March 30, 2023
What would happen if you chose to let ChatGPT control your life for a year? If products like ChatGPT help us (e.g.) to write something important, then should we give it credit as a co-author or merely act as though it's a high-powered Grammarly? How would you feel if you received a hand-written card from a romantic partner but later found out that everything they'd written had been authored by ChatGPT? How can we learn to get along with — and perhaps even form friendships with — people with whom we strongly disagree?
Dax Flame was one of the first YouTube stars. He acted in the movies 21 Jump Street and Project X, and has written three memoirs. After running out of money, he spent a few years working minimum wage jobs, but now at 31 he is attempting to become a full-time YouTuber again. Check out his YouTube channel, @Daxflame.
JOSH: Hello, and welcome to Clearer Thinking with Spencer Greenberg, the podcast about ideas that matter. I'm Josh Castle, the producer of the podcast, and I'm so glad you've joined us today. In this episode, Spencer speaks with Dax Flame about using ChatGPT as a life coach, dealing with early YouTube fame, and authenticity in the era of AI.
SPENCER: Dax, welcome!
DAX: Hello, Spencer!
SPENCER: It's really interesting having you on the show because I just heard about you recently, and I was really intrigued about what you're doing. But as I started investigating you more and finding out more about you, I discovered that you're kind of an enigma. A lot of people are really interested in your life and what you've done, and people try to make sense of it and seem to struggle with that. But I want to start with what really got me onto your work, which is your current experiment with AI. So do you want to tell us about how you prepared for this conversation and the general experiment you're doing?
DAX: Yeah. So basically, this year, I am just consulting with AI and letting it kind of control my life and my YouTube channel because that was just an idea I had. And I didn't really have any other ideas of which direction to take my life because I was very broke as the new year approached. And my luck has kind of shifted for now, so I'm really happy because my YouTube channel is starting to take off. So I'm just kind of consulting with AI all along the way. How that ties into today is, whenever I got an email or a Twitter message from you, I was very excited because I looked you up. And I was like, “Wow he’s on TED Talks.” And earlier this week, I was asking ChatGPT about coming on a podcast. I asked, “What's the ideal podcast guest?” And just a moment ago, I guess, I typed into ChatGPT, “I'm scheduled to do a podcast interview in six minutes. I'm very excited to talk to the host. He has done TED Talks and talked to very respected people, and I have been excited for this all week. But I'm also nervous, because I was hoping I would have more time to prepare what I was going to talk about, but I have been too busy this past week and just finished a video and I just released it. What should I do?” So then ChatGPT gave me a bunch of suggestions: review topics, focus on your strengths, be yourself, prepare some questions, take deep breaths. Then I said, “I don't have time to prepare anything, can you share some good conversation topics, etc.?” So it has some of those if we need them, but I'll just let you lead the way. But I have some of that. And that gave me some assurance. So maybe that's an example of how it works into my life at times. But mostly, I've been using it for YouTube at the moment.
SPENCER: Right. So why don't you tell us the story of how you started using ChatGPT to sort of control your life and make your decisions for you?
DAX: Yeah, for sure. So basically, it was a few days before New Year's Eve and I was just very broke. So I was just looking up jobs on Craigslist. And then I had an idea like, “What if this year, I just let ChatGPT control my life?” I just had that idea. And I was like, “Well, there's nothing to lose. So let's just try it.” So then, on New Year's Eve, I put out a video just about my life situation in general. And then on New Year's Day, I put out a video saying, “This year, ChatGPT is controlling my life and let's see where it goes.” And now, I've just been continuing to do that. And it's just been super helpful, because it's just this thing — it's AI — and it helps you. So I've been really thankful for it and enjoying it a lot.
SPENCER: Now I want to give a little background on ChatGPT, just for the listeners that may not have heard of it yet. So ChatGPT is made by OpenAI and it's based on the same technology they used in GPT-3. I've done some prior podcast episodes about GPT-3. The main difference with ChatGPT is they did extra training, to train it to essentially have an interface where it answers questions or it responds to whatever question you ask it. And it tries to respond in a way that follows a set of rules. They never want to say anything offensive, they want to avoid saying controversial things or speculating too much, they want to say accurate things, and so on. So whereas GPT-3 really is technology that says, “Given some text, predict what text would come next.” ChatGPT sort of adds this more user-friendly interface where you kind of talk to it like it's a person. And then it really tries to avoid these [quote] “bad behaviors” like speculating, saying untrue things, or saying prejudiced things, etc. And so, there's that extra fine-tuning on top. That's the layer that Dax has been interacting with. Now Dax, I'm wondering how seriously you take what it says. Do you have a rule where you just have to do what it says? Are you allowed to override it? Can you regenerate its response and so on?
DAX: Good question. I just kind of try to do what it says, but not if — I've had a couple instances — I will consult with it, and you can tell it's misunderstood what I'm trying to communicate and it just says something back that doesn't make sense or doesn't apply. So at times, it's really interesting because you can just say something else to it and then it starts to grasp it more. If it said something like, “You have to get a tattoo today.” I wouldn't do that. I don’t know where such a situation would come up. It's only given good advice so far.
SPENCER: Okay, so basically, as long as it says reasonable things, it sounds like you'll do them.
SPENCER: But if it says something completely crazy, like it tells you to commit murder [laughs].
DAX: No, I would never do that. And that's why I actually have this interesting thing called ‘code red.’ In that very first video I made this year, I kind of said, “Because I think this is something that you will be very aware of, that if I wanted to opt out of this experiment then I would say the words ‘code red’ and make a video called ‘code red.’” That way my fans and viewers just know that I'm opting out of it. Because I think that if you let AI control your life, and as AI gets more advanced — I don't think it's there — but I've just heard about different routes that things can go. It's just been super good and helpful. But some people have fears that it can go down a negative route as it gets more advanced, so I didn't want people to ever question if I was being controlled by AI to the core. And so, that's why I said if I make a video called ‘code red’ — for example, if I noticed AI said, “Go commit murder,” or something like that — then I'd be like, “Oh, ‘code red’, immediately stop this.” And I'd make a video called that.
SPENCER: Right. But as long as you're not doing ‘code red’ people could just assume that all your videos are something that AI told you to do, essentially, right?
DAX: Totally. And actually — I think it was the third or fourth video I made this year, it was something called “Letting AI Fix My Finances and Lovelife,” something like that — the AI had come up with the script for that video, the title of that video, and the thumbnail suggestion. And so, all of that was from the AI and I posted it without thinking twice. I was like, “I told everyone that this would be AI controlling my life this year, etc.” But then, after I posted it, I started to get this weird feeling. A lot of people were complimenting the video and saying, “Wow, you seem so eloquent in this video. This is an awesome video. I'm so excited. Congratulations.” And I was very excited to read that. But then I got this weird feeling that like, “Well, are they really complimenting me? Or am I being deceptive?” So I came up with a system like a week or two later — I apologized for not coming up with a system earlier — but basically the system is that: in my videos, I put at the end either a blue dot or a white dot. And what this signifies is if it's a white dot, then that means that the AI completely wrote the video and every aspect of it. And if it's a blue dot (which is the majority of videos, all other videos, and probably 90% of the videos I would make this year, I assume), that means that I've just consulted with AI about the video. For example, within the video, you can see me talking to AI, or I just said off-camera to AI, “What’s some good videos I can make today?” And it's like, “Do XYZ.” And then I go and do that. But it didn't write the script for me, so I have that. And I also have a comment system that's very similar. If the people watching the video want to leave a comment that's written by AI, I just ask them to put a white dot at the end of the comment. That way, everything is clear. And I'm curious if you think that in the future, people will need more systems like this.
SPENCER: It's a really great point. So I did this experiment for a while right after GPT-3 came out, where I told my Twitter audience that sometimes my tweets will be written by GPT-3. And I turned it into a game. Basically I said, “If you can correctly guess which ones were written by GPT-3, then I will give money to the charity of your choice. So I have some amount of money I would give away. But you only get one guess.” So in other words, if someone guesses it was written by GPT-3 and it wasn't, then they're out of the game, they can never guess again (as one of my Twitter followers). But if they get it, then they'll win a prize for a charity of their choice. And so I did that for a while and it was pretty interesting. A lot of them were caught by my Twitter followers. But occasionally a couple slipped or nobody caught them. And in addition to that, there were a bunch where they thought it was run by GPT-3, but I actually wrote it. [laughs]
DAX: That's cool.
SPENCER: But I think you're hitting on something really interesting here, which is that as these AI systems get better and better and better — and they're already getting so impressive — what do we do when AI has done a significant portion of our work? Is that a certain type of plagiarism? Is that something we should really call out, or can we just think of it as a tool? If someone uses a grammar checker like Grammarly, they're clearly not going to tell people because it's sort of like, “Well, yeah, of course you're using a tool to help you.” That's not a big deal, right? But if AI is actually writing 90% of what you're saying, then maybe you do need to credit it to the AI.
DAX: Yeah, for me, the best solution is just the blue dot and the white dot. But I wonder in the future, maybe people will not even think twice about any of it. I wonder.
SPENCER: Yeah, you can imagine as it gets more and more adopted, eventually, people just assume that a lot of stuff is written by AI.
DAX: Also, as AI gets more advanced, there's the potential that it could learn your passwords, replicate what you look like, and make videos as you. The reason I put a dot in the video — as opposed to just solely in the description box saying, “This is written by AI.” — is because in the video on YouTube, you can't alter something that's already been uploaded. So it's just really safeguarding against that, if there's a future in which people get scared of AI or something.
SPENCER: It's interesting that you're thinking about that already.
DAX: Yeah, just because I don't want anyone to ever feel like, “Oh, is this coming from his heart or not?” You know what I mean?
SPENCER: Right. And so, it raises interesting questions around why we care if something was written by AI or written by the person, if the person did end up approving it. So, if at the end of the day you're checking off on it, do we care? And I think there's at least a couple different issues here: One is, if the AI writes it, maybe it feels less authentic. Maybe it was close enough to what you feel that you were willing to approve it. But that doesn't necessarily mean it's what you would have written. And so, maybe it's sort of drifting from your authentic point of view, to some extent. What do you think about that?
DAX: I don't know, that's interesting. I think that maybe it can be a more eloquent writer than me. But maybe as long as it's like the message that I want people to know, or communicate, then I feel good about it. And then I'll just keep checking in with viewers along the way, if they feel good about things, and I can always change the system from the blue-dot-white-dot aspect. If people are like, “Well, we just disapprove of any white dot video.” But so far, people have just been really supportive, so I've just been super excited about that. But I guess like, if it's more authentic or something? Yeah, I don't really know. Do you have an opinion on it?
SPENCER: Yeah, I think it does lose authenticity. Imagine that you have a romantic partner, and they write you a really sweet and thoughtful message. And then you discover that the whole thing was written by AI. Doesn't that sort of undercut, like, “Do they really feel that way?” Yes, they signed off on it, but like…
DAX: Well a lot of people buy their partner a card from the store that's written by someone else.
SPENCER: That strikes me as just incredibly inauthentic [laughs] and maybe people are okay with that.
DAX: Some people really love it. Maybe your opinion is more the majority of what people would think, though.
SPENCER: It's always baffled me. I get the funny cards, because that's just kind of cute, but the ones that have a heartfelt message always struck me as very strange (to write a heartfelt message that’s actually written by someone else). But I guess at least they know it's written by someone else because it’s on the Hallmark card.
DAX: Oh, ‘cause you're saying if someone tried to take credit for it as well.
SPENCER: Exactly. Let's say, you got a message, and only later you found out that an AI actually wrote it and they just signed off on it. That feels like it would undercut the meaning and authenticity of that sweet message.
DAX: That's how I feel. That's why I feel like you have to be clear about where it's coming from.
SPENCER: Now, I think your experiment raises another interesting question around: At what point is it no longer within the realm of what you would do? Like you said, it's giving you good advice. But that doesn't mean that you would have done these things without it. So in some way, it's sort of changing your behavior. And I'm wondering how that felt for you. I imagine it must have pushed you to do things that you just wouldn't naturally have done or wouldn't normally be in your behavioral patterns.
DAX: So I'll share a little backstory. Basically, I got really famous when I was 15 years old for doing YouTube videos. I was just uploading vlogs talking about life. And then they just randomly got popular on the internet in 2007, which made me one of the first YouTube stars (I was like 16th most subscribed of all time at the time). I kept doing that for a couple years. And then eventually, I just stopped because — I don't really know why I stopped — but maybe I was burnt out. Or maybe it was something else. I'm not really sure. And then I got to act. So I was just kind of done for a bit. Then I graduated high school and immediately got so lucky, I got to be in a big Hollywood movie for Warner Bros. called Project X. And a year after that, I got to act in another movie called 21 Jump Street. So my career was taking off right after high school, then I struggled to find any more Hollywood roles after that, but those were very big movies at the time. They got hundreds of millions of dollars (the movies did). And I struggled to find more work slowly, just kind of lost savings. And then eventually, I just started to work at an ice cream shop and I did that for a little while. And then I got really lucky and had a YouTube comeback because this popular YouTuber named iDubbbz made a documentary about me, called “Ice Cream Man” and that rekindled people's interest in me online. And so then I had some momentum for a little while, but slowly over the year or two since then that has dwindled away. And then coming into this new year, I was just out of options again, out of money again. And then, as I mentioned, I was just like, “Okay, I'm just going to find another job.” So basically, the reason I say that is, there have been, I guess, three or four times in my life where I have found success and not been able to keep it. And I think that those experiences made me think that I have a lot of potential, but that I don't really trust myself so much with how to capitalize on my potential or how to keep it going, if I ever do find success. So now I'm 31 years old, and I was just like, “Okay, I'm broke again. So why not just try something new?” And that's where the Albert Einstein quote is — which I think maybe your audience will like quotes from him, just being a genius kind of guy — Einstein said the quote about “Insanity is doing the same thing over and expecting different results.” And so I thought, “Well, this is something I haven't tried — letting AI control my life — so I'll try it out because there's nothing to lose. I'm already just broke. So let's try it out and see what happens.” And then almost immediately, my YouTube channel started to pick up and take off. People were very interested to see what would happen, so I just felt super lucky. And I was like, “Okay, well, in previous times, when I've had moments of success like this, and moments of potential, I have squandered it — not in any crazy way or anything — it’s just that I don't always know the perfect move to make and that kind of thing.” So instead of just doing what my instinct would naturally tell me to do in these situations, I just consulted with ChatGPT. And so far, it's just led me to really good places. My momentum has been maintained since the beginning of the new year, which is awesome. That means six weeks of doing well on YouTube, which is crazy. And I think that basically, the first thing I asked ChatGPT (and the first video of the year) was like, “I just have all these aspects of life I want to work on such as dating, being a more put-together person with a better routine, or some kind of habits and stuff. Or if I just want to be someone who is financially stable, or just have a Hollywood career kind of going.” And ChatGPT was like, “Maybe your priority should be your finances.” And so, where do finances come from? Basically, I've been presented this opportunity to have finances come from YouTube. So then I said, “Okay, if a week or so later, how can I work on YouTube?” And ChatGPT is like, “Collaborations and staying consistent and doing this and that.” And previously, I've really struggled to stay consistent. And I'm not always the best at networking or collaborations. But I've been making myself do these things because ChatGPT has told me to, and it's working so far and I don't know if it will work forever. But previously, the other things I've tried didn't work, so I'm glad to just be trying this. And I'm just excited that it's a thing that exists, because it's been really helpful and I'm really grateful. And I'm just super feeling glad about that. And whatever happens next, we'll see. I don't know. But yeah, we'll see.
SPENCER: So has it pushed your comfort zone in a big way? I'm curious to hear examples where it got you to do things that you're pretty sure you wouldn't have done without it.
DAX: Yeah, so basically, staying consistent on YouTube is something I typically don't know how to do very well. But I've just been asking for ideas when I don't have ideas, so it's helped with that. Then trying to network more, I usually don't network just because I don't know how to do it that effectively. But it's recommended to reach out to successful people, so I've texted a famous person in my contacts and I've messaged a popular influencer who's messaged me in the past. And just today (the thing that I've been so busy on), I did a collaboration with a friend of mine who's very popular on YouTube. So I feel like now I've checked that box of doing a good collaboration. And I also went to the park and I just talked to strangers to see if I could get them to be in a video. And kind of like I was like, “Okay, well, people like to see me make videos with other people. Who can I make a video with today?” I drove over to the park and started to ask strangers, “Do you want to be in a video?” I did that for like 45 minutes and I finally found someone who said they were down. And once he said he was down, I said to ChatGPT, “Any video ideas for us?” We quickly filmed a challenge video as recommended. It recommended us to do a frisbee-throwing challenge video, just like a fun quick thing, because I had a frisbee with me so I figured let's do a frisbee challenge video. It wrote the whole script and everything. And then afterwards, the guy just had more free time and ChatGPT also recommended doing a documentary profile. So I started to do a documentary profile on him that was suggested by ChatGPT with a lot of questions suggested, and then I just veered off from what was written. So the frisbee video was all written by ChatGPT, whereas the documentary was suggested and prompted by ChatGPT with some questions written. But then it just veered off into being a documentary that was a combination of recommendations and just what I thought was interesting. So normally, in these situations, I might be like, “Oh, I don't know what to make today. And I don't know what to do.” But instead, I drove to the park, and just talked to the stranger and made a couple of videos all in one day. And that day, I tried doing my daily routine for the first time (which had been suggested by ChatGPT), and it just happened to be that the day I did it perfectly, was one of the most productive days ever because I made two videos and tried the daily routine. The daily routine is something that I'm not doing perfectly at the moment. But that's okay, because ChatGPT has recommended that I focus solely or focus primarily on career stuff first, such as being a full-time YouTuber. And once I really get that to a stable place, maybe I can put a little more attention on things such as daily routine, dating, etc.
SPENCER: It's clear that ChatGPT is generating ideas for you, and that's helpful. But is there something else that’s almost like playing the role of a life coach, where, if you meet with the life coach weekly, it kind of creates this pressure to act on the advice of the life coach. You're like, “Oh, I want to make sure I've done the thing by the time I meet with the coach next.” And I wonder if that's sort of happening for you too, where, because you should have committed to doing what it says, if it says to do a new morning routine that's healthy, you're more likely to do it than if you just came up with the idea yourself. What do you think about that?
DAX: Yeah, that sounds like it makes sense.
SPENCER: I wonder if you're pioneering something that's actually going to be commonplace in the next, let's say, 10 years, where we have decision-making tools like advanced versions of ChatGPT in our pocket that are helping us make our life decisions or thinking about what we want to do in life. You might just be the pioneer of this. You’d be on the bleeding edge of this.
DAX: I have no clue what will happen. It's a really big relief to not have to make all my life decisions right now, so that's kind of nice.
SPENCER: Presumably, at some point, it's gonna give you advice that you think is bad, but not disastrous. I'm not talking about ChatGPT telling you to murder someone, but just advice where you’re like, “I don't think that seems like really a good idea.” Are you just going to do it in that case? You just push through anyway?
DAX: Probably, because previously, my ideas — not to sound defeatist — haven't brought me to successful places. So for now, I will try to listen, but maybe it's helpful to hear an example of what you're talking about because I don't want to say that without having an example.
SPENCER: Sure. For example, it might recommend that you make a video that's really off-brand for you. It's really not the sort of thing that you would normally do in your channel, something that you don't think you're good at, something that your audience won’t appreciate, or something like that.
DAX: Yeah, I would do that. As long as it's not an offensive video. If it's saying, “This is what will help you with your goal of being a full-time YouTuber and your fans will like it,” then I would do it, just as long as it's not a bad subject matter. Because I said that typically I've not been able to lead myself to sustaining success. I don't want it to sound like I don't believe in myself. I know that I'm really good at certain things. I just need help with some other things, like some other aspects of success and maintaining it. So I don't want to sound sad about anything because I feel super excited.
SPENCER: So as you mentioned before, you've had these sort of breakout successes a few times in your life, right? Like at age 15, you had amazing success, you became super popular, then you got into these huge movies, then you had another surge in popularity recently. Why do you think you haven't been able to maintain it historically?
DAX: Good question. Um… I don't know. I think because I don't always know the exact right next move. And sometimes I overthink things. So it's really helpful to have something telling me to just keep being consistent, collaborating, etc. And typically, I just don't know how to do that. Or I guess, I don't know. Maybe the answer is that if I knew, then I would just do those things.
SPENCER: Right, well, sometimes we don't do a thing because we don't know what to do. But sometimes we don't do a thing because we find it difficult to do. And that's so different. But it sounds like you aren't sure. And you're hoping to kind of break out of that pattern. Is that right?
DAX: That's right.
SPENCER: I have a theory about you, I'm curious if it resonates. I had this idea of “singular people.” Imagine that you took millions of people and measured all the different traits that they have. And you plotted them as points in space. Most people, you'd find that there are other people nearby. If you plotted them as points in this high-dimensional space, you'd be like, “Oh, yeah, they sort of have other people close to them.” But there are these people I call “singular people” where there's nobody in their sort of vicinity. And in their personality space, they're just sort of an unusual outlier. And so, because they’re such an unusual outlier, it's hard to generalize about them or predict what they're going to do. Because they're genuinely not like anyone you know, and I suspect that of you. I mean, I've watched a bunch of your videos now. And I suspect that you are just a singular person. You're an outlier in this personality space, there aren’t other Daxes around. I'm curious what you think about that?
DAX: That sounds interesting. I think that if that's what you think, then that makes sense. Or that feels accurate.
SPENCER: Does that resonate with you though? Do you feel like you're quite different from other people? Or do you actually feel quite similar to others?
DAX: Yeah, I don't know. Maybe? Both?
SPENCER: Yeah. Could you elaborate on that, when you say both? What are some ways that you feel like you might be different from others?
DAX: I feel like I relate to a lot of people in a lot of different ways. And then other times, I feel like maybe this is a little more unique to my experience. And in other ways, this feels like exactly like anyone else.
SPENCER: One thing that kind of comes up throughout your career is that so many people on the internet are always speculating about whether you're playing a character. But to me, at least, it seems like you're not playing a character, you're being yourself. Maybe you're exaggerating certain things for comedic effect or for your videos just to make them more interesting, but to me, it just seems like you are yourself a unique person, and you're just being yourself. And people are just so unused to someone like you that they just sort of see it as a character. How do you react to that?
DAX: I think that feels accurate. I don't really exaggerate anything nowadays. I think when I was a teenager on YouTube, I would definitely just try to hype things up and really lean into my emotions more, because I just realized how much attention that would get me.
SPENCER: Another thing I wonder about: You were a kid, you had this huge success, you had all these people talking about you, but I imagined that that attention wasn’t always positive. Some people loved your videos and thought they're hilarious, but other people maybe thought you were strange. So I'm wondering how that affected you (getting so much attention at a young age and presumably getting a bunch of negative attention as well)?
DAX: That was interesting. I guess that at the time, I thought, “Wow, I'm extremely famous. And that's fine. And whatever.” At the time, even though I was extremely famous on YouTube, that didn't mean what it means nowadays. People didn't really care that much. If someone was famous on YouTube, no one was getting invited onto the Ellen DeGeneres show yet or anything like that.
SPENCER: Right, because it wasn't really a thing in people's minds. Like, “Oh, a YouTube star.” Now everyone knows what a YouTube star is, right?
DAX: Yes. So it was interesting. And then it was very helpful when I was in high school because it felt like I had a lot of support from the people watching my videos. And then it got a little confusing as I stopped wanting to make videos as much. And then I, just slowly over time, figured life out more as I became an adult. And just definitely tried to get a better grasp of things and that kind of thing.
SPENCER: What's your relationship to fame? Do you view fame as desirable? Do you view it as sort of a necessary evil in order to make ends meet doing what you love? What do you think about that?
DAX: I think that it's definitely attached to good things, such as, whenever my videos are being viewed by more people, that's really helpful because it means that I'm potentially making a bigger positive impact if my videos are positively impactful (which is what my goal is with the majority of what I make, and not every single thing is specifically like that, but there's that aspect). It gives you more money, which is helpful because then you don't have to work which means you can make more stuff. I guess that's all. Yeah. So I like it for those reasons.
SPENCER: What about the attention aspect?
DAX: Oh, that's nice for sure. People are always super positive and, definitely, I get a lot more messages whenever my videos are doing more, doing better. But that's also something that was more a part of my identity when I was younger. Whereas now, I like that, but I don't feel like I have to have that to be happy. But it is attached to what makes me happiest, such as making videos.
SPENCER: And what about negative comments? YouTube comment sections are notorious for being horrible and people saying really nasty things, right? How does it affect you when people say nasty things to you?
DAX: It doesn't affect me too much anymore. Sometimes I read something and it makes me feel bad for a little while. But after 16 years of being on YouTube, I'm pretty immune to a negative comment.
SPENCER: Another thing I wanted to ask you about: You have this new direction, potentially, in some of your YouTube videos about uniting America. Do you want to tell us about that? What's your approach there?
DAX: Oh yeah, I made this one video. It was originally going to be a series but I only ever made the first pilot episode, which was a video called ‘Uniting America.’ And basically, the premise of the video is that even though people can be very divided in the world — especially on social media — if you actually meet people in real life, often, they're much more kind than the internet makes it seem. So I wanted to make a video demonstrating this. And so in order to do that, I went undercover posing as someone on each side of a debate. So a hot button topic at the time was college loan forgiveness. So I went to a college (UCLA) and I interacted with people who didn't know they were being filmed. I told them after I finished talking to them, but they didn't know they're being filmed because the cameraman was far away. And I went undercover as a person who wanted to forgive college debt, and a person who wanted to maintain it. And I started conversations with strangers who had opposing views. And then the goal was to talk to them and establish our opposing views, and then demonstrate that we could still get along. So for example, by the end of the conversation, I had asked people if they wanted to maybe toss a frisbee around. And then there we were, with opposing views, and we were still getting along, and just tossing a frisbee around and having a fun time. Just demonstrating that, on Twitter, you think everyone hates each other. But in reality, it's not always like that. And when I was listening to an episode of your podcast the other day, I heard you talking about how you do some kind of experiment or test similar to that, is that right?
SPENCER: So I throw social experiments. And so the idea is that every experiment event, we have a role, where it's never allowed to be the same format again, so each one is different. I haven't done one like you're describing where we go up to people in the street. It's more like we have them at someone's home. But then what happens at the event is a kind of social experiment. And so, it could be playing with different ideas, like playing with the idea of social status, or playing with the idea of trying to outmaneuver someone, or playing with the idea of game theory. So that kind of different stuff.
DAX: Wow, is that for a science experiment kind of thing?
SPENCER: Well, there are multiple motivations I have. One is to give people experiential learning. So I used to do a lot of events where there'd be a lecture, but I felt like it wasn't changing people's lives. I felt like that's kind of interesting, but nobody's going to really be different because of it. So I started thinking, “Well, maybe a way to actually have a bigger difference in people's lives is to give them an experience where they're learning through doing something.” And so I started moving in that direction. The second thing is what fascinates me most in the world is how humans work. And so these are experiments, not in the sense that we're collecting data and trying to test a hypothesis, but more in the sense that we're putting people in an unusual situation and seeing what happens. And I do think I learned some things about human psychology from doing this.
DAX: So did you feel like you changed some people's lives, hopefully, like the participants?
SPENCER: Yeah, some funny effects have occurred. I mean, obviously, you don't always know if you change people's lives. One example, we did an event on disagreements where basically we match people who strongly disagree on a topic for one-on-one conversations, and we put them through a structured disagreement format. So the way it worked is that first, Person A would explain to Person B their view on the topic until Person B could restate Person A's view back to their satisfaction. And then once Person B was able to explain Person A's view back to Person A’s satisfaction, then they would switch and then Person B would explain their view to Person A until Person A could state it back to their satisfaction, so that we made sure that they truly understood each other's perspectives first before they were allowed to go into the disagreement. And then the last piece was that they would work together as a team to try to figure out the root of their disagreement, once they understood each other's disagreement. And so, it went really well. And I feel like people dug into some really controversial topics that they normally wouldn't talk about. But a funny thing happened as a result. A baby ended up not being circumcised because of one of those disagreements. One of the people convinced the other that they shouldn’t circumcise their babies. So that was just a funny effect.
DAX: Can you tell me about another one of those experiments you've done? I'm just curious to hear another one as well.
SPENCER: Yeah, absolutely. So we did another one on social status. When people arrived, they were assigned a social status as a number to wear on their lapel from one to 10. For example, if you were assigned to 10, and I was assigned to one, I would have to act during the night as though you're the most important person on earth, and you'd have to act like I'm the least important person on earth. And then we got to rotate throughout the night. So you get to try on different social statuses. And then, about halfway through the night, I gave a presentation on ways to signal higher or lower social status. So for the first half of the night, you had to just sort of try to do it on your own. The second half of the night, you had explicit strategies, you got to experiment with the top people [inaudible]. So the idea is to try to give people this experience of what it really means to have a different social status, and what it feels like to project different social statuses and to have other people project them. And it had some really interesting effects on people. One woman who tends to project low social status came up to me after and told me that it was just mind-blowing for her because she's never experienced what it's like to have people treat you like you have high social status. And that was really interesting.
DAX: That's so cool. Do you publish those?
SPENCER: Occasionally, I'll write them up on my blog. But honestly, I only write probably five or 10% because I just don't have time.
DAX: So you just do it, mostly just to change those people's lives and to learn yourself.
SPENCER: Yeah, I also find it fun. For me, it's one of my creative outlets. I view it as like, instead of painting or sculpture, I create social experiments.
DAX: Wow, cool. That's really interesting. Do you have any of those where you've combined multiple people in one room? Or is it always one person talking to one person?
SPENCER: Usually, it's many people together in a room, but sometimes it's split into pairs. Sometimes they're in small groups, sometimes they all..
DAX: But it can be a big group, too?
SPENCER: Yeah, generally, I target about 25 people; it's usually a good size. After 25, It's kind of unwieldy.
DAX: I used to make a game show called ‘Smoothie Madness’, where it's really fun to just bring people together and give them a positive experience. And to create fun games and situations for them to participate in. And so it feels very inspiring. I think that ‘Smoothie Madness' can be quite expensive. It was a game show I would make and I would typically lose money when I would make it, but I thought it was worth it to do that. Just occasionally, I feel like what you're talking about could be something that I would like to learn more about, and eventually see if I ever can do something like that someday. It's not something I would do immediately. But, I just would like to hear how you do that more. And if you ask people to come, or if they're all your friends, how much does it cost you?
SPENCER: Basically, I have a mailing list of people who are interested in social experiments and just send it out, generally. Because I limit it to 25, I basically tell people, “You have to RSVP right away if you want to come because the slots will fill up.” So what's good about that, is that it generally creates an event with the people who are most excited to be there. So people are like, “Yes, I'm coming.” It's very much an opt-in. I like to give people (when I send out the email about it) a little taste of what's going to happen because I want them to go in knowingly, since some of the events are pretty weird and I don't want people to be weirded out when they get there. I don't want to tell them exact details, because that would give away a surprise. But I give them an idea of what's going to happen so they can opt in.
DAX: Could you ever make a live theater kind of thing where there's some actors that are telling a story or a haunted house kind of thing? Would there ever be a way to combine it with that?
SPENCER: Possibly. I've never had actors in them. I've thought about strange ones involving actors. I thought about doing one where a bunch of people are actors, but nobody knows it, and allows us to create very strange effects where they think that this thing is happening to the group, when in fact it's staged. But they would assume it's just other participants.
DAX: What if they know that it's an actor, but it's like a story they can follow where it's like, “Okay, welcome to the next room and you're going to witness some mystery happen,” that kind of thing. And you have to figure it out.
SPENCER: I think, to me, that gets more into art and entertainment and away from the social experiment. Because for me, it's really about putting people in a situation that's novel to them that teaches them something about themselves or about humans, rather than telling a story. I'm not that interested in telling stories.
DAX: That's really inspiring to hear about.
SPENCER: You know, I've noticed that a funny thing is when you're interviewed, you always end up interviewing the other person. Have you noticed that? I have seen you do that in almost every one of your interviews.
DAX: Yeah, sometimes people tell me that. I actually became a talk show host in the show called ‘The Hot Seat’. So my friend produced it and asked me to host his talk show. That was pretty fun for me, because I just got to be the interviewer full-time in that. I don't know if we'll make more episodes, but we filmed six episodes last summer.
SPENCER: You mentioned ‘Smoothie Madness’ and I’ve watched a few episodes of it. And I think there's something that’s sort of emblematic of something about what makes you really different from others. So my perception is, when you do things, you do them with this sincerity that people are not used to. They’re so sincere almost that people can't tell if it's sincere, or something like it's so sincere that people are like, “Wait, is this a joke? Or is this real?” Do you know what I'm talking about? Does that make sense?
DAX: I think it's because I have some shortcomings in terms of my producing skills. And so, there are naturally some errors that occur in the making of my game show. Like, it's a show that if you watch the first episode, you'll be like, “Okay, this is good.” In some ways, it's entertaining, but it's kind of wonky in some ways. Then if you watch the most recent episodes, ideally, those are the ones that are the most well-produced and you can feel that it's the most put-together. So I think I just naturally will make mistakes, and I don't stress too much about making it perfect. Because I'm like, “Let's get better on the next one,” that would just be my motto throughout Smoothie Madness. “Okay, the camera died.” The main camera died in Episode Six. So we had to use all the side cameras, which just made it so crazy. It's like the Blair Witch Project, in terms of the camera not being steady at all, which is not good for a game show. Those were supposed to be the B-roll cameras, but the A camera died (the wide shot). A professional show would probably reshoot that, but I don't have the budget to do that. “Let's get better on the next one,” so that's kind of my motto, usually.
SPENCER: I agree that is part of it. I think it gives us a sense of, “Is this a joke? Or is this real?” But I think it's something else, too. You're making it a game show about smoothies, just because you really like smoothies. And you're like, “What would I like? Okay, a game show about smoothies. So I'm gonna make a game show about smoothies.” That's the sort of level of how I interpret it. Is that accurate?
DAX: That's accurate. And actually, maybe that brings up a good point. It was a dream show to make, I love smoothies, and I wanted to be a game show host. I don't even remember when I had that idea. But I just knew I had it for a little while before I did it. But even as I was going to make the first episode, I was like, “I don't know if this is going to be good or not.” And I'm kind of nervous. So I don't know if I want to do it. But then iDubbbz and Dane, the people who made the documentary and produced the first episode of Smoothie Madness for me, they just did it as a favor to help out as part of the documentary. They're like, “What's something you want to make?” I was like, “I've been wanting to make this game show.” They pushed me to do it. So then I overcame the doubt and did it, which is kind of what ChatGPT has helped me do in some situations, such as going to the park to film videos. And then I ended up making a documentary I really like that I wouldn’t have made if an outside source hadn't pushed me more. So that's something that's helpful in my life.
SPENCER: I think another aspect of your work that I pick up on — I'm curious to hear your reaction — is that I think you lean into awkwardness, or you're not put off backwards the way that almost everyone else is. Most people, when they perceive something as being awkward, will flail around trying to get rid of the awkwardness as fast as possible. Does that make sense?
DAX: Yeah, I'm definitely not the smoothest. I definitely have improved certain social skills along that front. But I'm not the smoothest on that front. But I don't mind. I'm like, “You know what, just keep going, keep getting better as well on that front.”
SPENCER: Right. I don't think you have the strongest social skills and I think that's something you've improved on. But that's not even really what I mean, though. What I mean is that in the situations that you create, most people do just feel this really strong compulsion to smooth the situation over to reduce the awkwardness, and watching your work, you just don't do that. I don't know if you've noticed that but you don't behave like almost everyone behaves in those situations. And I think that's actually one aspect of what you do that's fascinating to people, watching you just sort of sit in that awkwardness and not try to get rid of it.
DAX: I think I try to get rid of it when I can. But I don't always know the perfect way to do that. So I don't really mind. I'm like, “Okay, well, that's a little awkward. That's just part of life. That's okay.” And then you just keep enjoying yourself. And just as long as no one around you feels weird.
SPENCER: Well, no, it's just funny to me to hear you say that because I feel like that's such a central part of what you do. You're just willing to embrace the awkwardness. You're describing it as like you're trying to get rid of it, which is strange to me. Have you ever seen Curb Your Enthusiasm?
DAX: Yeah, yeah.
SPENCER: Because that's a show where the style of comedy is really awkward. Like, there's an awkwardness to it, but people love it. It's a really, really popular show. And I think that you're tapping something similar with a bunch of your work. But talking to you now makes me think that you're not aware of that, or maybe I'm wrong. What do you think about that?
DAX: Well, I definitely am aware of what people think. But regardless of how much or… Yeah, I think, right. I don't know what I'm saying. Yeah, I don't know.
SPENCER: No worries, no worries there. It's funny, because that's almost like a perfect example, which is happening right now. I don't know, there's awkwardness right now in the conversation, and you're just kind of like, “Yeah, I don't know.”
DAX: True, true.
SPENCER: Awkwardness hangs there in the air. But I don't think it's something you should get rid of. I actually think it's something really charming about you that I think appeals to people, actually.
DAX: If people mentioned that, it’s always in a positive light. So it’s never a negative thing.
SPENCER: Yeah, exactly. I think it's just your style. So, if anything, I think you should lean into it because I think people like it a lot.
DAX: Earlier in the conversation, I had looked up, “What are some good things to say in the middle of a podcast?”
SPENCER: With ChatGPT?
DAX: Yes, yes. And I got a list of five suggestions. But I didn't say any of them. Because I was like, kind of exactly what we were talking about earlier. I was like, “Wait, but what if I say one of these things?” Like if I said, “I'm glad you brought that up, it ties in perfectly with what I wanted to discuss next,” then I started to talk about the next topic. You might feel deceived if I later said that it was from AI. So in this moment, that was a moment where I needed a blue-dot-white-dot-system for it conversationally because I've established that system on YouTube with my viewers, but I haven't established it with you.
SPENCER: We're getting to the end of the podcast anyway, but how about for the rest of this you can feel absolutely free to use different AI? And we'll just know that you might be doing that. How does that sound?
DAX: That sounds good. Actually, though, maybe the better plan is, if I say something from AI, I'll wait 60 seconds. And then I'll tell you, because I don't want [inaudible].
SPENCER: [laughs] That sounds good.
DAX: But also, I probably won't use anything.
SPENCER: [laughs] Do you want to tell us what it told you to say?
DAX: It just gave me some prompts, some transitions such as, “I find really interesting. Can you give me an example of blank topic?” And then the topics I have that are available. I have a bunch of quotes, I could tell. I know your audience likes “smart” quotes. So I could read some of those. And then topics on things like personal growth and self-development, career and professional development, entrepreneurship and innovation, technology and the future of work, mental health and wellbeing. So if I had been like, you're talking about something, and I'm like, “That's a great point.” And it leads me to ask, “How's your mental health been lately?” And then 60 seconds later, I'd be like, “So I want you to know that that was prompted after you finished your answer.” Maybe I'd say, “I just wanted to clarify that it had been prompted by AI.”
SPENCER: Yeah, I think that's a good approach.
DAX: So if I said to you, innovation is a hot topic in today's fast-paced and constantly evolving business world. In your opinion, what do you think is the most important aspect of innovation? And how do you think companies can foster a culture of innovation within their organizations?
SPENCER: Are you asking me that? Are you just throwing that out as the thing that you could ask?
DAX: Yeah, that's just a thing I could ask. That's me asking ChatGPT word-for-word, “Hey, can you write a question for a podcast host about innovation?” If I had more time, I would probably write the prompt a little better. “Can you write a question for a podcast host who does a podcast about Ted Talks?” That kind of stuff. I’d give more specifics, then get a prompt that way. That said, I haven't been using ChatGPT to that extent in my personal interactions, but I think I'm bringing that up to raise the point of — in the middle of this interview — I had asked for some prompts. And then I decided not to do it because this could feel disingenuous.
SPENCER: Right, right.
DAX: So I was like, “But I'll mention it altogether, because it kind of ties into what we were talking about earlier.”
SPENCER: You know, when you read those prompts, they just feel so stilted to me. Does it give you that vibe?
DAX: I guess so.
SPENCER: Like this sort of thing someone would put on a conference flier or something? [laughs]
DAX: Let's try this real quick. So what if I said to you, “Innovation can be a gamechanger in many industries. But it can also be a little intimidating. If you were stuck on a desert island with only three things and you could only innovate one of them, what would you choose and why?” Does that feel less formal? I prompted it, “Less formal, more fun, please.”
SPENCER: That's just — I don't know about you — but that just seems like such a weird question to ask. Does that strike you as weird? Like, I'm not saying you're asking a weird question. I'm saying ChatGPT is recommending you to ask a very weird question. Doesn’t that strike you as very strange?
DAX: Well, maybe if I was a little bit of a faster typer, and I tied it into something you were talking about. Could you talk for a second about some miscellaneous topic, and then we'll see if I can generate a natural sort of thing to say. I'm not going to tell you whether what I say is ChatGPT or not.
DAX: It will be a game, like your Twitter. Just say a few different things and I'll give you three different responses. And you have to guess which one is ChatGPT.
SPENCER: Okay, great. Will just one of them be ChatGPT?
DAX: Yeah, just one out of three.
SPENCER: Perfect. Okay. So I'll tell you about one more social experiment we did. So just this past weekend, we ran a social experiment, where we used a data set I collected, where we assigned strangers to talk to each other, and then had them rate whether they want to see each other again. And then we analyze a dataset to build a predictive model of who would get along well. And we just ran an event yesterday, where everyone who arrived was assigned to a cluster or color that they'd have to wear as a wristband. And people in the same cluster were predicted to like each other based on the algorithm so that people could more easily find people that they were more likely to get along well with.
DAX: Interesting, could you tell me a bit more?
SPENCER: Yeah so the basic idea is to try to create, essentially, a party-like environment, but one where there's a really good reason to talk to strangers, and you also know who to talk to. Because normally, if you're at a party, how do you decide who to talk to? Maybe someone looks interesting or they just happen to be standing there and they're easy to talk to. But by creating this wristband system and using this algorithm, we're trying to actually say, “Oh, no, you should talk to that person, because you're wearing a blue wristband, and they're wearing a blue wristband.” And it also gives you an excuse to talk to them, because they know that they're there to meet other people that are likely to be compatible.
DAX: Oh, that sounds like such a fun experience. What was the coolest part about it?
SPENCER: [laughs] The coolest part about it? I don't know the coolest part about it. I guess it was just getting people to meet people that they’d unlikely to meet otherwise and hopefully making some matches. We also collected data. We had people report who they are interested in so that we could improve our algorithm, hopefully, by crunching the numbers.
DAX: So what was the biggest thing you learned from it?
SPENCER: I think it would have been a bit stronger if it had been a bit more structured. It was a bit more like a freeform party, where people could just go up and chat with the people from their own color. But I think a little bit more structure would have helped, where we facilitate, “Okay, now everyone go talk to someone that they haven't met before. Okay, now go report how the conversation went,” etc. So I think it just emphasized for me the importance of structure in these kinds of social experiments.
DAX: So can you tell me which one of those three responses you think was the one that was verbatim from ChatGPT?
SPENCER: I'm not sure [laughs]. They all felt awkward to me. But maybe the second one was ChatGPT?
DAX: Yes. Second one, good job.
SPENCER: It just felt like maybe slightly less sort of a thing a person would really say. [laughs]
DAX: Okay. Yeah, true. Well, it's interesting, because I haven't used it in that way yet. And I don't know if I will ever, but maybe that's the future. Maybe people will someday use it more like that. Maybe me, included. I haven't yet. But that's something I'm exploring for the first time as well, right now. Maybe people will need a blue-dot-white-dot system for conversations.
SPENCER: You can imagine, especially with an AR display (augmented reality) if you had a Google Glass type of thing you were wearing, and it just was showing across the screen what you could say is your response so that you don't even have to move your head, you can just read it in real-time. That'd be pretty odd, but maybe useful in some situations.
DAX: And then on the other person's AR glasses, a white dot could flash to signify that that has been selected by AI.
SPENCER: Yeah, someone was posting on Facebook the other day about how, eventually, we'll just, instead of meeting people, have our AI talk to their AI and figure out if we get along well first.
DAX: That would actually be good. That's kind of what dating apps are really, like match.com or whatever, but just automated. How was your TED Talk? What was it like to give a TED Talk?
SPENCER: So I've given TEDx Talks twice, which I should point out that TEDx Talks are much less prestigious than TED Talks because TED Talks are on the big main stage of TED.
DAX: But yours had like 200,000 views or 300,000 views or something, like crazy.
SPENCER: To be honest, I hated doing a TEDx Talk. I think it was terrible. The reason I hated it is because it felt like a performance and I'm not a performer. It felt like a performance because I had to memorize what I was going to say. It was all completely regimented, I had to know every single sentence. And I hate performing. It’s just completely antithetical to my nature, whereas many people love it. And there's nothing wrong with being a performer. Lots of people love being a performer, it's great. It’s just not me. So I would much prefer to give a talk that's more casual, where I'm just explaining to my audience the idea, not memorizing my lines, and remembering when to use this vocal intonation, or to step forward, or whatever.
DAX: Gotcha. Okay, cool. That would be really hard. I did not get on a stage. That's interesting to hear about.
SPENCER: Well, you're an actor, though. So presumably…
DAX: I've only been on stage a couple times. I haven't done stage performances. That's a lot harder to do, being in front of a live audience.
SPENCER: Yeah, it was. I honestly really hated it. I would do it again, just because it's a good way to reach a large audience. But it was pretty miserable. I spent probably two days before practicing my lines for it. Every hour be like, “Okay, I have to review my lines again.” And then even during [quote] “performance,” I forgot the most important lines, I literally didn't say them. I just skipped over it. And it was the punchline or something. But the bizarre thing is, nobody noticed. People didn't realize there was supposed to be a punchline there. So it's fine. But I was like, “Oh, my God, I can't believe I left that line, that was like the big punchline.”
DAX: If ChatGPT had existed at the time, would you have used it leading up to it? Just asking for tips?
SPENCER: I doubt it, I doubt it honestly.
DAX: Gotcha. Last question. Do you want me to read any quotes?
SPENCER: Sure, why don’t we finish up with you reading some quotes that you like?
DAX: Sure. I mean these aren't even what I like. They're just general quotes that have been recommended to me.
SPENCER: What do you mean by recommended?
DAX: Like when I asked if there's any interesting quotes to say in a podcast.
SPENCER: Oh, from ChatGPT?
DAX: Yeah, yeah.
SPENCER: All right, let's do it.
DAX: “Success is not final, failure is not fatal: it is the courage to continue that counts.” Winston Churchill. “The best way to predict the future is to invent it.” Alan Kay. I'll just give a couple more. “The only way to do great work is to love what you do.” Steve Jobs. And then, “Your time is limited, so don't waste it living someone else's life.” Steve Jobs. And then, “The biggest adventure you can ever take is to live the life of your dreams.” Oprah Winfrey.
SPENCER: Thanks, Dax. I really enjoyed chatting with you. I really appreciate you coming on. And I wish you great luck.
DAX: I was excited when I saw your message. I was like, “Well, I haven't been invited on a podcast like this before.” And thank you very much for the talk. It was a really easy podcast to be a part of. I was a little nervous beforehand, so thank you.
SPENCER: Yeah. Did you use ChatGPT to decide whether to come on or did you just decide on your own?
DAX: No, I just decided on my own. I was just like, “That sounds good.”
SPENCER: I was wondering about that. It's like, “Did he ask ChatGPT whether to come on the podcast or not?”
DAX: In a way, you could call it a blue dot decision. And that ChatGPT has recommended for me to collaborate more and put myself out there, like do more interviews. So I suppose, in that sense, anytime I get a good opportunity along those lines. That said, it's a decision I would have made the same with or without it. But there's that aspect as well. But that said, if ChatGPT was like, “Focus solely on your daily routine for the next week,” then I probably would have said, “Okay, I can't do it this week.” But that hasn't been advice that's given to me. It's just been good advice that aligns with what I was trying to do.
SPENCER: Yeah, makes sense. Well, I'm gonna follow with great interest your ChatGPT experiments, and yeah, wishing you the best of luck. I hope you're able to make this sustainable, long-term career on YouTube.
DAX: Yeah, I hope so too. We'll see what happens next. And I hope that you get to do some cool experiments soon and some more of what you enjoy doing most. Thanks.
JOSH: A listener asks, do you ever censor yourself or your podcast to avoid social justice ire to some extent, or do you feel you express yourself pretty openly and honestly?
SPENCER: I feel that I express myself quite openly and honestly. I would say the main thing is if I'm talking about a topic where I feel like people are likely to be sensitive around it — let's say we're talking about rape — then I'm going to try to speak more carefully to avoid accidentally upsetting people unnecessarily. There's no reason to upset people if you don't need to. Obviously, there are times you have to upset people if you're challenging a system, but I try to avoid unnecessarily upsetting people for no benefit, certainly no benefit for them, no benefit for myself. I do try to speak carefully, but I don't feel that I'm taking some belief of mine and just suppressing it or something like that. I will also add that on the podcast, primarily, I'm more focused on making sure that I get the guest's opinion than I am on making sure I get my opinion out. So often, I will not say my opinion, but merely because it's more interesting to focus on the guest's opinion, right?
Click here to return to the list of all episodes.
Sign up to receive one helpful idea and one brand-new podcast episode each week!
Subscribe via RSS or through one of the major podcast platforms:
Apple Podcasts Spotify Stitcher Google Podcasts TuneIn Amazon RSS
Host / Director
Broke for Free
Quiet Music for Tiny Robots
Please note that Clearer Thinking , Mind Ease , and UpLift are all affiliated with this podcast.