. Anyway, there's lots of stuff.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Oh, I mean, like cognitive science is explicitly overlaps with computer science. I mean, cognitive science as a discipline is usually philosophy, psychology, neuroscience, and computer science, right? Like artificial intelligence and stuff. And so when you're reading papers that are talking about like, I remember one paper I was reading, which don't ask me to cite it because I've lost the link now. But here's a guy talking about a computational model for how chunking works. And I mean, if I didn't have a background in understanding computer science, this would be a very difficult paper to read because he's talking about how, well, you make these nodes and then you make these links and that's how you can represent a data structure of conceptual information. I mean, if you've only done like, you know, site classes, this is a kind of a hard paper to understand, but if you've spent classes doing algorithms where you're trying to design systems that do just that, then it makes sense to you. It's a natural language. You're like, oh, yeah, it's a linked list or, oh, yeah, that's a tree structure. Oh, yeah, that's, you know, that's doing this.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“I'm actually glad that I did it that way, but I mean, my preferences are probably different from people who are like, well, you study CS so you could get a computer science job so you could be a programmer so you can, you know, do that. And that was just a different situation for me.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“From a practical point of view, do you want to do the heavy programming assignments to become a real programmer? Yeah, probably. But I think if your goal was, could I get the kind of high theory sort of conceptual education of MIT, then I think it works. And to be honest, because I'm not a practicing programmer, I just program for fun, the conceptual stuff is really where all the value was for me. Like knowing how to design websites is not super valuable for me, but Understanding how information works in general actually translates really well when you're understanding like cognitive science and stuff.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Just final exams, nothing else. And that was sort of my justification for like, could you do it in a year? Because well, you have to learn it and you have to do a hard final exam, but you don't have to do all this other like busy work that you're in school. You don't have to do every single problem set or essays or, you know, group projects and stuff. And I was getting some flack in the beginning, like it was like the first week or two about like, well, why not the programming assignments? But it just turns out that you can add the programming assignments to that challenge and it doesn't materially change the amount of work. I probably only added like an extra week or two to my schedule to also do the programming assignments just because, you know, you do the programming assignments and they're very tight at MIT. Whereas, you know, people were questioning me because they went to different CS programs and they're like, well, I spend months working on this program. There's no way you could do it. And they're like, well, that's not how MIT does their course. They're like, it's all like proof by induction and you're doing things like it's all like drawing graphs on pencil and paper. So, I mean.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Oh, that's another thing, too, that came in my favor is that MIT's computer science program is very math and theory-based, which means that it's intellectually more difficult. So it's like harder to understand the MIT curriculum. Like, you know, when they're doing analyses, they just assume everyone knows calculus. Whereas in university, sometimes they'll go easy on you. Like they won't make you do the calculus. Whereas they make you do the calculus in every MIT class. Like you take the intro microeconomics class. Like when I did it in my school, they never made you do the calculus. They were always like, well, you find the intersection on this graph or something like this. Whereas in MIT is like, okay, so now we take this integral of this and find the Laplacian of this. They just do that in MIT because they assume everyone has a very strong kind of math background. However, the way that works into your advantage is that like in a lot of other universities CS programs, they have these really long, tedious programming assignments, whereas MIT like.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Like just some middle of the road universities computer science curriculum, people would be like, ho, hum about it. But I mean, most of the work is just doing the curriculum. It's not the fact that it's MIT. Like I only picked MIT because they post their material online for free. Like if, you know, the University of, I don't know, like Wisconsin's computer science program uploaded their material. I could have used theirs too. And people would have been less impressed. I think I'm really kind of sneakily leveraging the fact that people associate MIT with being super, super smart. And that's because of how stringent their application process is. And that's the very thing I didn't do when I did this project. So I definitely, you know, looking back on it, like I definitely am proud of the project. But to me, it's kind of funny that's the thing that people fixate on because I also really like the language learning project, but it definitely, I think the MIT challenge is the one that captivates people for some reason.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“It's clearly something that the reason people found it impressive is that when they imagine doing it, they can't imagine doing it themselves. And maybe I get the benefit of that. I don't know whether there's the technique that Cal talks about where you're kind of like, he sort of talks about how you kind of make these sort of inroads into a field and then you find new opportunities. And so the issue is just that there's all this kind of serendipity that pushes you forward to an outcome that people wouldn't have expected. I'm not sure whether that's true at the MIT challenge. I mean, I can explain how I did it and it's not like someone else could attempt it. There isn't a lot of like, well, I just happen to find the right guy at MIT who gave me access to, it wasn't like that. I just did it. But I think that at least in this sort of idea of like, why do people find it interesting certainly because of that? I think I also benefit from the fact that if I had done”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“His example is that if you're in 14 different clubs and you're like the president of the debate team and you're on athletes and you did a bunch of AP courses and this kind of stuff, that's objectively a ton of work. Like it's like to be the high school validatorian to do those things, it's a lot of work. But we can imagine ourselves just being a grind doing all of that. Whereas when we see someone, let's say like one of the examples he gives is a mutual friend of ours, Minish Setsi, who published a book for programming for teens like through an actual publisher that sold fairly well when he was in high school. And like that's the kind of thing. Like, I can't imagine doing that. Like, I don't know, how does it, you know, grade 11 kid do that, right? And so that makes him seem much more impressive, even if, you know, the amount of work is like strictly less than the amount of work that it takes for someone else to do. Now, I think the MIT challenge is a little bit of that example because”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, so the failed simulation effect comes from a good friend Cal Newport, and the failed simulation effect was from his book, How to Become a High School Superstar, and the idea was that the impressiveness to which we assign things is not based on how much work that thing requires, but on how hard it is for us to imagine doing that.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Okay, what are the insights that when I breed these two ideas together, I get like a super idea and not just some like garbage idea that's like actually much worse than what people are currently doing?”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“In how you learn things. But again, I think one of the things I talk about in the book is sort of, all right, given the kind of goal that I'm optimizing for, how do successful people do it? And so again, go to academia and you say, how do I make things that inch forward the field? Well, you have to specialize. You have to figure out, okay, what's a problem that I will understand the best in the world? And then I can find ways that I can make improvements over it. And that's how our whole academic system is organized. And there's sort of no surprise for that. That may not be the thing that we want as a society. Maybe we want these sort of weird cross-pollinating insights that no one saw coming out of left field. But I think as just a pure strategy on your own of just doing it, you're definitely going to get into situations where, okay, well, I learned these two things and there's no obvious overlap. And so I've mastered like it's been kind of useless, right? So it's hard to pick those two things. It's hard to figure out.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“I mean, this is also another consideration as well. Like, you have the whole kind of we're talking about this like everyone else in the world doesn't exist, but the fact that everyone in the world is also kind of pursuing some sort of strategy can sometimes mean that doing something that's sort of objectively suboptimal but is kind of nobody does that I think gives an advantage. I think even my own life is an example of that that like you know obviously it's better if you can actually go to MIT and get an MIT degree and get the alumni network and all this kind of stuff but the fact that nobody was doing the MIT challenge gave me kind of a claim to fame like if only one person is doing it that makes me kind of more unique so the value to my career has probably been higher than getting an MIT degree but I mean if every single person was doing the MIT there's hundreds of thousands of people doing it well then okay maybe the difference changes so I think there is a sort of strategic rule”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, again, I think that one of the key things there would be to specialize, and you see this at top level researchers. Part of it, I think, is just luck. You have to be in a fertile territory, but I think you also are kind of looking for, well, because everyone else is specializing, then I have to do something competitively unique.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Completely different strategies about going about it in my mind, like, I don't think the, you know, even if there's some overlap, like the thing that you do to be able to read ancient Chinese texts and the thing that you do to be able to chat with people on WeChat or something are totally different goals. And so one of my main kind of points that I'd like to make is that you can't even really begin to think about optimizing unless you say like optimizing for what and learning tends to I think learners tend to kind of discount that just because they're sort of like well, I want to know X or I want to learn this topic and sometimes those assumptions about that tend to be the problem”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“A different way of like building around this. And there are more robust strategies. So, like a strategy that works better if you have multiple goals. So if you not only want to pass your class, but you also want to be able to invent things and like you have this sort of wider set of goals that you're considering, I do think that there are strategies that are more robust. So like if I'm learning Chinese, for instance, and my only goal is to look really good for 10 minutes of like an audio thing that I can prepare, then obviously the strategy is to script it and to rigorously practice and have someone is to actually not learn Chinese at all. But that's not a very robust plan. So like if I'm doing like I would never do a project that way because to me it's like, well, that's not really what you want to do, even if that was one of the outcomes that you wanted to have. But I think that, you know, if you're learning Chinese and your goal is, well, I want to be able to talk to people versus I want to be able to read Confucius. Well, now you're talking.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“To not do enough depth, like to not really go into it and understand things. And so in my mind, that's just because the specific constraints of learning a language tend to suggest this is the right strategy, that that's the optimal strategy. Whereas, you know, learning physics, for instance, I think that the problem is just that most people won't get nearly enough depth, right? Like that they just understand the concepts too superficially. They're not seeing, they're seeing the pulleys, they're not seeing the conservation of energy. And so when you're asking about like, what's the right way to approach a subject? I think it's really hard to do that in a goal-free setting. It's really hard to say, well, I want to learn physics the best way possible. I don't think there's really an answer to that. I think there's an answer to, well, I want to be able to pass these types of exams in my university class as the best we possibly, or I want to create groundbreaking physics research, or I want to be able to discuss physics intelligently to other physicists. There's all these different kinds of goals where it would be like, okay, this is.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“So, I think the way I view sort of projects and learning projects in general is that they all kind of have to start with this, what does success look like? What is the outcome that I'm trying to generate from this learning project? Because analyzing things like this abstractly are hard because if you're optimizing for a different goal, the answer might be different, right? Just even to take the language learning as an example, I found for the goals that I had that trying to figure out, let's say, the etymology of words or spending a lot of time figuring out where words come from and what it is, it quickly hits a point where you can be wasting a lot of time. It quickly hits a point where, okay, yes, but actually what I need to do is just memorize more words. I don't need to like really dive dive on that. That being said, that's not a general principle. There's other subjects where I find that the tendency.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Brilliant thinker, but I really admire him because he's the one who kind of like points out these sort of, well, why don't people just do this? And he has the kind of, he has the right kind of mind to even like consider things that seem absurd to normal people. But I would say that's the issue for something like that. And I think it's probably true for all kinds of compound growth. At some point, at some point, it has to stop. You can't just go on forever. But if there's some meaningful range where you get compound growth. Even if it's a short range, it can still be super, super important to pay attention to.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, I mean, the classic reason for that is that that's one of those sort of engineering fallacies where it's sort of like, well, yeah, but obviously once an organization started to have that kind of power, not only would you have corruption from the inside so that like the human agents that are controlling the agency would start diverting the resources and breaking the company's mission. But everyone from the outside would start to kind of dominate it. Like you don't just dominate the world without also having an army and also having like wide popular appeal and like doing this kind of thing. And so, I mean, I could see some scenario that happens where you're creating an organization that's like a new church or something where it like becomes the mass religion and then controls the universe. But there's no kind of secular ambitions organization that manages to concentrate so much wealth for thousands of years and people are like, oh yeah, well, we'll just honor that contract.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Performance for memory, and so now you're doing little tweaks, you know, you're it's a little bit like I would say athletics is similar that if you've never gone running before and then someone first says, hey, there's this thing called jogging. And if you put your shoes on and you kind of run for a bit, like you can probably get quite a bit better for a while. Like there's quite a bit of gains. But then once you're at the level where you're, okay, you're regularly competing in marathons. You've run the Boston Marathon. Well, now you're kind of shaving seconds off your time. So there's probably an S curve there. I think the argument I'd like to make is that most people are before the gains. They're the people who've like, I've maybe jogged a couple times in my life, but they've never taken it seriously. So I think that for the intended audience, I think this book's true. But I think, again, whenever we talk about compound growth, we just have to keep in mind that an unending compound growth is a very, very scary thing in the world because it just implies that, you know, one person”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Johnny Depp movie where he becomes like he becomes the universe computer like no obviously not like there's obviously some kind of okay now you've learned and mastered sort of the edge of what we know about like”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Could think of all these other types of projects that I could tackle now if I was really serious about them. And so I think. There's this sort of improvement not only in your confidence but in your overall strategy. You know, if you've learned a couple languages, then learning more languages just becomes like this routine activity. I mean, it still takes you the same amount of time to get to like that level of mastery, but the way people feel and talk about like, oh, I really want to learn Spanish, but I haven't made any progress in like five years. Like they don't deal with that, right? Like, okay, yeah, I'm going to work on Spanish. And I know I'll be able to get to a level where I'm having a conversation after, you know, a few months just because that's just how it works, right? And so I think that there is this benefit of getting this compounding. Now, if we're talking about like to just the people who become smarter just become this like eclipse level.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“And you try to do this. And I mean, you're reasonably smart. You have the kind of background of stick to like actually get a project done and this kind of thing, which is not a small assumption, but I mean, I'm kind of assuming you have enough self-efficacy to like, okay, I'm going to sit and work on this for three months and then you actually do it. And it's not like a week later. And you're like, ah, I just threw that book in the garbage. That was too boring. But if you're actually able to get through it and follow up on it and actually kind of do the things that I'm talking about in the book, then you can often get to this sort of like, oh, wow, I didn't know that I could do that. Like this is a lot of the people that I interviewed and I talked to who kind of went through this in the book, like Tristan Montebella was just this great example. I mean, his outcome was pretty extreme, but his was kind of like, he was kind of reasonably competent, smart guy, but he'd never tried to do things this way before. And then he does them and he gets this huge result and he's like, oh, this is like crazy.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Confidence to go from one to another S curves. They're probably S curves. They're probably a case that like So, one of the main points that I try to make is that I think that there is this kind of compounding confidence curve, and especially at the beginning, which is sort of where I'm focused on, which is that if you've never done sort of aggressive, self-directed learning projects before.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Learning, like, yeah, if you want to be able to lecture in a language or speak so fluently that your entire life has lived in that language, then yeah, it's going to take you probably a decade. But you get to a level where like traveling in the country is pretty frictionless after maybe a couple months. And I think that's the kind of interesting zone of like, oh, I could get like decent at these skills in a relatively short period of time. How does that change my calculus about what kinds of skills I pursue or what kinds of things I invest in in terms of projects?”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Maybe that's the kind of thesis and ultra learning too is that having kind of like, okay, I can capture the 60% of the value of like or utility of this field in a relatively short period of time. Because if I do kind of plan my learning out very effectively, I can kind of capture it. I don't want to say that that's always the case. And certainly for professional careers, you often need to be in the top like 1% of a skill for it to matter at all. So I do think there's differences. But, you know, definitely there's the case that if you wanted to become good enough as a programmer to do a lot of your own programming for personal tasks, but maybe you're not like what you do is programming full-time. You can easily do that in a year. Like that's something a person could easily do in a year. Whereas the way people often think about it is, well, I have to go to school for four years and then maybe work in an office for, so I have to do it for like a decade to be good at. And similarly with line.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“About truths about the world, there's some basic mental models or concepts which are very fundamental. And once you understand them, you kind of get like the 80-20, you get quite a bit of understanding. And then you get into like the esoterica of academia. Now suddenly they're like debating the finer points of some BS problem that doesn't really matter. And like, yeah, spending a lot of time studying that may be necessary to advance the field, but like clearly from a utilitarian standpoint of like, where are you going to get the most benefit from your learning? It was in that first phase.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Well, I don't think that compounding works generally. I think compounding is as very seductive idea where you just get more and more returns, but really the areas where there is true non-stop exponential growth are like vanishingly rare. And they're vanishingly rare because when they apply, they like totally transform the situation you're dealing with. So like startup growth is an example of compounding return. And it's where one guy in his basement can rule the world after 10 years of grinding or 20 years of grinding. So that's a situation that's vanishingly rare. Most things have kind of regions of exponential growth and then regions of diminishing returns. I would say that, you know, classic Cowan economics thinking is that you tend to think in terms of diminishing returns. So that diminishing returns tends to be the kind of default way of viewing things that you get most of your growth in the beginning. But I think it's certainly”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“And sir, that's his point later in the book his idea about how we overcome this transfer problem is that we teach more theory. Can't actually recall them, then that's useless for you when you're speaking because you're not recognizing sentences. What you're doing is actually speaking. And so I'm very critical of dual lingo there because a lot of the exercises they do are not recalled. They're just doing multiple choice from the word bank, but actually speaking involves recall. It also involves pronunciation. It also involves working around words you don't understand and these kinds of things. And so I think the more you analyze skills you're trying to learn, you realize, oh, this is why this thing doesn't work because the actual thing that I need to do in the performance situation is not what I've been training.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Yes, you can do that transfer. But the problem is that most people don't have this sort of like richly abstract sort of principle-based reasoning about things. And so they just see all these superficial details and they have nothing to do with each other. And if the superficial details of one don't have to do with the superficial details of another, there's zero transfer there because they don't have to like at all. And so I think, you know, one of the books that I read that was sort of a major source of the research on transfer was, I think it was called Transfer of Cognitive Skill. And I'm blanking on the author's name right now, but it's mentioned in my, I think it's Haskell maybe, is mentioned in Ultra Learning.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“That it learns things quite specifically, and that is how it works. It's not just like a defect, but like that's why we're smart is because we can make very fine grain discriminations between things. And so you can get transfer if you're thinking about it in terms of like, well, I understand domain one well enough that I can see abstract pattern here and I understand domain two well enough that I can understand abstract pattern here and I can see that these are the same abstract pattern and I can make that linkage. But the problem is that if you haven't chunked the first domain enough to get up to that pattern and you're trying to talk about the next, they don't match because superficially they're quite different, right? So this is like this issue where yes, if you understand physics problems well enough that you can see that this conservation of energy is this principle that holds and then you start learning a new domain and you realize, oh, this is actually like conservation of energy. You get it up to that level of abstraction.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, I mean, I think so if we're kind of visualizing it a little bit here, the principle of intuition is sort of recognizing that because of this sort of hierarchical structure of chunking, you have this kind of building up layers of abstraction on top of each other. And the thing on transfer is showing is that when you go from one domain to another, the problem with the problem with the transfer idea is that when we talk about learning skills, we tend to use fairly general labels for things. So we say kind of like, well, I'm going to get good at X. And X is just this sort of broad category of skills. And what is sort of missing from that is that to actually perform those skills quite well, you have to do something very, very precise. And so it seems to be a general feature of the brain.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“It seems like you probably do this through chunking. So you probably learn all these smaller patterns and you build it up so that when you see it, it just all comes together and you can see what the situation is. And so I do think it's probably the case that if you're learning something that has a lot of abstraction, really spending time to kind of familiarize yourself with the ins and outs of the more basic pieces, allow those things to kind of lock together faster. And so, you know, the vignette I chose for that one was Feynman just because he had spent so much time kind of playing around with math and he just had so much familiarity that he was like this encyclopedia of like weird math trivia. So, you know, like just random stuff comes up and he can come up with the answer because he has all these patterns that are stored in memory.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Of energy aspect of it is a kind of abstract property of the problem. And so whether it involves a pulley is an obvious aspect. And so what you need to be able to do is look at all those obvious aspects and then figure out what the abstract or higher thing is.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“Concepts, yeah. I don't know whether it's the case that you're like, so there's two issues there. So the issue of being able to look at what the deeper principle is, I think, is part of this problem that you need a lot of exposure to a field and to the knowledge in the field in order to build up this kind of repertoire of patterns so that when you see it, you can actually see what the principles are. And so that's one of those sort of things that there's, I forget the name of the study, but it was like one of those classic cog size studies where they took physics, novices, and experts and showed how they look at problems in the novices focus on stupid stuff like, oh, this involves a pulley or an inclined ramp, whereas the experts are like, oh, this is a conservation of energy problem. And so the kind of naive way is thinking, oh, well, it would be nice. We should just tell people how to recognize whether it's a conservation of energy problem. But the problem is that the conservation”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“If you have kind of, if you feel like you're more distractable, if you feel like you struggle to understand concepts, I mean, those are all things that would apply. So that would be my connection between age and learning. But I think the broader thing that I think matters is just that when you're learning something, using the correct strategy is mostly going to be the same. So I don't think it's the case that there's some strategy that works for really smart people and some strategy that works for people that don't. The same thing's probably work. It's just it's going to be easier if you're more intelligent or you're younger or you have those advantages.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source
“For instance, there are studies that show that if you're making flashcards, for instance, and you have to do more manipulation to show how the parts connect with the flashcards. So like in this particular example, I'm working off of memory here. They were talking about like learning Chinese words. And so there's the characters in the pronunciation. There's two characters and two pronunciation. And you put them side by side, what you have to do is take the first part of one and link it with the first part of the other and the second part of one and link it with the other. And it turns out this makes it somewhat harder to remember than if they're on top of each other and there's just an easy visual link. And so you could think of maybe if you're older and you're struggling more with these kinds of issues organizing your material better, making the connections between things you have to learn more explicit, having that kind of pre-processing work might be a little bit better. But I think those are also things that would benefit people who just struggle with learning more.”
2020-11-16 · Dwarkesh Podcast · Scott Young - Ultralearning · IDENTIFIED FROM THE TRANSCRIPT · source