Below are the highlighted quotes from the 4th episode of The Inside View Podcast.
“I feel like a lot of us are just being cheated out of the most sort of interesting or beautiful things in this world. You have this track that most people go on, from grade school all the way up to high school, and then to college, then to some kind of professional environment; the “life highway”. And the further along I get, the more I'm realizing that all of the interesting stuff seems to exist off the menu.” (youtube)
“There's a way in which the experience becomes less and less interesting the closer it is to some kind of center of society. As a thing becomes more and more mainstream, it becomes boiled down to its simplest, most easy to consume form.” (youtube)
“I'd never say that “I believe in the scientific method”. There's something a little bit strange about people saying things like "believe in science". To me, “believe in science” almost sounds paradoxical. Science is not this thing you’re supposed to believe in. It's a mechanism for testing hypotheses, for troubleshooting our world, right? And so it's like, well, “believe in the debugger” or something like that. What if the debugger is wrong?” (youtube)
“People crave meaning. And religion seems to be the thing that is best at giving that meaning to people. When you take that thing out, then you end up with all of these different mechanisms that now, all of a sudden, feel the need to give you meaning, right? Now, politics is the thing that gives you meaning, or now science itself is the thing that gives you meaning. And it somehow corrupts all of these things, because it basically turns them into quasi-religions, whereas if you have a religion, these things are free to not be quasi-religious and they can actually function properly.
When your political party doesn't have to be the thing that gives you meaning, we can actually have a much more constructive discussion about what the proper policy should be, because now you're not emotionally attached — your identity is not attached — to these political causes, for example.” (youtube)
“You get to the university, and basically — at least this was my experience — you get told to specialize, right? But imagine you enter this big makerspace. There's all of these tools on the wall, and you can explore all these tools. You can go make whatever you want. And then this guy comes up to you and says, "Well, actually. You're just going to be using the hammer. And there are very useful things you can do with this hammer. If you properly train yourself in using a hammer, you'll be useful to people who actually need somebody with hammering skills." And that's kind of... I don't know. I don’t think that’s a very healthy way of taking advantage of people's skills and interests.” (youtube)
“A lot of people say that Steve Jobs had a “reality distortion field”, where he would just say things that weren't true, but he convinced himself and everybody around him that they were true. And sometimes that led him off a cliff, but sometimes actually that led him to push through and persevere and actually achieve these very impressive things. And so, having too big of an ego is certainly a problem: if you are way, way, way more confident than your beliefs actually allow you to be. But if you place yourself slightly above that line, where you're slightly overconfident, I feel like that's a much better position to be, because you'll actually end up pushing through and persevering even if you're wrong, because that energy will just kind of carry you.”
“It's almost like you're distorting your reality a bit, and then you're letting reality catch up to your distortion of reality.” (youtube)
“I kind of just approach [Twitter] as making friends. I try not to overcomplicate it. I just look at it as like, “is this the person who I can potentially vibe with and learn things from?” And if the answer is yes, then probably it goes both ways. And so, I kind of use Twitter as a platform for making friends — and maybe there'll be opportunities to work on cool projects with these friends. That's really as deep as it goes for me.”
“The big problem I have with people using the word “networking”, for example, is that it reduces human relations down to this very utilitarian thing. People become resources.
And I don't know, it's just— if somebody were to ask you, "Why are you friends with somebody?" and you have a very clear answer to that question, you're probably not friends with that person. You're probably using that person in some way.” (youtube)
Sav (explaining Tim Urban’s “Traffic Rule”):
“If you're hanging out with someone and it comes time to either drive them home or call them a cab or something, you end up hoping for traffic, because it's an excuse to talk to that person more. I found that to be a very good rule of thumb for friends, I think.” (youtube)
Michael (another rule of thumb for friendship):
“If I give you $1,000, is there a way to reach out to you by your mom, your friends, where I can ruin your life if you don't give me back the $1,000? Or are we just too distant for me to have an impact in your life and I would just need to spend $1,000 in plane travel to just get back my $1,000?” (youtube)
“I feel like somehow, on a spontaneous or intuitive level, friends — good friends, top friends — would spontaneously reach out to each other. And even if you don't reach out to each other for a while you know that's because the other person is busy and not because the person has lost respect in you or interest in you. I feel like it all boils down to some kind of mutual respect. So then there's mutual respect and knowing that the other knows that, estimating that the other likes you or something, the same way you like him.”
“There's this saying: a friendship is some path in the sand where you need to walk every time to make it alive, otherwise it will be erased by the water. I feel like why this path exists is because you'll keep walking over it and going on walks with those friends and having conversations with them. You can try to maintain those relationships as part of a professional network, but it will always be artificial. I think a good rule of thumb is: do you respect this person? Do you think this person respects you? Does this person spontaneously reach out to you? Do you want to spontaneously reach out to them?” (youtube)
“Whenever I took psychedelics, I would enter some kind of state where I would want to stay because it was very cool and we had all this vibe going on and we were like, "We're in the same room. We're listening to this music. We hope this stays forever." And then, at some point, the effects go down a bit. You start being tired, you want to eat, or maybe actual life — like there's an actual job that you need to do. They're the distinction between vibing and having fun with your friends, and then actual real life.” (youtube)
Sav (from “Ayn Rand, The Fountainhead”):
“What you feel in the presence of a thing you admire is just one word – ‘Yes.’ The affirmation, the acceptance, the sign of admittance. And that ‘Yes’ is more than an answer to one thing, it’s a kind of ‘Amen’ to life, to the earth that holds this thing, to the thought that created it, to yourself for being able to see it. But the ability to say ‘Yes’ or ‘No’ is the essence of all ownership. It’s your ownership of your own ego. Your soul, if you wish. Your soul has a single basic function-the act of valuing. ‘Yes’ or ‘No’, ‘I wish’ or ‘I do not wish.’ You can’t say ‘Yes’ without saying ‘I.’ There’s no affirmation without the one who affirms. In this sense, everything to which you grant your love is yours.” (youtube)
Michael: Hello, Sav. How are you doing?
Sav: I'm good! Nice to meet you.
Michael: Yeah. I'm glad to have you here. What time is it where you are?
Sav: Well, it's currently 1:30, which is... Just kind of nice, I guess. I only woke up like an hour ago. I'm in Calgary right now.
Michael: Where's that?
Sav: Calgary, Alberta. Canada. I lived here for, quite a while. My parents moved here when I was about six years old. So I spent most of my life here. And then went to Vancouver for university, in UBC.
Michael: And now you're back?
Sav: And now I'm back. Temporarily, yeah.
Michael: And you're spending most of your time, from what I've been reading online, writing blogs, and learning about cool stuff?
Sav: Yeah. I guess, what the big overarching theme of everything that I've been trying to do is identify the things that I'm interested in and sort of... I thought that through writing about different topics, I can learn a lot about the different topics that I'm writing and also get a better sense of the kind of things that I actually feel pulled towards. And it's the same with all of the various side projects that I'm doing. I'm still just very much in this process of just trying to learn as much as I can about this world and just trying to see how I can fit in and be useful and do something that is not only interesting to me but also helps people.
Michael: And you're also learning about learning?
Sav: I am learning about learning, yes.
Michael: Yeah. I found your blog about 'top-down learning' and how to optimize your learning, quite interesting. So, essentially, if I can steelman one of your points, you were saying that if we were in a perfect world, learning would be easy. So, people would essentially know your brain and give you the shortest program that makes you understand concepts. And, also, you give some definitions of knowledge as words that bind to each other. So, yeah. Have you evolved? Did your definition of intelligence or knowledge change from that moment you wrote this post or something like this?
Sav: Well, yeah. I guess those... Let me see how I want to say this... A lot of my posts about learning how to learn, just kind of come out of frustration than anything with just like the current way in which we teach people things. And I don't know. I've spent, I guess, like three years in the university now and I've kind of became kind disillusioned with the whole process. And that got me into thinking about how, well, in a perfect world, how we should learn. You asked about the definitions of knowledge and intelligence. Sorry, can you remind me again what you're looking for?
Michael: So you're essentially saying that knowledge is... So you're extracting knowledge as something that is useful. So every bit of information is useful to you, and then you go on about knowledge being coupled to other knowledge. So when you have concepts, you understand better than if you can relate, connect each different book, different authors, different ideas. And then there's intelligence in kind of being able to... Maybe we can just go on what is the... Maybe you can define those terms if you want, or I think when your definition, let's say, your definition of knowledge - bits of information that make us more capable of doing stuff is kind of connected to one definition of intelligence as our ability to solve problems in general.
Michael: So, we could say, does intelligence or, I think in your blog, you have intelligence... So there's this efficiency in learning. So I think in Chollet's paper on intelligence... is intelligence a program. So.. his Definition is like how efficient you are at sampling over information or in like an... Filtrate through information, depending on your priors. Also, how difficult it is to generalize from new data. And I forgot the third one.
Sav: I guess the way that I remember it is: it has to do with your priors, which can be either like your genes or something along those lines, and then the environment that you find yourself in, and the problem. So the environment around you kind of shapes your way of thinking as much as your priors do, and the problem space actually shapes the way that you think about it as well. Cause it's like, imagine being in a perfect environment, having good priors, but working on a problem that you just don't care about, right? You probably wouldn't do as good of a job on it as something that like, truly grabbed your attention.
Michael: So yeah, I think there's being good in some kind of training environment. And then this environment is good for generalizing to the real world. I just re-read the abstract of the paper and it says scope, generalization difficulty, priors, and experience. So scope, I think is how big is the problem. And experience is perception, history, and memory. So I think what you're saying about the environment falls into scoping experience.+
Sav: Yeah, I think that's a good way to break it down. Yeah.
Michael: So, if I try to summarize your moto from just reading five or six of your blogs, I would say something controversial as you're trying to mix like... understand how to better understand the world. So, what are good tools for thought and good ways of thinking about the world in general and, the other part would be becoming smarter and becoming smarter at being smarter. So that's like the two ways I see where you're heading at.
Sav: Yeah. There's certainly that kind of through-line to it. I don't know, maybe this whole sound, I feel like a lot of us are just being cheated out of the most sort of interesting or beautiful things in this world. And I feel like the stuff that is being offered on the menu of like... You have this track that most people go on and you go from grade school all the way up to high school, and then to college, then to some kind of professional environment; the life highway. And, and I guess the further along that I get, the further I'm kind of realizing that all of the interesting stuff seems to be off the menu.
Sav: I used to think that you basically go in and order everything on the menu and that's basically the best stuff that life has to offer. It's like if you do everything properly then actually you'll end up where you want to go. And I guess I've just been more and more kind of like disillusioned with that and that kind of leads to this... Well, it's like, okay, if the stuff being offered on the menu isn't actually the most interesting, then how do we learn all this other stuff? How do you pursue your own interests? And then that gets into the efficiency of learning and all the rest of it.
Michael: So I guess what's important is what's on the menu and what are we eating? So I guess we can be eating knowledge or at least pleasurable experiences. So maybe what we're trying to aim for should be interesting conscious experiences, either of knowledge or perception or thoughts, where, as I said, when you were in the lifestyle highway, you're trying to kind of maximize for social capital and social status and trying to fit into those criteria that you think people will find you attractive or respect you. And then you think this will maximize your social connectedness and interestingness in conversations. So I guess what we're saying, what I thought of when you were saying this is mostly: one, there's a way of bootstrapping social relationships and work, where you can just like email people and talk to them without being a professor at Harvard.
Michael: And the second point is that actually being a good economic maximizer, a good economic agent, for the economy is not as productive or, a pleasurable experience as just like taking LSD. So there's like this, and kind of quoting Nick Cammarata you start to realize that there's much more. Like we're pretty close to very cool, great experiences. And it doesn't take much more than just like a couple of molecules for us to experience very different experiences. And maybe we should think about how we can change the perception, or at least how the machine works inside and become smarter, or like those things. And not just do what people have always done and research papers, you know, a career and all the things. I think that they have like hacking our brains and then hacking society and connections.
Sav: Yeah. That's an interesting way to put it. Yeah. I think that's a really good way to describe it, because if you think about if we’re to take this sort of menu metaphor further there's a way in which the experience becomes less and less interesting the closer it is to some kind of center of society, as a thing becomes more and more mainstream, then it kind of becomes boiled down to its simplest most easy to consume form.
Michael: So, Goodhart's Law, but for society.
Sav: Yeah, I guess in a way it's really just this kind of... If you've ever heard about that metaphor of like our knowledge or like scientific progress is like this bubble. And then it's, like keeps on expanding. And the surface area, the stuff we don't know keeps getting bigger and bigger. But also that center becomes more and more well-established and it's like the point people usually make is, it's more interesting to be on the fringes of science, right? But, I feel like it's the same way with everything. It's like this with art, for example, with pop culture, the summer blockbuster movies are usually less interesting than indie films.
Sav: And in the same way, if you think about career paths and sort of what people are doing in life then the most commonly traveled road is usually the I guess, the least interesting road. Or if you think about, like if you were to talk about psychedelics the most common sort of conscious experience, in some ways, the least interesting conscious experience. And you can just like go and apply this to every single thing. And you just start realizing this is well, it's just more interesting to be on the fringes of all these different things. It's like, if you're going for an interesting experience, you're usually on the fringes.
Michael: I guess, by definition, interesting is something we've not considered often. So by definition, something that is out of distribution is interesting.
Sav: In some sense, yeah.
Michael: Yeah. I get what you mean. Your definition is more like interesting is a pleasurable experience.
Sav: Yeah. Now the most awful experiences are also on the fringes. Bad trips, being a homeless person, or like an especially terrible movie... These things also exist on the tail end. Right. But it's just the opposite tail end.
Sav: And that kind of ties into like, what I wrote about in my post 'Is a Whale a Fish?', talking about the Pre-Trans Fallacy, and just talking about how there's the conventional viewpoint on something, then there's the pre-conventional viewpoint, which is the less advanced viewpoint and the post-conventional which is more advanced, more enlightened than the conventional viewpoint. And a lot of people mistake the pre-conventional for the post-conventional because the one thing of course that they share in common is that... If you hold a viewpoint that is either pre or post-conventional it's not like the conventional viewpoint.
Michael: I get what you mean because I've read the post, but can you give a simple example of, like an interesting example for the listeners?
Sav: Sure, sure. Well to take one from the title of the post, and this particular example, I think comes from Eric Weinstein. I think I got that right. Basically, you can ask the question: is a whale a fish? And if somebody answers well, yes, the whale is a fish...
Michael: Yes. A whale is a fish.
Sav: Is that person more or less enlightened than the person that thinks that, well, actually the whales are mammals, right? Like you might just be a person that doesn't know that a whale is a mammal. And therefore you're saying, oh, well, it looks like all the other fish, therefore it's fish, or you might have sort of a more advanced take on things when it's like, well, actually, you know that whales are mammals and they have this interesting evolutionary history where they used to be on land, and then now they're back in the water, but you just define fish as being anything that swims in the water. And so, when you're given this answer of yes, a whale is a fish, it's like, well, does this person know more than me? Or does this person know less?
Michael: I think there's a bunch of different ways of looking at it. And I think the most obvious... So one very interesting parallel analogy for me is making jokes. So irony. So what you say about pre-trans: there's pre-conventional then post-conventional, or trans-conventional, in some way pre-conventional and trans-conventional say the same thing. They say that the whale is a fish, right. But for different reasons and the trans-conventional knows, the reason why they say mammals, but they say fish because they're trying to prove they're smarter than most people who would say that it's a mammal. So essentially we have similar things with feminism where let's say pre-conventional would be something from decades ago where people will say, men are superior to women, so men should vote and not women, and then conventional would be, you know, actual feminism and equality of rights.
Michael: And then trans-conventional would be people arguing that men and women are not equal, and giving arguments that try to contradict the bad arguments from feminism. And so yeah, as we see, there are two ways, there are also two binary things where we can oscillate between those two. And I always save stronger and more complex arguments, same with irony where you can be serious than sarcasm then meta - like being serious because sarcasm becomes boring. And then you can level up as many layers of irony as you want.
Michael: And there's also this post by Scott Alexander called Intellectual Hipsters and Meta-Contrarianism, where he's making the claim that so essentially pre-conventional is people who don't know what to wear. Don't know, don't know how to have good looks. Conventional is I think good looks like you're wearing good colors and so on. And trans conventional is just like wearing stuff that looks weird, but actually has good taste. So, he thinks that there's like intellectual hipsters, so people that are against them - as in, people from the dark enlightenment - they're mostly, I feel, intellectual hipsters. There's always the normal take, the contrarian take, and then you can go meta-contrarian against the contrarian view.
Sav: Yeah. Yeah. I guess a couple of thoughts as you're kind of talking about that. The first being that of course, like you're saying, you can always go a step above. And so if you imagine a particular frame where you think that you hold the post-conventional view, let's say a year, one of these dark enlightenment people, and you're kind of an intellectual hipster and everything, right. Well, actually in your community, you hold the conventional view, right? Because you're surrounded by people that also hold these views. And if you were to think about this in kind of relative terms, well, you could always ask the question: what would be the post-conventional view of my belief, right?
Sav: It's not that there's this one mainstream conventional opinion. What I've found is that to any opinion, there are people that know less than you, and there are people that know more than you, and you should generally try to listen to the people that know more than you. Or at least, whenever you're having a conversation, you should always assume that somebody is making a more nuanced point instead of a less nuanced point.
Michael: I think politics is a good concrete example where if you have both very smart liberal - a liberal sense of American liberal - versus a Republican, also very smart, and they're both completely opposed in terms of values - in terms of ought - then they both can have very complex arguments about what they believe, but I don't think they will reach an agreement. So I think maybe they will understand each other's arguments, but they will not agree on the premises.
Michael: And I feel like most kinds of debates about feminism are about premises. So basically the premise of feminism would be women, people should have equal rights, the premise of equality. And the premise of the opposite view would be arguments should be good, and definitely scientific evidence. And we should describe what the world is instead of what it ought, so it's essentially like ought vs is debate in my view. And I don't think you could find someone smarter than you at any level that will tell you something, and then you can change your view. I think at some point there's like some kind of very separate sides where you need to take a side or at least arguments come from those premises that are different.
Sav: That might be true. At some point, obviously, this has to end, right? Like at some point you will reach a person or a viewpoint that is - quote-unquote - the best viewpoint on a particular issue or the viewpoint that understands the complexity of the issue the most. I guess just the mindset that I go into it with is that well, I'm probably not the person with that viewpoint. And so it's interesting to just consider, if I hold a particular viewpoint on, or if I'm making particular arguments, what if I'm wrong? Right? And that asking of "what if I'm wrong? What if there's actually a better read on the situation?" I think that's something that is valuable to do. A lot of people don't really do it. And they just kind of assume that what they think is correct. And of course, everybody who's two has an opinion that's correct. They're all going to be like, no, this opinion is wrong, but I'm going to say it anyway.
Michael: Okay. So I think I got most, I understood better the second time. So essentially, you're trying to steelman, like as an ideological Turing test, and try to understand that others position well enough to be able to pass the Turing test for this particular view. And try to come up with counterarguments that are more elaborate to that steelman of your opponent. So it's more like, are there better arguments for my position that would defeat my opponent somehow. And then if there are not, then I might just change my position and think that the opponent has better arguments, and then you can move on later.
Sav: Yeah, yeah. Although I wouldn't think of it in terms of defeating my opponents or whatever. Maybe that's a perfectly fine way to think about it. I just find myself coming from a perspective where I'm trying to figure out how this world works and I'm trying to have the most sort of accurate understanding of how this world works. And it's like, well,
Michael: It's like a cooperative game. I feel like.
Michael: I said winning and losing, just because of, I think in any discussion, even if it's like collaborative like we're having a collaborative discussion, at some point we polarize and take a side to kind of go on this debate. So I think there's like some winning and losing, or at least I'm trying to add the best arguments competition.
Sav: Yes. And those can be helpful actually. Kind of engaging in local arguments, but really trying, doing it for the sake of whatever point is the strongest and I think that's worth doing but, I think that a lot of people tend to get into this mindset of like us versus them and it's an.
Michael: In-group versus out-group.
Sav: Yeah. A lot of people would pick teams, and they stick to those teams. It's like being a sports fan or something like that, right? You're not a sports fan because you're making this rational decision. When you think that team A is better than the other team, it's more like, this is your team and so you clearly want your team to win, and then you want the other team,
Michael: what about your team? [crosstalk 00:29:58]
Sav: And that's not a very productive way to understand the world.
Michael: Well, okay. Most people will support football teams or like sports teams, but it's kind of that bit of a ...
Sav: Or political parties.
Michael: What about people understanding all those bad aspects of football or any sports games, and that rationally decide that they want that dedicated brain are happier when they're part of a social group and decide to be in a team because they know that it will produce positive outputs in terms of neurotransmitters. So I think there's some steelman of people actually enjoying sports in a way to trick the brain into believing they're part of a group.
Sav: I, yes.
Michael: I think it's a minority. So sorry if I'm like going to edge cases.
Sav: Well, no, no, that's interesting. That'll be interesting. I haven't met anyone like that, but I could see how that person could exist.
Michael: I've talked to someone like that just today.
Michael: And so the other part would be people trying to be very convinced about their arguments and really taking a team, and in terms of intellectual disagreements. I'm just like... I think if trying to optimize for learning and trying to steelman my position and sticking to it and trying to win a debate is kind of a good way to... a productive way where you have a good feedback loop. And so if we're trying to maximize learning efficiency, then it might be good because also, people tend to trust more the people who are more confident. So if you kind of train yourself in building this confidence in taking some positions very strongly, then I think, I think there's some kind of social feedback loop that can make you learn faster. Somehow, there are some social incentives and evolutionary pressures into being confident in your positions, but yeah. We both agree that it's better to cooperate... not better, but the aim is to learn more about the world and not win or lose.
Sav: Yeah. It might be a good idea to... I mean, whatever beliefs you hold, to be kind of ready to defend those beliefs and to hold them strongly, but I think there's a... I can't remember who this is from, but there's a saying I came across where it's like, you basically want to sort of hold strong beliefs loosely or something like that, so just have strong beliefs, but at the same time, if there's evidence for the opposite, then not be so attached to these beliefs that you can't let go.
Michael: This is like every LessWrong post I read. Epistemic status, strong belief, lightly held.
Sav: Yeah. Yeah. Maybe that's where I got it. Maybe I got it from LessWrong forums, but that's... I guess, the thing that I noticed the most with people who are just very attached to a particular cause or a particular topic, really just becomes this emotional thing for them. It's not like you're rationally adopting your position because you want to be part of an ingroup or something, it's just that you think that you're the good guys... sort of holding out this sort of kind of outpost of civilization, right? And it's... I don't know, this kind of emotional attachment that gets in the way.
Michael: I adequately agree that there's some emotional attachment to sports and also politics, and I think we both agree on it's more people wanting to update and-
Michael: If you take any social cause, you'll find people who are very emotionally invested in these social causes, whereas if you look at it some kind of scientific topic... I mean, sure, this also comes up in science, but it's much more kind of... It's not as intense in science. It's like if you're studying some kind of scientific phenomena, if your hypothesis turns out to be false, well, you'll probably just accept the data. You won't go in and start putting out a bunch of propaganda about how you're right, and the other... There's just not that kind of emotional investment.
Michael: I think we can make... I think there are links to what you're saying. In the post, you wrote about religion, which is called, TLDR: Religions, Values, and Peak-Experiences, where essentially, you're making the point that... I think maybe like, kind of sports could be the religion if we're doing analogies. And there's this kind of scientific bubble... not bubble, sorry, scientific mainstream view of atheists, who think that we cannot prove anything about religion, so there's no evidence of religion, so we should just straight up to avoid those things. And then there's like, people who are more interested... were kind of interested in, maybe, more Hume's ought, and also conscious experiences and what might be possible.
Michael: So I think that's where... So if we're trying... Like what you were saying about people not being able... So people being in teams, not being that rational and somehow, smart people getting their view and trying to figure out what is true. I feel like you were kind of defending the scientist's view, whereas in your post about religion, you kind of attacking the scientific view and saying like, oh, you can choose a religion, or at least experiment ways in believing in things, where I think maybe... if you believe in a social value, then you're kind of believing in some kind of religion in some way. Where are you at now? And you're be in between both.
Sav: Yeah. Well, it's cool to see you bring that up because you're right. It does appear to be a contradiction. I mean, really, that whole post was, I guess... I read this book by Abraham Maslow, where he was basically making an argument in favor of some kind of unity between science and religion, and he was writing this in mid-20th century. And he was just describing this world where he's just seeing science and religion diverge, and then you get this quote, unquote "rational" scientific world, and then you get this sort of superstitious religious world.
Michael: Is it called Religions, Values and Peak-Experiences by Maslow?
Sav: Yeah. So that's what the book is called. It's an interesting read. It's quite a short read. I'd recommend people to read it.
Sav: Yeah. I guess, the way that I kind of think about... the role of emotion on the role of faith in the stuff that people do, I've come to the realization that we kind of have a religious part of our brain. We tend to... It's almost like we're hardwired, or predisposed, to think about things in religious ways. And when you take religion out of the equation and you still have that need to satisfy your kind of drive to religiosity, then it seems that that thing gets channeled towards ideologies, social causes... like red versus blue. Just like picking it... because it's really... A lot of people are treating these things as if they're religions, almost.
Michael: What about the meta... kind of meta-religion would be believing in the scientific method? I feel like most atheists believe in the scientific method, and in a way, their religion could be to be against religion.
Michael: Yeah. Yeah. And really, that's just another religion, right? I'd never say that I believe in the scientific method. There's something a little bit strange about... People recently say "Believe in science". And to me, believe in science, almost sounds paradoxical. Science is the thing that you're not supposed to believe in. It's a mechanism for testing hypotheses, for troubleshooting our world, right? And so it's like, well, believe in the debugger or something like that. It's like, that's really not... What if the debugger is wrong, right?
Michael: Yheah. I guess there's some interesting things in your posts from Eliezer Yudkowski on belief about beliefs. So I think he talked about believing you believe something, and there's also all those meta preferences of wanting to believe something, or wanting to prefer something over another. But yeah, sorry. It was just something in my mind and I interrupted you.
Sav: Yeah. Yeah. I guess just to kind of put up a bow on this point, I've kind of come to view religion and the mainstream religions, like Islam, Christianity, et cetera, et cetera, as kind of the most effective things at satiating this drive, or this craving, for religion that we have. And while they do come with their own downsides, when you take these things out of the equation, then you actually end up with... If there's no religion, we basically end up religifying everything, right?
Sav: Because we're still chasing some kind of religious need, and so I feel like that's where you get the whole like, believe in the scientific method or believe in democracy or believe in the Republican party or-
Michael: My problem with your point is that essentially, you could say the same about the need of people to make sense of the world, finding meaning in the world. So it's not like various... I don't find truly specific to religion, as you basically saying, people crave for meaning, in general.
Sav: Yep. [crosstalk 00:42:57] I think that's a cleaner way to say it. People crave meaning. And I guess, the way to frame my arguments, in this case, would be is that well, religion seems to be the thing that is best at kind of giving that meaning to people, and when you take that thing out, then it's like you end up with all of these different mechanisms that, now all of a sudden, feel the need to give you meaning, right? Like now, politics is the thing that gives you meaning, or now science itself is the thing that gives you meaning. And it somehow corrupts all of these things, because it basically turns them into quasi religions, whereas if you have a religion, these things are free to not be quasi-religious and they can actually function properly.
Michael: So you're saying that we built infrastructures out of like churches. I didn't get the last part, can you repeat it?
Sav: Yeah. Basically, if you have religion serving as the thing that gives everyone meaning, then it frees up all of these other institutions that do not have to do that anymore, right? Now your political party doesn't have to be the thing that gives you meaning, and now we can actually have a much more constructive discussion about like, well, what would the proper policy be, right, because now you're not emotionally attached and your identity is not emotionally attached to these political causes, for example.
Michael: Yeah. A bit like I would say, the end of religion or at least the kind of Luminaries period. It's called enlightenment, I think, maybe enlightenment, where people would kind of stop believing in a religion that much. Is it enlightenment or... I don't know the English term.
Sav: Maybe. I don't know my history at all.
Michael: [foreign language 00:45:02]
Sav: That's one of the things...
Michael: Yeah. Yeah. I guess-
Sav: I guess you could probably trace it back to the Renaissance.
Michael: From middle ages to modernity. Well, yeah. In my country, we cut the head of our king, and we decided to make everything digits and not... remove every symbol of God everywhere, and just make law the most scientific way. So that's what I called maybe enlightenment. So I think so be we kind of disrupted this whole churches, and then people didn't have those like, the infrastructure of social status from churches. So they kind of built another religion somehow of science, and now, most atheists kind of made me believe in science or at least believe in social causes and all this stuff about... The other stuff is about meaning.
Michael: We're always in a need of those things. And when you were saying things about building churches and giving people meaning, I think it's kind of connected to this concept of Chesterton's fence, where we try to see why there is a fence and not completely erase it. So I think the value of the church is not... Obviously, it's not only giving meaning, but it's also all those societal... and the gathering and talking to the priest and being a way of meeting people and so on. So when we replaced... Churches add this special role in the life of people, where they would meet there and talk there, and that when we replaced churches, we also need to replace all the other stuff. A bit like with education, when we remove schools as a way to learn, we also need to replace the need to have the kids out of your room. I think that's my parallel, I don't know how much I agree with it.
Sav: Yeah. And of course, it's always much harder to do than just to say, right, because you're dealing with this complex system, right? Society is a complex system, and it's built on hundreds and hundreds of years of just us tinkering and trying to figure out what works best, and so, one of the ways you can think about tradition, for example, is traditions are kind of like a hotfix to a particular issue that that society had, right? And then as it gets carried through the ages, you basically... You're left with this practice, but you forget why you started the tradition in the first place.
Michael: Yeah. I think one important thing to note is that this tradition... So all of this starting to... We start from being molecules on a random rock, and then there's evolutionary pressure after years, and then this evolutionary pressure biases our brains towards social approval for optimizing, replicating those genes, and surviving. So I guess it's worth stressing that old traditions, in my view, evolved from a kind of some evolutionary psychology, where we're trying to optimize in fitness. So in a sense of religion, it would be some kind of... Our beliefs in a better future would make us happier and make us more connected, so we're more able to protect from other tribes. I don't think-
Sav: Interesting. Yeah. I guess you can look at... Well, there are many different causes for why traditions come up. And as you just pointed out, is well, I guess, the sort of social cohesion and sort of togetherness and everything plays a big role in why religion... or just traditions, in general, still stick around. But there's some quote, I forget what it is exactly, but it basically says that some... Tradition is the solution to problems that you didn't know you had until you take away the tradition.
Michael: Oh, right.
Sav: And so, it's like, that's another... Parts of this is like... So on the one hand, you're trying to kind of create this sort of social group, a social environment, on the other hand, you're literally just trying to construct a society in a way that is actually stable and doesn't fall apart. And so you create a bunch of these hotfixes that you call traditions, and then you forget why you've made them in the first place, and that... You feel that because there isn't a reason for why these things were there in the first place, that kind of gives you the license to take them out because you don't need them. And this whole like, well, there's no obvious reason for why this is there, therefore, we can take it out because we don't need it. That kind of should, I guess, remind you of 20th-century medicine, and how many... Like, oh, we're going to delete this sort of vestigial organ, because we can't think of a reason for why it's there, or like, we're going to go and lobotomize people or something.
Michael: I think that's precisely what the Chesterton's fence is about. It's exactly like having this fence, and after society evolving, we don't really know why there's this fence, and we want to destroy it because we don't see any meaning on this fence. So the basic kind of the idea behind just a fence is just to pose and understand why do we have it here?
Sav: Exactly. Well, and it's also like, maybe there is no need to have a fence there, right? That's actually a really interesting problem. Maybe we can actually remove this fence and there won't be any consequences.
Michael: We were thinking about it before. I think... I don't have much more thoughts on religion, because I haven't thought much about it, but I think... I guess, there's this kind of perspective of shaping the society in a way that optimizes better for understanding, meaning, and acquiring skills that make us happy and improve the world, and in a sense, I meant the parallel between... When we removed churches, we also need to fulfill all those needs. And I think today, with the pandemics we've been experiencing, there's also this shift between, I think, kids in schools, and we're understanding that schools are not just about learning. They're about all the things I mentioned before, social connections and credentials.
Michael: I think you have very interesting perspectives about education and how to change and reform education. So how would you transition from... It's very hard to change stuff like religion, to give meaning, and then transpose that reasoning to... How do you see how we can change education. I think you were students still, and so you have maybe a bunch of zoom calls and understand that there's more to learning, to education.
Sav: Yeah. So, yeah. I guess my frustration with education comes from a couple of different places, and we can certainly get into that soon. And it is an interesting kind of parallel that you draw between trying to replace the churches and sort of trying to replace the universities because in a way-
Michael: They're forged by the same thing.
Sav: Well, yes. And even today, the universities are kind of like a state church, especially if you look at the way in which every university basically has the exact same positions on things as every other university, they... Yeah. It's like...
Michael: You mean things like, oh, COVID is not a bad thing. You should definitely go to school and pay one more year of...
Sav: Sure. Stuff like that or stuff like, specifically on this day, we're all going to decide that we have a big problem with racism, so we need to fight this racism with antiracism, and somehow, all of these different institutions that are independent of one another, all come to the same conclusion at the same time.
Michael: Well, maybe that's more like cancel culture and society as a whole than... I don't see... So do you see a difference in how fast universities adopt social constructs from cancel culture than other institutions, or is it like most institutions try to be like... kind of protect themselves from cancel culture?
Sav: I think it depends on the institution. I don't think construction firms care very much about all of these different social issues, right? I think it depends on... I don't know very much about this, this is another thing that I have to kind of... that's on my list of things to learn about, but I think you can basically look at which institutions belong to which classes in society, right? It seems like universities, newsletters, or news publication companies, and tech firms and all of these things, they kind of... The people who inhabit these places kind of belong to the same social class, I guess you could say, right? And because of that, certain ideas propagate within these places that wouldn't propagate your average construction firm. I think in effect, all of these things kind of function like state churches, but... Education isn't limited to it.
Michael: So there's a segregation... Okay. So I guess, I think you're pointing at, somehow, some echo chambers, at division between classes, where it's hard to climb the social ladder, somehow, and social reproduction... it's called like this in the U.S., I don't know, but when you're a kid from a teachers, then you're most likely to read books and maybe apply to Ivy League schools. So, yeah, maybe there's society cleavage and differences, and clusters, where we can't really move between those. But maybe we can just stop thinking about those clusters and society as a whole, and focus on, especially education. So you have this cool stuff. I think your post is called The Future of Learning, so if we only focus on the learning part of education and not all the societal stuff...
Michael: You have this really cool picture about the wall of mysteries. Maybe you can summarize it, or I can just give my one sentence summary. It's like, if we accelerate and speed forward towards 10 years after, in 2030, and we have this big wall with all the stuff in science that we've discovered, or even in history and other fields, what would it look like, and how much could you optimize right now on your connections and trying to find good mentors to better fit into the society? Did I summarize it well, or I didn't?
Sav: I'd say so. Yeah. Yeah. Well, I guess the point that I was going to make before going in that sort of tangent, was that, well, if you draw this metaphor between actual churches and the university as a church, if you get rid of the university, in the same way, you're going to have to think really hard about what you replace it with, and there's probably going to be a lot of, sort of unintended, kind of problems that pop up, because you just didn't anticipate certain things that the university was doing, certain problems that it was solving. So, I guess... Just like to preface that Future of Learning post by just saying that well, there are probably hidden things that the... sort of hidden problems that the university has solved that I don't really know about, but if we were to kind of... My whole idea with that wall of topics and sort of wall of research areas was meant to kind of solve the problem of...
Sav: Like you get to a university, and then you, I guess if you want to find out about the world, and if you're eager to get out there on the front lines of some kind of research field and you're eager to contribute to that stuff, you get to the university, and basically get told, at least this was my experience, but you get told to specialize, right? And...
Sav: You get told to specialize. Right? And so it's like instead of looking at this big sort of wall of active projects, imagine you're in a... You enter this big makerspace. There's all of these tools on the wall, and you can sort of explore all these tools. You can go make whatever you want. And then this guy comes up to you and says, "Well, actually. You're just going to be using the hammer. And they're very useful things that you can do with the hammer. So if you sort of properly train yourself in using a hammer, you'll be useful to people who actually need somebody with hammering skills." And that's kind of... I don't know. I didn't think that was a very healthy way of taking advantage of sort of people's skills and people's interests.
Sav: And I thought if you have this big sort of actively updated list of all of the ongoing problems, kind of like, I guess, a GitHub Issues, but for science, right? You just have the freedom to choose, "Oh, I'm interested in solving that problem, but I want to now get introduced to the people that are working on that problem." And so it's a much different approach to this kind of like, "We're going to mold you into a tool and then go find yourself uses." This is more [crosstalk 01:03:39].
Michael: [crosstalk 01:03:41].
Sav: You can do this now. It's just that you can't do this in an institution. You kind of have to do by yourself. Well, and that's kind of what I've been, or at least in my experience.
Michael: I mostly am agreeing with what you're saying, and I think all of your argument makes sense. If I'm trying to be the devil's advocate and make the steelman for the opposite position, there are several points. One is if you're just being a researcher, doing your PhD and trying to write a thesis, I don't know, in neuroscience, let's say. You can obviously reach out to other researchers, reading their papers, and connect them today. And so you can kind of not specialize and make a very general thesis. And so I know the point is that, which is a bit different point, is that when you're saying that we tend, there's some incentives to make us more specialized. It's a very specific education, which is maybe grad school and PhD. And this is because, again, there's this bias of people that teach us what to do.
Michael: I've been through this route of academic research. So if you're in grad school, maybe your teachers are doctors. And so there's this kind of reproducing loop where there's academia, kind of preaching their religion of science, and people following them and becoming the new priest somehow. But also, in power of that academia, there's a society maybe in tech, people understanding that we don't really need that specialist knowledge in today's economy because we have the internet, and we can Google stuff and when we need to. So I'm sorry. A big shift that I see my country is programs, master programs, or engineering schools making explicit courses about how to learn and how to adapt to different situations, how to make the best use of internet. And obviously, it is very slow change. Those are very slow changes compared to how fast we need to actually go to catch up with technology. But at least, there is some kind of tendency towards, "Oh, maybe we should learn how to use the internet and not how to learn by heart medical textbooks."
Sav: Yeah. I've noticed that as well, especially in... I don't know, if you look at tech companies, more and more tech companies are kind of hiring people based on, well, how good can you code? And not necessarily if you have this very specialized knowledge that you've been taught. But I don't know. I don't have experience with the graduate program. I only have experience with undergrad and specifically undergrad at the University of British Columbia. But I did kind of notice this, at least in engineering, what happens is that you get more and more narrow in what you study, more and more specialized as your degree progresses.
Michael: Sure. Yeah. I just want to counterbalance your point about specialization with your own story. I'm sorry. And I'm treating you with your own line. Where the last conversation we had, you told me that one thing, one particular project you did, I don't remember if it was in high school or in university, was actually building robots. So some kind of robotic competition every year where you had six weeks to build a robot from scratch. And at the moment of this competition, which was project based, and maybe where you learned the most, it was mostly horizontal learning. You have to learn to do everything. So essentially, okay. So it's true. Society optimizes, university optimizes for a specialist knowledge. And they ask you to specialize to do good in academia. But surely, you discovered new ways of learning horizontal by project based. And I think it's interesting to just point out that what society optimizes for, to make listeners happier and more productive, maybe we can discuss ways of learning that you learn that is much more efficient than university, I think that could be helpful for me.
Sav: Yeah. Yeah, yeah. And the point about [crosstalk 01:09:07].
Michael: And maybe you can explain, summarize maybe, or you can go in depth into your robotic experience. Because I think that's one of the most valuable experiences you have. I don't know anything about your life, but this sounds very impressive to me.
Sav: Yeah. It was very fun to do. I would probably describe it as one of the highlights of my life. Yeah. So basically, what FIRST robotics is for those who don't know about it, it's...
Michael: Maybe we can ask you that people know... Maybe you can define the more precise way. Go ahead.
Sav: So FIRST, is this nonprofit organization that is very popular in the States, and to some extent, Canada. I don't know how popular it is in Europe, probably less so. But basically, it's this organization that...
Michael: What's it called?
Sav: It's called FIRST. And it was founded by this guy, Dean Kamen. Dean Kamen actually invented the Segway.
Sav: So if you've seen people ride around on those, well, now you know who invented it.
Michael: I'm not sure how useful it is. But, yeah.
Sav: Yeah, yeah. Well, I mean, I don't know. It seemed to be trendy for a certain amount of time. I don't know.
Michael: My first LessWrong post - let me Google it - I think my first LessWrong post that I wrote three years or four years ago, maybe three years. It's something about the Segway. So it's called Schule and AGI build a telescope to spot intergalactic Segways, where I kind of argue that Segways are something very specific to humans. And should we kind of try, as humans, should we try to... When we learn about other lives in the universe, will they have Segways or something similar to Segways? Or are Segways very specific to humans? The internet and computers are... You could find something similar in other species. But Segways? Come on.
Sav: Yeah, yeah. They're very specific to Dean Kamen. Maybe if that other planet has a Dean Kamen equivalent, then they'll have Segways. I don't know. Yeah. It sounds fun. I'll just check that post up. But, yeah. Dean Kamen, the guy who invented the Segway, goes off and... Amazing guy, creates this nonprofit that puts on robotics competitions for high school kids. And all the way from, I guess, from very young grade school up to high school kids. And there's different competitions for different age groups. And every year, a new game gets announced. So completely new rule sets. And then from January 1st ish up until mid February, you have to come up with a design for a robot that can play this game.
Michael: Can you give examples?
Sav: For example... Well, so the three years that I participated, the first year involved... It was this theme of recycling. And so you had these big plastic crates that you would have to stack on top of each other, and then you'd have to put a recycling bin on the top of those crates. And so you'd make these big towers and...
Michael: Of plates?
Sav: Of crates. Crates.
Michael: What's a crate?
Sav: Big plastic boxes.
Sav: Like a box. Yeah. And, yeah. And so you basically get... It'd be a 3v3 match. All of these matches are 3v3 matches. So you have your team, with your big 120 pound robot, on an alliance with two other teams competing against three teams. So it's 3v3.
Michael: [crosstalk 01:13:48] robots.
Sav: Pretty much. Yeah. And it's whoever could stack the most boxes would get more points.
Michael: So, yeah.
Michael: I'm just not understanding... Maybe it's obvious for you but... so there's two towers of boxes. And you're just... Is there... Can you hit the other team's tower, or do you just need to move faster than the other? And it's quite independent.
Sav: So that year, you couldn't actually cross to the opponent's side of the field.
Michael: Mm-hmm (affirmative).
Sav: So in that sense, it was kind of, I guess, a little bit of a boring game because you couldn't actually go knock over people's towers.
Michael: How tall would the tower be? Because I guess robots cannot just reach out their arm independently, right? So I would guess maybe five meters tall or something.
Sav: Yeah. I don't remember exactly in terms of meters, but basically, most...
Michael: Oh, sorry. Not meters.
Sav: I mean, it's... I guess we use the metric system, but FIRST uses the imperial system because they're from our neighbor down to the south. But basically, most robots would... There'd be a... They'd stack from the bottom. Right? So they would pick up a box, and then they would somehow have a belt that pushes it up so that they can pick up another box. So the stack, you could actually get it arbitrarily high, as long your mechanism was powerful enough to lift the whole stack, to put another box up there. Right? So you wouldn't...
Michael: Oh, right.
Sav: It's like a different way of stacking.
Michael: You need to put it from the bottom. Okay, so it's like your weight [crosstalk 01:15:40].
Sav: Yeah. Yeah, yeah.
Michael: And how... I guess you said those are kind of plastic boxes, so it was kind of light, maybe?
Sav: Yeah. But they were big, so maybe several kilos.
Michael: Wow, using the metric system.
Michael: And how many feet tall? How many inches?
Sav: I'd say somewhere... Maybe, I don't know, three meters. But the point is, every year, it's a new challenge. So this year, it was boxes, but maybe another year, you have to collect tiny balls and shoot them into a goal. Or it's a combination of things.
Michael: Like a basketball? I'm sorry, [crosstalk 01:16:40].
Sav: Maybe a basketball. Maybe a basketball. Maybe something bigger than a basketball. Maybe a Frisbee. Maybe there's an added challenge in the last 15 seconds of the game where a rope comes down from the ceiling, and your robot has to climb the rope. So it's just this crazy, big diversity of engineering challenges that you have to solve.
Michael: Wait. So is there engineering problems during testing at the end of the competition where you just compete with other robots? Or do you know all the rules of the game beforehand? Or do they say, "Oh, you need to build robots that can throw stuff." And at the end, you're trying to build the more general programs. And then you're like, "Oh, this environment is very adversarial." And you throw ropes at other robots. And you have no idea. Was is obvious or not obvious, the last rules at the end?
Sav: So you know the rules in advance, and the rules don't change. I mean, sometimes they make little updates to the rules where if there's some kind of mistake, or if people figure out a way to exploit some kind of rule in a way that they didn't make intentional, they're going to be like, "You know what? You guys can't actually do that."
Michael: Mm-hmm (affirmative).
Sav: But for the most part, people know the rules in advance. But even then, I don't know. You basically go... Over the course of six weeks, you go from not having a plan whatsoever because the game was just announced to a complete robot. So design done, manufacturing done, everything assembled, the code done. Because at the end of the six weeks, I don't know if this is still the case or not, but at least back when I was doing it, you would put away this robot in a big bag after the six weeks. And you couldn't touch it. So you put this ziptie on it, and you sign it and promise to not touch it until the competition.
Michael: You just have this nice curl or wget in your robot that can download from GitHub your update?
Sav: Yeah, yeah. Yeah, pretty much. Well, you can still work on the code, but you can't work on the hardware.
Michael: Oh, okay.
Sav: So the rules don't change, but it's still kind of hectic because you have to do all this in six weeks.
Michael: Oh, sure. And I guess my intuition about this competition is it comes mostly from kind of competing in mathematics. And I also competed in Go and competed in tennis and also for coding interviews and so on. And so I feel there's like... So what you're saying... And also, I also did a bunch of hackathons, where you're trying to do a project in a few days. Where... Sure. I agree that if you give yourself a specific subject, then we can kind of measure the skill acquisition efficiency from Chollet. Right? And you see how fast people can learn new domains in robotics. But I would kind of guess that robotics as a whole is maybe five big themes, subjects. And if you're good at those five ones, those ideas, then you can go into a sub field, then you would kind of know how to handle... Knowing how to handle boxes may be one of the most mainstream things that robots do.
Michael: I have no idea. But, yeah. So for this kind of competition, let's say math, of course, there were mathematicians where you could throw any geometric or algebra problem. Right? But then knowing good equations that were mostly useful in most problems would help you go much faster. So how do you prepare for it? And did you learn general things about robots? And also, please explain all the details about the robots. Did you win? Did you lose? Were you happy with it? And I don't know. What's your proudest achievement, maybe the coolest robot you built in these competitions? I think you did three of them, right? Or two?
Sav: Three, three. One on somebody else's team. And then I kind of went off and did my own thing and actually founded my own robotics team and built two robots with my own team. And that was my high school life. That's what I spent my time on.
Michael: So two robots? But then it's only six weeks, right? So you need to spend the rest of the time learning about robots and building ground, right?
Sav: Yeah, yeah, yeah. So six weeks to build it, but then you basically, the next two months are competition. So you end up going to various different events. We were lucky to have one in our city in Calgary, but most people go to than one event. So we flew our whole team down to... Well, in the very first year I did it, we flew down to Salt Lake City for a competition in Salt Lake City. And then in my last year, my grade 12 year, we actually flew down to Arizona, to Flagstaff, Arizona and competed in the event there. Because of course, the more events you compete at, the more chances you have at tweaking the robot and the more chances you have at winning. And if you win, then you get to go onto the world championship. So if you win a regional event, then all the people who won regional events go to the championship.
Sav: And we made it to there twice.
Michael: Oh, you made it to the world event?
Michael: Wow. That's super impressive. And so how many countries are there?
Sav: A lot. So most teams are in the States. There's a lot of teams in Canada, but then you also get teams... I remember we had teams coming down from Mexico, from Turkey. There were a couple of teams in the Ukraine, one team from Israel.
Michael: So how many tournaments do you need to kind of win or achieve top score to go to an international one?
Sav: Well, you actually just need to win one.
Sav: And most of these events have 30 teams competing or more. 30 to about 40 or 50 sometimes in the big ones. And so you can either get to the championship by winning the tournament, or you can get two awards, either the chairman's award or the rookie all-star. The rookie all-star award is basically the best first year team, the best rookie team. And the chairman award is kind of the...
Michael: The beautiful, the public's...
Sav: Well, it's... No, no. It's more how much did you do for the community? And how much did you like spread the gospel of FIRST in the community? So people would talk about the amount of money they raised for charity, or maybe they opened up some kind of summer camp for kids where they teach them things about robotics. And so that's also a way to get to worlds.
Michael: Well, I'm super impressed that there's the solution, either build the best robot at building towers or be the best human at giving ethical speech.
Sav: Yeah, yeah. Yeah, exactly.
Michael: It seems very... I guess it's typically an Ivy League... To apply to an Ivy League, in my opinion, is like, "Oh, yes. Yes."
Sav: If you don't have the top grades.
Michael: Yeah. No, no. If you have a very interesting life and you raise money for charity and those things, and maybe your cover letters are really, really good, and you will enter Harvard. But then you can also pretend, kind of raise money for a kind of charity in kind of a sneaky way to optimize for just winning this robotics competition. So, yeah. I'm not sure how much those people were genuine in their speeches. So, yeah. Do you have any takes on that?
Sav: Well, the FIRST community is very good and very friendly.
Michael: Oh, okay.
Sav: And that's one of the sort of best things about FIRST is just the quality of people who are in it. So I think, for the most part, people were genuine. Of course, you tend to get the cynical team or something, where they just have, for example, big egos and don't really want to help anyone. But that kind of thing is fairly rare. Most... You have a lot more people coming to these events than you really need to actually fix the robot. So you have a bunch of these high school kids there that don't know what to do. And so everybody just goes around and helps out each other.
Michael: That's so cool.
Sav: So, yeah. But, yeah.
Michael: Did you help people also, or did you mostly work on strategy and how to build a better robot?
Sav: Me personally?
Michael: Of course, you personally.
Michael: How many people...
Sav: And that kind of laser focus, actually, it was just... 12 in the first year, 30 in the second year.
Michael: Wow. So you were managing 30 people to go to international events?
Sav: Ish, but these are high school kids. And so you can imagine the power law, the 80/20 rule of who actually does the work. It's 20% of the people do 80% of the work. It's kind of what we got on this team as well, where we had maybe the core team of six people working on this thing. And then you kind of see this drop off. And that's kind of whatever. That's to be expected. So, yeah.
Michael: And you always went to the way of winning the competition, or did you also do the other ways? Did you also win first on a 30 ish team competition to go to the international events? Or did you do where do you get the speech thing?
Sav: Well, so in the first year, we won the Utah events, which let us go to worlds. In the second year, when I founded my own team, we won the rookie all-star award for a rookie team at the event.
Michael: Oh, wow.
Sav: So that's how we got to go there.
Michael: And you were a second year, so not so rookie?
Sav: Well, I was a second year doing this competition, but it was the first year of us as a team.
Michael: Oh, okay.
Sav: I went off and started my own team. And the third year, we ranked fifth or fourth, I think, at the regional event. But unfortunately, didn't win it. So that last year, we didn't get to go to worlds. But paradoxically, we did better that last year than we did in the first year of our existence.
Michael: I guess there's some... There's a way to some stochasticity. And it also depends on other teams being better. And so it's not only about your team being skilled, it's also... There's some randomness about the particular project. Okay. So I'm sorry to kind of change subject or do these tangential things. But I have watched the movie 21, multiple times. I don't know if you've seen it. It's called Blackjack maybe, the movie.
Sav: I don't think I have. Can you give me a brief?
Michael: So the brief thing, it's just a guy at MIT that is supposed to work on a robotics competition. But instead, he learns how to play blackjack. And so he goes to Las Vegas to play blackjack and learns blackjack from counting cards instead of playing with the robotics competition team. So, yeah. If you have seen it, then it would be fun to talk about. But if you've not seen it, it's just a kind of meme, a very bad meme about robotic things, where the whole of his friends at MIT are very kind of nerds. And he has two lives when he's talking to the nerds who are designing their own chips and writing software and so on. And the other is counting cards at Las Vegas and winning millions of dollars.
Sav: Right, right.
Michael: So, yeah. In my kind of mainstream cultural memory, I just have this just video of those nerds. And that's why I think it's good to hear from the actual people doing those things. I've watched this movie three times. I think that's a problem.
Sav: Right. Yeah. I'll check it out. Yeah.
Michael: Yeah, don't. Don't. But, yeah. So, okay. So first year was building towers. Second year, it was balls. Third year was what kind of task?
Sav: So second year, actually, was basically, you had these dodge balls that... It was a medieval theme. So you had these... The balls were kind of boulders. And so you had robots that had catapults that would shoot them into [crosstalk 01:31:57].
Michael: No way!
Sav: And you also had, basically, these...
Michael: How far? One yard?
Sav: I don't know, maybe two meters. You really had to get up to the tower to shoot the ball into the tower. But in order to get to the tower, you had to go through all of these different obstacles. And so kind of, if you think about a castle having different defenses on it, you had different obstacles that you had to have mechanisms on your robot that would let you get over the obstacles. So for example, there was a bridge that would like come down. And so you have to somehow reach the tip of it and pull it down, for example.
Michael: How do you even start programming the software? Is it reinforcement learning? Is it just basic instructions of... Do you need to solve? How does the competition problem solved? What's the algorithm?
Sav: Well, yeah. So it's really like...
Sav: Well, yes. So it's really like... the software kind of goes... it's really as hard as you want it to be. Most of the game is actually played with human players, like drivers controlling the robots, but the first 15 seconds of the game is autonomous. So you have to just basically have the robot by itself, carrying out certain tasks. For example, our team didn't spend too much time working on the autonomous mode. So we have a simple algorithm that basically used... I guess the simplest thing you could do is just count ticks and run the motors for a certain amount of time and then stop, and then rotate. A step up from that would be having sensors on the robots and having potentiometers, for example, on your motors, we keep track of the rotations. And so instead of saying run the motor for a certain amount of ticks, you just say run it for a certain amount of rotations. That's more accurate. And have the- [crosstalk 01:34:18]
Michael: And you need to know how far is it the tower or the obstacles, right?
Sav: Yep. Yeah. Yeah. And so this kind of gets us into the third, most advanced approach, which would be, you have some kind of computer vision recognizes the hole in the tower. So usually what they would have, is that they have reflective tape on the goal, on the top of the tower and you shine the lights at the reflective tape, and then you basically use that to identify where it is, and then you- [crosstalk 01:34:53]
Michael: I'm very bad at physics. I'm very bad at physics but if you have a tower and then on the top, you have some reflective tape. Then when... if you throw a laser, then it will go up right? Not down. Or maybe there's some reflection that goes down also?
Sav: Yeah. They're just like... you don't need all of it to reflect. You just need a little bit of it to reflect [inaudible 01:35:11] [crosstalk 01:35:11] for the object. Right. Or you can go completely crazy and build an AI that sort of drives around the field and the most... the best teams in the world did this. So I remember in that game, there was this team, they're called the Cheesy Poofs. Team 254.
Michael: Cheesy Poofs.
Sav: I think it's a south park reference or something like that. But, they had... they have-
Michael: Wow, they have a website.
Sav: They do have a website. I'm a huge fan of these guys. This is the best team in the world, super impressive. So they would basically get a ball from one side of the field, go across the obstacle and shoot it. So, that was their self-driving autonomous mode. However, there was a point where the robot lost the ball and it rolled off somewhere. So it stops, turns around, it goes, finds the ball comes back, and then it gets over the obstacle and shoots the ball. So, I was like... One of those, just really impressive things, but you get to see a world that you don't get to see it in regional competitions.
Michael: I think the most impressive thing for me, from working in computer vision at the moment, is that papers about object detection and learning, being efficient, at least in training, and with online data coming in as is, is kind of recent. And maybe not as much like the competitors more, maybe more like 2012 than 2020, but you're saying that this happens somehow, like in high school, so maybe like 2016, 2017. And I think the first time RetinaNet, so the object detection kind of first breakthrough. Not the first breakthrough, but one of the biggest breakthroughs during the past five years, I think it's like 2016 or 2017. So it's pretty impressive to have something in 2016 already using learning well, and I'm mostly curious about the kind of the constraints on energy. Did it last like five minutes to throw the ball or one minute? So, did you have constraints on, "Oh, I should not use too much GPU on my little machine or something?".
Sav: Its something like 15 or 30 seconds.
Michael: Mm-hmm (affirmative).
Sav: That's how long the autonomous mode takes.
Michael: Mm-hmm (affirmative).
Sav: [Crosstalk 01:38:21] I don't know. Sorry, what's that?
Michael: When you drive around by yourself with your joystick or something?
Sav: Yeah. So, the match is, if I remember this correctly, something like two minutes or like two minutes 30 seconds. And then the first 30 seconds of the match is autonomous mode. So the robot drives itself. You can't touch the controller. After that autonomous mode is up, you get allowed whistle and that's when everybody rushes up to the controllers and then they pick up the controllers and start driving. I don't know how Cheesy Poofs did their autonomous implementation at the time, but they're this team out of Silicon Valley. They have mentors that work at all of these big Silicon Valley companies. This other team, 118, for example, they get sponsored by NASA and actually, they get to build their robot out of NASA hangar. So NASA gave them a whole hangar-
Sav: where they can just work out of, and their mentors are NASA engineers. So, that's pretty convenient.
Michael: Okay, okay. I guess I'm not that surprised that you... your team in Canada... I guess I don't know how many mentors you had, but I guess maybe less than, you know, NASA engineers to competition.
Sav: Yeah. Very, very tough. Although it's pretty... if you know what you're doing, it's not very hard to win a regional event. And the level of competition at regional events varies a lot. Right?
Michael: Mm-hmm (affirmative).
Sav: If you go down to San Francisco and try to win a regional event there. You're probably not going to make it, but if you go down... I don't know if you're in Calgary, Alberta, or if you go down to Mexico or... you pick the events that you compete at. [crosstalk 01:40:29] Kind of a strategic game as to which events you choose to go to and how many events.
Michael: Surely, you're you're not the first one to think about going to Mexico.
Sav: Of course not. So you're... The system kind of balances itself out, if you go to San Francisco, you're going to get your ass handed to you.
Michael: Sure. I guess Americans wouldn't go to Canada, by principle. So you kind of can win in Canada.
Sav: Yeah. Although, you... we got a few U.S. teams come up, but not very many.
Michael: I'm curious, by pure... I know it doesn't reflect how strong your team were, and your abilities or skills in any way, but I'm curious, how did you rank in the World's events? And if you're happy with it? Did you make top 50 or something like this, or....
Sav: I don't remember exactly how we ranked in the world events. I remember we didn't rank very well.
Michael: Its a tough competition, so it's quite understandable.
Sav: Yeah, yeah, yeah. [crosstalk 01:41:52]
Michael: All won their original events. So they're literally the best teams in the world.
Sav: Right. Yeah. And, and so one other detail that matters here, with who wins the events. The way that it works is... so you have all of these teams play qualifying matches. So you get you and two other randos playing against three randos. You don't get to pick who your teammates are. And so you play all of these matches. Maybe you play 5 or 10 matches or something. And there's a total of 80 matches that happens over the course of several days. But at the end of that, you kind of have this ranking of who scored the most points, or who won the most games. And so you have this ranking of 1-40, let's say. If there are 40 teams competing and then the top eight, get to pick two more teams to be part of their alliance. [crosstalk 01:42:57]
Sav: And this is for elimination matches. Yeah. So the way that quarterfinals, semifinals, and finals happen at these regional events is that, now that you've got this ranking, the top eight teams now get to pick two more teams. And there's all of this strategy that goes into, okay, which teams robots compliment your robot the most and how can you pick-
Sav: the most effective alliance? And so the way that the selection goes is like-
Michael: So Meta.
Sav: It's crazy.
Sav: Well, strategy is probably one of my favorite things to do there. You'd have to have an entire subsection of your team devoted to scouting. To basically looking at the performance of other players, robots-
Michael: No way.
Sav: To gauge how they compliment yours. So that if you were in the top eight and you're in a position to pick your teammates, then you know who to pick. Where if you're not in the top eight, then you can go strategize with one of the people at the top 8 and then make your case for why they should pick you.
Michael: Did you strategize to be the best compliment for team 254?
Sav: I wish. Well, yeah. So, I think they were competing down in California, and I think they did an event. They basically go all over the states.
Michael: In the World's event they were there. Right?
Sav: In the World's event they were there. The World's event... So, the World's events is actually split into different fields. So different divisions compete with different teams. So, if you made it to the World's event, not necessarily you're going to be playing against all the other teams, because there's just so many teams there that they have to split them into different divisions. And so you win your division of World's and then the winners of the divisions go and face off on one final field.
Michael: Oh okay.
Sav: The point I make that whole tangent into how the pick system works is that, it's the Alliance that wins the championship. It's not necessarily the top three robots to win the championship. So-
Michael: So you can just lose at the beginning and then go to the better Alliance.
Sav: And get picked by the best team, and then win the event. Exactly.
Michael: Surely the best teams will choose sandbaggers, like people who were playing badly, but lost, but then very good for... I don't know what even... qualities of robots, like oh they're fast robots or they're light, or they're good at throwing balls. How did you try to optimize for being pick-able by any other big teams, like having a bit of any skills? Well, what even are the skills?
Sav: Yeah. So see, now you're starting to pick up on it. There are quite a few drama... there's so much drama behind the scenes, with regards to sandbagging and just secret agreements between people. It's this bottomless pit, but-
Michael: So the Meta-Game is politics and money.
Michael: Object level?
Michael: Robotics, like science-
Sav: naive. Yeah. Like robotic science level. You absolutely think about how pick-able you are when you make decisions about how to build your robot. So let's say there are three big challenges in a game. So, that last year that we competed in 2017, I would say it's actually my favorite year out of the three, basically the three challenges were... it was a steampunk theme. So the goal was basically you have... each team has their own Airship. And so you have to get this Airship off the ground and quote unquote, fly away. Right? And so there are these small plastic wiffle balls that you would have to collect and rapid fire them into a goal. And that goal was like the fuel tank. And so you're filling up the fuel of the Airship.
Sav: And then they're these big plastic gears that you'd have to also pick up with your robot and carry to the Airship. Then you'd have a human player actually pull the gear out of your robot and put it in a part of the tower that was symbolizing the Airship. So basically you have to collect these balls, you have to shoot these balls. You also have to collect gears and you have to carry them to certain locations. And at the end of the game, there'd be these ropes that come down from the Airship and your robot would have to climb up. So it's... Imagine this 20, 120 pound thing, giant metal box pulling itself up a meter into the air. And so it's these three tasks. Right? Okay. Of course, if you're-
Michael: Pulling from a rope, you just... you grabbed the rope and you try to escalate like a human would, or what do you do?
Sav: Well, that depends on what kind of rope you have. So you can actually design your own rope as well. So what we ended up doing for example, we had a Velcro rope, and so we had a little Velcro drum on our robot that [crosstalk 01:48:58] was attached the Velcro rope, like a... I don't know how to... you know that material that has loops and then another that has hooks, and then it just kind of sticks?
Michael: Oh yeah, yeah right.
Sav: Like it, people use it in shoes or... So, basically, we had this rope come down and this drum would catch onto the rope. And then just basically the whole robot would get pulled up by this one motor on our robot. Others had different mechanisms for doing it. And so you have these three tasks, right? So collecting balls and shooting balls, picking up gears and delivering your specific areas, and climbing up the rope. Right? Those are the three big things.
Michael: Do do the gears help for anything? Or just like-
Sav: Yes, so all of these things give you a certain amount of points.
Sav: And there's very, sort of complicated interactions where if you deliver X number of gears and you climb, and all three teams on your lines climb the rope, then you get some kind of... Points multiply. And so it's this complicated strategy that goes into it.
Michael: And you were managing 30 people to get this.
Sav: Something like that.
Michael: Three teams of maybe 30ish people trying to... Would they be interacting with the other team or would it be just-
Sav: Oh, absolutely.
Michael: Wow. So it seems very different from the first scenario where you just build a tower, now you just have this crazy StarCraft game with a hundred people trying to- [crosstalk 01:50:50] Its crazy.
Sav: Yeah. I mean, it's usually not... You have the top three representatives from each team and you coordinate this stuff. You obviously don't try to involve your whole team in the decision making process. That'd be crazy, but it's usually just the drive teams cooperating with each other. And before the game, you have several minutes to strategize and say, "Okay, I'm going to go do this".
Michael: Only several minutes?
Sav: Yeah. So in the qualifying matches where you randomly get matched with people, it's not a lot of time. In the elimination matches you actually get breaks where you can talk to them for an hour or so.
Michael: Even a Meta-Game would be talking to people before the game tournament.
Sav: Yes, yes, yes, yes. And that also happens.
Michael: I'm just curious about like... when you were saying that you started by managing 10 people and then 30, I'm kind of curious if you... in those three team leaders, they might be some kind of team leader of the three teams. Right. So did you try to become the leader or you naturally emerged as the leader? Or was it somewhere else?
Sav: I mean, it really depends.
Sav: So if you're stuck with two teams that don't know what they're doing and don't have any plan, then obviously somebody has to make the decisions. And so I would find myself naturally stepping up to that role. On the other hand, you can run into situations where you just have a team that's just so much better than you and you're playing with them. And, and at that point it's like, well, you should really just defer all the strategy to them and just do what they tell you to do. Not to say that it always works out cleanly and perfectly in these ways. I certainly got into my fair share of shouting matches between different teams and just... it gets pretty intense. A lot of people there have a very big ego's. [crosstalk 01:53:10] Sure. And you know, a typical drive team consists of two drivers. So you actually have two people at the controls of the robot. So one drives and then the other one operates the shooter or something like that. And then you have a drive coach, the drive coach can be either a student or a grownup, like a mentor, and a human player, who does some kind of human player task.
Michael: So did this kind of NASA guys would come and pilot the robots?
Sav: No. So, the students would have to be the ones who pilot the robots, but you can have a NASA guy be the drive coach. So the person who mentors the drivers.
Michael: Come on. Yeah. You just have Novak Djokovic telling me how calibrate your right-hand like, it doesn't make sense, but yeah, go ahead.
Sav: Exactly. So, I'm just saying like sometimes, I found myself arguing with kids, which is fine, but sometimes you have to argue with adults, and the adult drive coaches have probably bigger egos. And so it doesn't always cleanly split into "Oh there's a"... everybody knows what to do, and everybody agrees. [crosstalk 01:54:31] It's fun. It's fun to-
Michael: [crosstalk 01:54:35] Did you have... Before, I get the sense that maybe you experimented with other things after high school that made you... Maybe like those kind of psychedelics we talk about, or like other ways of learning in books and stuff that made you very humble now? I guess my picture of other high school students that win robotics competitions or something, might get in the way of earning some ego, even if it doesn't want to. So did you actually become very confident and add an ego? Or did you have to fight this person in you?
Sav: Ego's an interesting question.
Michael: [crosstalk 01:55:25] confidence- [crosstalk 01:55:27]
Sav: and I feel well. Yeah. So, confidence is like the... if you're familiar with Russell conjugation - emotive conjugation, right? It's this notion of being able to... It's like having different words that mean the same thing on a content level, but emotionally mean different things. So it's like maybe one of them-
Michael: Oh ego is bad, its negative.
Sav: Is emotionally bad, negative. Right. Right. So, so ego is-
Michael: Its like techy versus nerd.
Sav: But confidence is good, but somehow they're both the same thing. Right?
Michael: Wait, They're not. Okay. Maybe just words, [crosstalk 01:56:00] but-
Sav: A lot of people use them to mean the same thing. So it's like, having an ego is bad, but having confidence is good, but really-
Michael: Okay, okay so my-
Sav: you can't really be confident without having an ego I feel like. [crosstalk 01:56:16] It's a balancing act between keeping your ego in check.
Michael: For sure. Maybe I have misunderstood the word ego. So it may be worth clarifying the definition here. Because when I hear it is like, pejorative word ego, I kind of hear people were not aware of their own calibration. So, I think there's self-esteem where you're trying to guess... To get the best estimate of what's your value.
Michael: But what are your skills? What can you do? And then self confidence and confidence in general, which is how, how can you talk to people and say, "Oh, I'm good at this", like how much can you appear skilled? And so there's the skill of appearing skilled. And then there's the skill of calibrating yourself and knowing where you are. Right? And the ego in a sense, is what I understand from it is, mis-calibration.
Michael: So people who have no understanding, not like did they have high confidence, but maybe bad self-esteem and bad calibration on their confidence. When you think about people with strong egos, they would not really listen to what you're saying, and just think they have the right argument. Right? So maybe I misunderstand, but for me, confidence and ego are somehow different. Maybe in mainstream culture, they're the same thing. But in my understanding, they are a bit different, but we're just arguing over words here.
Sav: Well, think about it this way. So you... There's the question of how much you believe in yourself, how confident you are of yourself and the things that you believe and the things that you value. And then there's, how those beliefs actually correlate with reality.
Michael: Mm-hmm (affirmative). Interesting.
Sav: Right. What I've heard a lot of people say is, " Well, you should aim to be exactly on that line of your beliefs correlated with reality". So you should be exactly as confident as your beliefs let you be. Right? Or your understanding of the world lets you be. But what I kind of noticed in robotics is that... and this gets into like... maybe you've read about Steve jobs. A lot of people say about Steve jobs that he had a reality distortion field.
Michael: Mm-hmm (affirmative).
Sav: Right? Where he would just say things and they weren't true, but he convinced himself and everybody around him that they were true.
Michael: Mm-hmm (affirmative).
Sav: And sometimes that led him off a cliff, but sometimes actually that led him to push through and persevere, and actually achieve these very impressive things. And so I would always... having too big of an ego is certainly a problem where, if you are way, way, way more confident than your beliefs actually allow you to be. But if you place yourself slightly above that line, where you're slightly overconfident, I feel like that's a much better position to be, because you'll actually end up pushing through and persevering even if you're wrong, because that energy will just kind of carry you. [crosstalk 02:00:04]
Michael: True, yeah.
Sav: If that makes sense. You know? [crosstalk 02:00:06]
Michael: I think there are a bunch of instrumental reasons to appear confident, especially when you're in a position of leadership. Like Steve jobs. Then, there's this whole self fulfilling prophecy where if you believe you're confident, or you believe you're strong, then people will think of you... will have higher beliefs about you, they will respect you more and help you more. Then you will become more the leader and so on. So I think you're right. It's a good trait to have- [crosstalk 02:00:53]
Sav: Its almost like you're distorting your reality a bit, and then you're letting reality catch up to your distortion of reality.
Michael: Exactly, it's known in the memes as, fake it until you make it. Sorry to bring that in. I guess in Twitter I read it also. A lot of people saying, "Oh, you know, confidence. I just realized that confidence is just a thing that you can decide to have". I read this once every month or something... people realizing this. I think I've spent a lot of time... So the problem with calibration is if you're mis-calibrated, then somehow.... sometimes you will achieve great outcomes. And then your model will be, "Oh, it's from update", and, "Oh, I'm great at this". And then you will have the defeat and you will be like, "Oh, no, I suck at this". And then you will oscillate between those two, and your mental health or your confidence will vary. There will be too much variance. So at some point you need to deeply understand what's your worth. Otherwise there would be too much problems and difficulties in managing your confidence. I don't know. How did you feel?
Michael: Okay. So to go back to your original question, did you feel that you become more confident after winning... going to the World's twice and then did you feel less confident after not making it to the World's at the last time?
Sav: Yeah, I would say that's accurate.
Michael: Where are you now?
Sav: I think that depends on what I'm trying to do. How confident am I that-
Michael: [crosstalk 02:02:59] 3:00 PM Canada time-
Sav: That I would actually have reasonable success versus, how confident am I, if I pick up some new skill. I-
Michael: Let's say the confidence in general. [crosstalk 02:03:12]
Michael: So confidence in your ability to learn about the world, meet new people, and have an impact in general.
Sav: I'd say that I'm reasonably confident in that. I think the interesting thing with robotics is that it's the... there's a framework in place. So your goals are very clearly outlined for you. And all you need to do is just design a robot to the spec of the competition, show up to the competition and win the competition. So it's the... your path forward is very clear. And-
Sav: Your path forward is very clear. And I guess one thing that I sort of underestimated coming out of that experience and I'm only recently just kind of coming fully to grasp this now is, like how much in the real world, the path forward is not clear at all. And you kind of have to come up with your own path almost-
Michael: Or meaning, somehow.
Sav: ... You find your own meaning. Exactly. You can't rely on these institutions to hand down meaning to you. Of course you probably could, but it's just like less and less convincing. I don't know. When I was in FIRST, I kind of bought into the FIRST propaganda, sort of the things you get told about, like recruit kids to your team and like teach them STEM skills.
Sav: And I still very much believe the FIRST propaganda. But I just kind of found it harder and harder to believe the propaganda of a university, for example. So found that well, I have to come up with my own meaning for doing things. And so there was certainly a period in my life where I was kind of like, I guess, lost, and not really knew that well what I wanted to do, what I wanted to work on, coming off of robotics. But, talking to you now, I feel like I, for the most part, know my interests and sort of the stuff that I want to go and chase [crosstalk 02:05:48]. So because of that, I'd say that I'm reasonably confident.
Michael: Sure. I think we have pretty similar kind of ideas of what is useful to focus on now and what would be like significant in the future and will have impact. And so I think maybe like to trace back from to this stack, to pop the stack and go back to why ask this question about robotics is about learning, right? So essentially I was interested in how much you learn from robotics and those competitions, because of course this was like a very optimized way of learning horizontal skills, robotics. And also our management, psychology, meta games, like essentially life. And then the question is like, how do we... And then you found other stuff, right. So other stuff to learn. And I think there's... So, to bring back to just like other parts, there are other aspects to robotics than just science.
Michael: There's also this kind of social connectedness and group and beliefs, and sportsmanship of being the same team and all the things where there's reward. So there's kind of motivation and, as you said, there's like this strict goal of winning and you know how to win and there's like deadlines and stuff. So I guess what I'm interested in is in education. So, my experience is that I've spent three years with the best, the top high schoolers in mathematics. That's where I'm doing math competitions with FIRST and friends or something. And we were told that if we were the best at math, we would go to top engineering schools. And then we would like win at life. That would be like just the end of it. Then you just like add the maximum salary. You have maximum the best jobs. Everyone respects you. So I kind of did this like competition of math for a while. And that motivated me because you have three years or something of rewards of like teachers encouraging you, deadlines. Everyone wants the same thing, where everyone wants to compete to have the best engineering schools.
Michael: And after that, as the same as you, I tried to find external motivation to do other stuff. And so, I'm curious about your ways of kind of giving yourself reward, of pursuing long-term projects when there's less social constraints? During pandemic, you don't have any teachers, any class to attend. Maybe you need to attend some classes, I don't have no idea, but you essentially need to be both the agent walking and the agent that puts the carrot in front of you.
Michael: So yeah, it's mostly like coming back to your posts on the future of learning. Yeah, you made some points about education trying to make everyone fit in the same bed, the Procrustean bed, and then you want education to be more something different that would be like... Just to throw some more at you, that does elaborate personalized, decentralized, self-based, group-based with peer tutoring, peer collaboration to have like some monotorial system where, as I understand it, everyone kind of mentors each other a bit in some kind of same level hierarchy. So same with your robotics team where you're like free leaders instead of having a teacher and top-down architecture. Yeah, so how do we solve education and what would... In five years, if we've solved education, what would it look like independently of AGI and different phases, like just education?
Sav: Oh, man.
Michael: There's a lot of so many threads there.
Sav: Yeah, yeah, yeah. Where to start? Well, first of all, I don't think we're going to solve education in five years.
Michael: Yeah, I don't think so.
Sav: I think on an individual level you can solve education for yourself.
Michael: Definitely. Sure.
Sav: When it comes to reforming these big systems-
Michael: Yeah. Impossible. Yeah, okay.
Sav: ... Maybe something-
Michael: Okay. What-
Sav: ... But for yourself-
Michael: ... Okay, so, what's a good education system for yourself that you will be using in some short-term range, like six months or a year, that could be better or ideal, or that you're moving towards? Or that you want to-
Sav: Well, so I found finding small groups of people who are into the kinds of stuff that you're into-
Michael: ... Twitter.
Sav: ... and like working on Twitter. And working on projects together really helps because it's very hard to motivate yourself, sitting in a room by yourself to do all these things. But if you have a small group of people, then you can create some kind of like social accountability, right?
Michael: How do you find those people? Why do those people will stick together?
Sav: I mean, so far I've just kind of been reaching out to various people who I find interesting [crosstalk 02:12:08] on Twitter specifically.
Michael: Does it work? Do you actually get work and long-term projects for like months with those people you-
Sav: Yeah, yeah, yeah, it works. It actually works. Well, for example, very recently, actually, I got the opportunity to... So I met these three people online who all focused on sort of building new software tools, like building new tools to help people think, to help people learn and to help people be more productive.
Michael: Modern day Steve Jobs.
Sav: Yeah. Or we're like more trying to pick up the torch from where folks like Doug Engelbart, for example, or Alan Kay, or some of these very early personal computer pioneers have left it. And basically saying, well, there's so much creativity back then, like so much potential for where a person in computing could go. And then it kind of like stagnated and became very sort of monotone and almost trying to bring that back.
Michael: I'm not sure if Alan Kay was the same as like the Bell Labs era. Is it-
Sav: Yeah. Yeah, yeah. Same era, same time period. I don't remember exactly what his affiliations were. But yeah, like Doug Engelbart, for example, founded this research lab, I think one of the first research labs to really think concretely about personal computing and he was very focused on augmenting intelligence. Like the lab was called ARC-
Michael: Oh, was it the one where-
Sav: ... Augmentation Research Lab, I think, or something.
Michael: ... Because I remembered something when reading Steve Jobs' biography where he actually goes to another lab to kind of pick up the stuff about desktop interface.
Sav: Yeah, yeah, yeah. Yeah, I think-
Michael: Was that the ARC?
Sav: ... Jobs went to Xerox PARC.
Michael: Oh, yeah, yeah, yeah.
Sav: Xerox PARC took a lot of ideas and I think a lot of employees actually from the original ARC lab-
Michael: Oh, okay.
Sav: ... that was started by Doug Engelbart. So yeah, just basically like a community of people who are very interested in that kind of stuff and want to pick up where these great computer pioneers left off. And so, the four of us were like, "Hey, let's try to meet in person." And-
Michael: In Canada, I guess? Are they Canadians?
Sav: ... One of them was Canadian. Another one is from San Francisco and the other one's from Columbia. The person from Columbia was like, "Hey, you guys should all come down to my home in Columbia." And let's just work out of my home and try to come up with different software ideas. And so I did that. I got the opportunity to go and do that.
Michael: You went to Columbia?
Sav: I went to Columbia in mid-February of this year. I actually got back from Columbia on April 1st of this year, 22 days ago.
Michael: No way. No way-
Sav: And that whole thing-
Michael: ... I don't buy it. I don't buy it. When I last talked to you, were you already in Canada? Yes?
Sav: ... I was already in Canada. Yeah, yeah. Yeah, most of that-
Michael: And you were just like recently in Canada? Okay.
Sav: ... In Columbia. Yeah.
Michael: And did you have parents that are supporting you or did you make money in Columbia or, how-
Sav: I had savings. And so I-
Michael: ... Cool.
Sav: ... was able to support myself on those savings.
Michael: Okay, so when you're saying-
Sav: My parents definitely did not support a spontaneous trip down to Columbia. But my point is, this whole thing happened-
Michael: ... It works.
Sav: ... right?
Michael: I thought it was like a one-month project. It was like a one-year project with random guys on the Internet and you traveled to another continent. Or it was in the same continent, but very far away.
Michael: And the guy you just talked to him Zoom, or something like that?
Sav: We met over Twitter. The guy who organized all this, his name's Azlen Elza. Very like creative, generative guy, I learned a lot from him. But basically, he had this idea of like starting learning groups where we would just like meet up on Zoom, talk about a particular topic. For example, we'd talk about the future of learning. There was another group where they talked about architecture and another one where they talked about learning Chinese. And so through these learning groups, we kind of got to know each other and became friends over these groups. And then after that, we were like, "Hey, let's meet up and work on projects together." [crosstalk 02:17:28]. Exactly, exactly. And it just turned out to be this like really cool opportunity to just kind of sit there and learn from each other and just brainstorm ideas and...
Michael: And build your own [crosstalk 02:17:51] in Columbia.
Sav: And it wouldn't have happened without Twitter. Yeah, yeah, yeah, exactly. Yeah. And I mean, unfortunately it didn't kind of go for as long as I was kind of planning it to go for, But it's still such a cool example of being able to just meet people on the Internet and then get together, actually, in the physical world and then build stuff together, learn together, and bounce ideas off of each other.
Michael: I'm not that surprised that... I guess I'm pretty surprised that you went far away and spend one year there. But I think one-
Sav: Not one year, a month and a half.
Sav: So I went there in mid-February, came back April 1st. Yeah.
Michael: Oh, sorry. I thought it was-
Sav: So, it was kind of short.
Michael: ... last year, February to-
Sav: No, no, no.
Michael: ... two months and a half. Okay. Well, it's still impressive. And, I guess, the other Canadian and the American, they also went there as well, which is even crazier, right? Or, did they stay...
Sav: Yeah, yeah, yeah. Like there's basically three people who came down and one was already there. Yeah.
Michael: Can you repeat the name of the guy, because I've had to google on your photos, but I'm not very good at it. Maybe you can tell me afterwards, so it's not publicly-
Sav: Sure, sure.
Michael: ... Yeah. I think it's better that way.
Sav: Okay, all right. Sure.
Michael: Do whatever you want.
Sav: Well, yeah. Okay. So, the person who was also from Canada, his name's Azlen Elza and he's like... I don't know, I'd recommend everyone go to his Twitter because he's just doing so many cool things.
Michael: Oh, Asla Aslin? No? I'm very bad typing. Just send me in the chat.
Sav: Yeah, yeah, yeah. And yeah, he's friends with this person down in the States named Matthew and they were part of this other kind of, I guess, experiment in learning called School 2.0. And basically, from what I heard of it, it's this big, I guess, meetup of, I don't know, like 30 people or so down in New Mexico. And they like rented out this big Airbnb house in New Mexico, and they just spent several months there just kind of making things.
Michael: I'm googling it and it says, it's from The Thinking Stick, is it the same school? There's a School 2.0 from a website called The Thinking Stick, where I partner with organizations in helping us to understand the changing nature of learning, but working together in long-term, embedded professional development. Maybe not.
Sav: I'm not exactly sure. I don't know that much about them.
Sav: So I don't want to say something that's wrong. But yeah, the point is there are a lot of these kind of groups that are forming now where people with common interests meet online and then they get together-
Michael: In Discord? How do you actually talk to each other? Discord?
Sav: Yeah. Yeah. Yeah, Discord or Zoom.
Michael: And so, you just like have this like weekly meeting where you discuss how to improve education, how to match learning? And you just reached out to this guy on Twitter, because he had learning in his bio or something, or had full-time-
Sav: Yeah, something like that, something like that. Yeah, yeah. The learning groups would basically consist of picking out several papers to go over and read and then have a group discussion about those papers. So example in the future of learning group, the papers would have to do with the ways in which people learn. Or maybe it's like some excerpt out of a book or something, and then we would all discuss it.
Michael: And is it part of... Because I've listened to, I guess, one of your podcasts on Spotify, there was one on perception and I guess you were discussing one paper with someone, was it part of this thing or something-
Sav: No, that was separate. That was just out of interest.
Michael: ... Not like connections? Interesting. And yeah, okay, so what's your relationship with Twitter? How long have you been there and how do you meet people there?
Sav: [inaudible 02:22:54].
Michael: Oh, okay, okay. Should we stop? So-
Sav: I think I can hear you now. Yeah. Yeah, okay, good.
Michael: Yeah, you can still talk and hear me?
Sav: Yup. Yeah, yeah. Yup.
Michael: Cool. Yeah, I think the sound quality is fine.
Sav: Okay. Yeah. Sorry, can you repeat what you were saying?
Michael: I was essentially saying that yeah, how do you approach Twitter? When did you start and when did you start reaching out to people?
Sav: Well, I think I started when most people started. At the start of the pandemic when, I don't know, I was just sitting at home and I was like, "Well, I want to still meet people and socialize with people. And so, hey, I have this Twitter account that I never use. Maybe I should start using it." So, yeah, and through that, I just kind of found a lot of people with similar interests. For a while, I would just follow people and not really interact with them all that much. And then decided to just message some people out of the blue and be like, "Hey, let's meet up." I guess, kind of like how we got to meet each other.
Michael: Because I had a calendar link on my Twitter bio. But if I didn't have, will you have messaged me and sent that DM? Maybe like more high effort, right?
Sav: Yeah, yeah. Well, I don't think I actually noticed the calendar link in your bio. I just kind of messaged you.
Michael: Oh, really?
Sav: And that's kind of how I do it with-
Michael: Oh, we should put in more evidence. There is a Calendly.
Sav: Big, bold font. Put a bunch of clock emojis around it and it will help.
Michael: A bunch of what?
Sav: A bunch of clock emojis.
Michael: Oh, yeah, clock emojis. Yeah, yeah, exactly.
Sav: Yes, yes, yes. Or I've noticed people hosting events, organizing events, like the whole story about how I got in touch with Azlan was just he started this learning group and it seemed like an interesting thing. So I signed up for it and that's how we got around to talking. And so, yeah, it's just, I don't know, I guess there's a number of ways to get involved with people.
Michael: Okay. And now I want to know... Maybe it's too personal question, but I've kind of reached out to some people on Twitter. I met some really cool people and very smart and some ways I feel that one of the smartest people I've talked to weren't people I reached out to, but people who reached out to me, because I was being open about talking to people and having this kind of link. And so, yeah, maybe, yeah, you can share your experience with how much time you spend on Twitter or... Personally, when I've had to call or email people I don't really know, or haven't interact with on Twitter before, it's pretty difficult to actually have a call because you need to find a time, you need to find a reason, you need to negotiate something. So, yeah, how has it been for you? Was it mostly like if you're similar bios, then it's easy to meet on that call? Or was it like a half ratio or one in 10 or something, I guess, what...
Sav: I don't know, it varies. I think being familiar with the person beforehand helps a lot. If I just naturally read somebody's blog or have seen things that they posted or know a bit about their interests, then it's just easier for that conversation to happen. Or if there's a specific, I don't know, reason to have the conversation, then that also helps. For example, with Azlen - I told him this - because I'd been following him for a while, but I couldn't really think of a reason to actually reach out to him. And then he was organizing these learning groups and so I thought, wow, like this would be a great opportunity to just learn together and get to know each other. And so, the context for the conversation kind of naturally arises out of like whatever the activity is. Yeah, so I don't know, part of the reason that I've been kind of getting into this whole podcast game is that it's really just an excuse to talk to interesting people.
Michael: [crosstalk 02:28:13]. It's to save the world, please, save the world.
Sav: Of course. Yeah, yeah. And so, I don't know, maybe saying, "Hey, do you want to be on my podcast," is a better excuse for conversation than like, "Hey-"
Michael: Do you want to be on my podcast, and I'm recording this.
Sav: ... Well, yeah. So, I don't know. I don't know. I'm still very new to this whole thing, but at least it kind of like creates a context for conversation, right?
Michael: Mm-hmm (affirmative).
Sav: Because we're having a conversation... By the way, this is very meta. But we're having a sort of a conversation. To the extent this conversation is interesting or insightful, we're actually doing something useful or potentially interesting for others too. And maybe that's one way to facilitate a conversation.
Michael: Yeah. Actually I think it's worse... So I think a weird way to look at it is what would you advise your younger self to do? I think you've quoted Alexey Guzey in one of your essays and I talked to him a few years ago and-
Sav: He's great.
Michael: ... Yeah. He read those posts about how to reach out to people, when do you follow back? When do you either follow back or when you send one more message? Is it follow back? Anyway-
Sav: Yeah. Like follow up, I think.
Michael: ... Oh, follow up, follow up, right. Follow up. How many times, you could follow up four times and if the person hasn't answered you after four times then maybe you should stop sending the messages. So yeah, I think he was a great inspiration for me, to actually reach out to people and be open and have a... I had a blog before that, but I think just sharing our experiences online can make people more confident about reaching out to people. So yeah, for me, I think something that wasn't in Guzey's post, but I think he highlighted those, but was something like most people will not want to talk to you if it's not something specific. So if you're just cold emailing someone... I think Twitter is different because Twitter you have mutuals, you have trust. People if they have their games open, then mostly they want to interact with other people.
Michael: But if you want to email for jobs or something, then it's pretty hard. If you want to email researchers about the papers it's pretty easy. If you want to DM people on Twitter, it's pretty easy. If you want to cold-email a girl on Instagram with like a bazillion followers and is super cute, it's pretty hard. So I think those things are rules that you learn somehow doing it. Yeah, do you have any rules like that, that you learn or... Or maybe like hard lessons?
Sav: Yeah, probably. Although I'd say that they're not as well-formed or concrete enough for me to write a blog post about it or... I haven't really... I don't know-
Michael: Or there's some social negative points in saying them out loud.
Sav: ... No, no, not, it's not even that. I think, for me, it just kind of all exists in this intuitive level of what feels right and what doesn't feel right. And I haven't spent too much time yet trying to actually formalize that to some kind of like-
Michael: So, what feels right?
Sav: ... Well, reaching out to people on Twitter with similar interest fields, right.
Michael: Yeah, I agree.
Sav: I mean, to your point about what Alexey wrote, like probably stopping messaging someone after he doesn't reply to you four times, feels right. But like, I kind of just approach it as making friends. I try not to over-complicate it and I just look at it as like, oh, well, is this the person who I can potentially vibe with and learn things from? And if the answer is yes, then probably it goes both ways. And so, I kind of use Twitter as a platform for just making friends and maybe there'll be opportunities to work on cool projects with these friends. That's really as far, as deep as it goes for me.