Ep. 238 - Boys Club LIVE: The Pope on AI, Enhanced Games, Ferrari's First EV and guest Jacob Taylor from Brookings Institution on "Context-maxxing"

[00:28] This week's show. [00:30] you [00:32] I have something... [00:34] that I want to [00:35] call you in on. Oh, okay, great. Yeah, that sounds really inviting. I'm excited. Okay, so you and I, [00:42] Share locations, sort of. The devil's in the technology because it constantly keeps breaking. [00:47] Yeah, you tend to blame the technology, I would say. I tend to blame the technology, but I'm not doing anything. So it feels, for me, it feels it must be the technology. [00:55] You're not doing anything, so you're not putting forth any effort to share your location with me. Yeah, that sounds right. Okay. And you tweet a lot about Nomad Flatiron area. Yeah, totally. I love that area. I refer to it as Flatiron slash Nomad. That's, like, really important for me. Okay, you're starting with Flatiron, then Nomad. Yeah, because Flatiron is better, and then Nomad is just kind of, like, it got tacked on because of the business improvement district. There's some history and lore that goes into that, but Flatiron is more important. [01:26] while looking at your location to see you in Flatiron slash NoWhead? You know, that's fair. I tend to go there for like... [01:33] one hour or less. I just go to Madison Square Park basically and then I leave. Do you go to the Shake Shack there? [01:40] I have never... [01:51] Very good. And then, like, sometimes I'll visit offices. It's just, like, where a lot of tech startup offices are. Yes. So sometimes I'll... [01:58] Okay.

[01:59] But not for very long. Just kind of pop it in. Yeah, if anyone has an office in Flatiron that you just, like, want someone really fun and exciting to go to, like, I would say maybe just DM me on Twitter. Bring the electric atmosphere. Yeah, exactly. You're the guy. I can do something. Yeah, just tell me what you need from me. Okay. I just wanted to shout that out. Trying to call me fake, basically. Basically, everything you see on the internet is not real. Okay, so this live stream is brought to you by two wonderful partners, Polygon and Vercel. [02:29] And we're going to talk about Vercel a little later, but do you want to talk about Polygon? I... [02:34] Literally love talking about Polygon. Let me just figure out what I'm going to say by scrolling my phone. You do the ad reads from your phone. That's cool. That's maybe something I should start to implement. [02:45] Yeah, I just find that it's hard for me to use a laptop. I used to do my whole job on my phone, actually. Oh, really? Yeah, that was amazing. But no, I don't have a job, so I don't have to worry about that. Kate, do you think that you can pull up the Polygon graphic while this goes live? [02:58] the polygon graphic while this goes live. [03:02] And Quasimatt is going to do a great job telling us about. Okay. Let me tell you something about polygon. [03:10] Every company moving money globally hits the same walls. Legacy rails are slow, expensive, and built for a pre-internet world. Stablecoin rails are fast, but they're fragmented across vendors that don't talk to each other. The plumbing to deliver on the promise of stablecoins hasn't existed. Polygon's open money stack is fixing that. One unified stack that puts global money movement onto proven rails that move at the speed of the internet.

[03:31] And why Polygon? Because they've been moving money. Check out the link in chat to learn more. [03:36] We love Polygon. [03:38] With a fly on. [03:39] Great job. We really do. I do have to bring up another geographic moment that I had today. If you would have checked my location this morning, you would have seen that I was in my neighborhood. And I do sometimes see celebrities who stay at the Trendy Hotel near my apartment. Oh, I was there this weekend. I stayed there this weekend. Well, you stayed at that hotel? Yes, I did. Why? My mom was in town. We had a little girls weekend. Oh, okay. Yeah. Wait, your mom's so trendy. She's a trendsetter. Yeah. She's really, well, she asked me where should we stay. [04:09] Well, [04:11] And if we had our location shared, you maybe could have met my mom. Yeah, that's so true, actually. [04:16] Anyway, Addison Rae grabs a city bike right outside my apartment. So I've seen her twice this week. Whoa, twice? That's huge. Okay, wow. I really would have loved to... [04:27] If I had spotted Addison Rae at the breakfast... [04:30] situation at this hotel. It very well could have happened. It could have happened. That would have been really cool. I would have been really excited. I... [04:38] wouldn't have done anything but i would have yeah i didn't do anything i just like turned to my friend and i was like oh that's addison ray [04:43] That's a cool move. Yeah. Today, this morning, she was wearing her hat really low. So I was like, oh, like either you're worried about like the sun or you like really don't want to be seen. So I was like, okay, I'll let you grab your city bike. [04:54] Yeah. She took a while to do it, though, I will say. Like, I saw her getting it, and then I went off into my apartment, and I looked out the window.

[05:01] And she was still getting it. Okay, so Addison Rae is slow on a city bike is what you're trying to say. Grabbing, yeah, grabbing a city bike. I love her. I'm literally obsessed with her. It's not her job to get a city bike. It's her job to make, [05:12] amazing music and to entertain and she does that better than maybe anyone else on planet earth yeah like i would say [05:18] Yeah, she's certainly the top 1% entertainer. So if she's not the most efficient at getting a city bike, that's okay. That's really fun. And I will have to say, sometimes it's the technology. The city bike is very slow. Yeah. Yeah. Sometimes you'll try to dock it. It doesn't work. Or you'll try to get it, and it's just locked in there. Exactly. It's a dock sometimes. Yeah. Okay. We have a lot to get through. We have a great guest that's coming on, Jacob Taylor, who will be here in 20-ish minutes. [05:42] Perfect. 20 inches. Thanks for letting that slide. I was getting into my AI safety researcher bag in preparation for that. Okay, great. I feel like I never really... [05:53] read about ai or anything but now that i've read i've read like i think it was actually three pages and i literally read it i don't know how much of it i remember but now i feel like i can go to parties and say yeah i'm actually an ai safety researcher you are after this conversation i'm like a research collaborator really exactly you you're an academic we are yeah we are okay me too um okay yeah he is a literal academic he is working at the brookings institution and he [06:23] great conversation. Before we do that, you and I have some things we're going to share and yap about and talk about here collectively. Some global news. Do you want to go first? Should I go first? I think you should go first because I need to do my two back to back. They're married in some ways. Okay, so mine is a reference to

[06:42] Williamsburg, really. Okay. So there's this thing that happens online that I really love called [06:46] Dating discourse. Okay. And this time, the discourse is about whether you should... [06:53] It's like this girl. Here, let me just read it. Let me read it. Okay. We're finding it. Okay. Okay. We're pulling it up here. It's a Reddit. [07:00] post. While I'm doing that, how do you feel about, do you participate in dating discourse on [07:05] X, do you find it fruitful? What is your experience with dating discourse at large? Okay. I wouldn't say that I participate. [07:12] I would say I sometimes keep up to date on the discourse. Okay. I see what people are saying. I like to see how people are weighing in. [07:20] I don't think I've ever shared my own opinion before. [07:25] All around it. Oh my God, we're getting a first from you because you're going to have to share your opinion on this dating discourse. Maybe on this live stream, but like I've never written out my opinion, I guess would be one way of saying it. But I find it really interesting. I love to know how other people do things. I think people are so weird and wonderful. And so I want to know. [07:43] all of the things. You know what I'm saying? Literally. Okay. I'm really, really struggling to pull this up. One second. Okay. I think I can actually read it on the screen. Oh, I have it. I [07:54] Williamsburg dating discourse. Here we go. Okay, here we go. This is a Reddit post. Do men commonly send Ubers to bring women to dates in Williamsburg? I was talking to a woman, 27F. [08:03] for a few weeks, had been on two dates in Manhattan and FaceTimed a few times. I invited her to Williamsburg, my hood, [08:09] for a brunch date. She said yes, and then asked how she'd get there and if I would send a car for her. I assumed this was a joke since she lives in Greenwich Village, so like 15 to 20 minutes away. So I laughed and said I had full faith in her to make it to Brooklyn. She quickly called off the date and broke things off. Is this a common expectation of NYC dating? I could understand calling an Uber for my date at the end of the night if it's late.

[08:31] There are no good subways, etc. But for a daytime date, when there's an extremely easy subway trip, makes no sense to me. [08:39] Mm, okay. [08:41] I want to hear your thoughts before I share my thoughts. Okay. Well, also, it blew up. [08:45] Yeah, everyone's commenting on it. My thoughts are this is entitled, narcissistic, disgusting. [08:53] Um... [08:54] Gross, uninformed. [08:58] So I basically think like... [09:01] you should take the subway for three stops. So Greenwich Village, famously off of the L train. Williamsburg, also famously off of the L train. Like, you can take the subway for three stops, and it will literally be faster than an Uber. Right. So I think there's, like, a... [09:13] lack of logistical knowledge and that's like where we're starting but then okay let's transcend that and talk about the sentiment which is like I expect that [09:20] the guy that I'm going on a date with is going to get me an Uber to the location. And then I'm still kind of like, well... [09:27] Not really. Like it's one of those things where like if someone is offering it, that's electric. Like that's amazing. But like if you have to like do this weird negotiation to get what you want, then maybe just like calm down. Okay. Yes. Okay. Okay. Really, really interesting. Okay. A few things. [09:45] This is Princess Behavior. [09:47] From this woman. Right. 27F. Yeah. And I... [09:53] Would never... [09:55] do this? Never. In one million years, like... [09:59] But... [10:00] I kind of feel like this is a test.

[10:03] He failed. [10:04] That's actually good because it means they're not a match because there's some women who, [10:10] who expect this and want this and some men who, [10:14] want to do it. And those people need to find each other. Yeah. These people are not meant for each other. Her asking, how do I get there? [10:22] is honestly, I would be like legitimately embarrassed if I had sent that text to somebody. I would feel shame. And I think it's like we're getting the guy's perspective and his framing. Like, I think there is a genuinely normal way to go about this where you're like, [10:34] hey, like, I'm wondering if you'd be willing to grab me an Uber, like, maybe I don't know, come up with some fake reason. Like, I'm really scared of rats on the subway. Like, I don't know, put a little bit of like, courtesy into the request and then be like, totally chill if he says no. Right. Or, or what you're saying is be like, oh, I actually have like certain standards for how I interact with people on dates. And like, this isn't meeting yet. Like, [10:54] You know, it's kind of a callback too. It's that video that we watched of the person who was like, I... [11:00] Um, [11:01] That was like walking around their apartment and apartment building. Oh, the rich auntie. Yeah, it's very rich auntie behavior. It's rich auntie behavior. Okay. It's like entitlement behavior. It's entitlement behavior. I find that really, I really don't. [11:17] like it, but I also don't want to be judgmental around it because I do feel like [11:21] I... [11:22] I think it is net positive for people to elevate their expectations around how they are treated generally. [11:30] But I think that this is not something I would do, and I think it's ridiculous. I think...

[11:35] This is just dating advice for heterosexual men generally. [11:40] Offering for the Uber home. [11:43] is like the easiest $10 to $25 you will spend. And it's such a great move. Right. High ROI. Just do it. On that behavior. Just be like, can I get you a new home? [11:52] a lot of people will probably be like, "Oh no, I like to walk or I like to take a subway or whatever." [11:57] Some people won't and they will take you up on it. And those people will be like, wow, that was so, so nice of him. [12:04] just do that. Just do it. And like, if the $25 is a big stretch for you, [12:10] Then maybe there's another way to go about it. Stop dating and get your bag up. What are you doing? [12:15] There is this other greater thing that I think is like, I don't really think. [12:19] It's bad to like what you're saying. It's not bad to have preferences. What is bad is to like perpetuate ambiguity about your preferences when they're nonstandard. So it's like people think people think that. [12:30] it's okay for like me to have my expectations in my head and the other person have their expectations in their head. And then we just like, [12:37] expect each other to execute on that when really it's like okay you're using a dating app that has a profile that gives you an opportunity to list what your preferences are i would say [12:46] Just do all the filtering right there. I would maybe even say write like a two-page PDF. [12:51] that says exactly what you want so you can just filter people out quick because now... To be PDF. Downloadable PDF. Yeah, that has some expectations and then maybe list what's negotiable and what's not. And then before you go to the first date, I would suggest, [13:04] calling, having a negotiation, and being like, hey, this is what I'm looking for. What are you looking for? And having everything really established. I think people love to think this is just going to happen to them. But it's like, well, you're on hinge. So it's not. You already lost the magic of the moment. You might as well just admit that it's a negotiation, and all of your relationships are really transactional, and dating is just kind of a way of scamming someone into adding value into your life and stuff like that. And just work from there. Work from there. I think...

[13:33] I always feel when someone offers me an Uber to a date when they're like, [13:38] Would can I call you a car? Yeah. [13:40] I always feel like it's a test. [13:43] Like, I feel they're testing me. And how do you pass the test? Like, what do you imagine? By saying, no, I can take the subway there. [13:50] No problem. [13:52] I would assume the opposite. I would assume that if they are into providing something, then they're going to want you to say yes to it. But if it's like... [14:03] It's so complicated because it's like maybe they are assuming an expectation that you have and meeting that expectation despite the fact that they don't really want to. Right. So then in that case, what you're doing is actually great and amazing. [14:16] I think that the only answer is to be very true to yourself. And so if there's a moment where I'm like, actually, you know what? I would like an Uber there. [14:23] I'll take them up on it. Never in a million years will I say, how do I get for the question? Yeah. Oh, my God. It's so beautiful that we like, oh, what's the game theory of dating? Well, maybe it's not about that. Maybe it's just about being you. Being real. What was the general take online? Well, there were a lot of people who were saying like, oh, this is like something that [14:42] uh sex workers do and like normal people shouldn't be normal people that non-sex workers i would say dating is sex work but whatever um shouldn't be doing that okay uh so that was like a and then people were like oh well is that really true like what does this actually have to do with that at all and like okay it seems like you're just trying to like make it seem like this extremely like okay not extremely normal but you know this relatively innocuous behavior you're

[15:12] Thank you. [15:13] side. Yeah, I mean, you're also on [15:16] a really masculine coded, like you're on Reddit. [15:20] Oh, I didn't even think about that. You're going to have a lot of men weighing in as opposed to like if that were on TikTok, I think there would be like a very different response. Yeah. Potentially. [15:28] Yeah, that's true. Because on TikTok, there's the whole universe of... I added another tweet. I don't know if you pulled up the other tweet, but there's like this universe of... [15:37] what is it called? Sprinkle, sprinkle? Because there's this lady that gives dating advice on [15:42] on TikTok and she'll just be like, scam him, scam him, scam him, like use him. Like that's like her whole shtick. And everyone is in the, like, I know women who like stan her and are like, yeah, like, whoa, I'm going to like, yeah. [15:57] Use him. Yeah, like get anything I can out of him. Okay, okay, okay. And I'm kind of like, okay, this is like a... [16:02] I think we've talked about this before. There's a certain type of person that can get away with doing that because you're just so clouded and amazing and everyone always wants to be with you and you're basically a celebrity. And everyone sort of imagines that that's them. And it's simply not. See, the reason why I would refuse the Uber app [16:20] initially is because my worst fear is being perceived exactly as you just communicated. [16:25] Someone who thinks that they are this type of person and they are really not. And so – [16:30] I'm doing some mental gymnastics around that's, that's what's happening here. And I need to make sure that they know that I know that I'm not. Yeah, no, that's so that's real. Yeah. Thank you. Yeah. Okay. Really interesting stuff. I think that we should move now to the Pope.

[16:46] Um, [16:48] What does the Pope have to say? Seamless transition. [16:52] Okay, so the Pope weighed in on... [16:56] on AI on Monday. So Pope Leo the 14th, [17:00] issued a [17:01] His first major, I had a moment where I said his, and I was like, oh my gosh, what is the pronoun? Is it Papa Woman? [17:09] Am I being unwoke? Did I misgender the Pope? Isn't it literally like the Pope actually has to be a man? I think, yeah. I don't think a woman can be with the Pope. Because people were freaking out when, wait. [17:19] when the Pope was going to be like... [17:21] black or something. I think people were freaking out. And then this pope is actually [17:26] an American. And that's a really big deal. Oh, yeah. And people think he's going to run for president, I think. Oh, really? Wow. I mean, I just mean, at least one person thinks that because someone told me that. But I'm not really sure. Okay. So on Monday, Pope Leo, the 14th, issued the first major theological text of his [17:45] Papacy. [17:46] Okay, we're going to define some things. I know this is not your world. Yeah, this is new terminology. Warning about the growing power of artificial intelligence and calling for stronger regulation of the technology. Okay, so he released, every time I say he, I'm nervous. I don't know why. Okay, he released something called an N6. [18:06] Encyclical, which is essentially like a formal letter from the Pope to address the people and the church. And it is 42,000 words.

[18:23] This one on AI. And a papacy is basically referring to like the authority of the Pope as the head of the Catholic Church and sort of. [18:35] what the institution of the Pope is like. [18:38] decreeing essentially for their time in office. God goes, stop AI. And then the Pope goes, stop AI with a big AI slot paper. That's 42,000 words long. Got it. Yeah. Basically. No, I mean, I, I mean, I guess that's one way of saying it. Yes. I feel like that's what I heard. I don't know what nuances you're interested in adding. Okay. So his encyclical, [19:03] covers. [19:05] Basically, theology and doctrine... [19:08] Ordinarily, generally, as like a piece of literature, theology and doctrine, social justice, ethics and morals, politics and society and sort of like the environment or the care for creation. So like everything. Everything. Yeah, for sure. Big remit on that one. [19:25] Okay, and so it was called Magnifica Humanitas. [19:31] Okay, pretentious. Okay. And basically... [19:36] what the whole paper is on is that obviously AI is ushering in a new industrial revolution. So [19:46] Him, he, as a pope, is... [19:51] Hearkening back to the previous pope before him, Pope Leo XIII, who wrote his...

[19:59] paper his decree on the industrial revolution on the industrial revolution okay so he's kind of like harkening back to that and like it's a call it's a callback he's like remixing exactly um [20:12] And I wouldn't say it's all anti-AI. I would say that there's some aspects of it that are somewhat like pro-technology, but it's definitely like a warning call. And some of the things he writes, we're living through a rapid crisis. [20:28] phase of transition, a change of era in which while some are vying for the future of new technologies and others dedicate themselves to reflecting on the matter, some people are watching and waiting, observing from afar and merely hoping for the best. And it was sort of like a call to be like, [20:45] collectively deciding like we need to engage with this issue yes the issue of artificial intelligence yes um and that's that's cool um so basically a few key points here um [20:59] He... [21:02] Sort of. [21:04] references, references, [21:05] This is going to be a deeply biblical reference. So I don't know. I love. I mean, I have so much biblical education. I honestly might be able to clock it. Oh, really? I might be able to tell you what book it's from. Oh, my gosh. Wait, what? I mean, the chances are like 1%, but let's try it. Okay, okay, okay. [21:21] Constructing Babel and rebuilding Jerusalem. Oh, well, the Tower of Babel, I do know, was when God made all the languages. Exactly. Yeah, people were building too high. Oh, my God, wait, this is so, I love that he used this. Yes. It really is the same thing. It is.

[21:35] Gobble. [21:36] Humanity builds high into the sky in an attempt to touch God. Yep. AI. Humanity goes on their laptops for a really long time in an attempt to become God, perhaps? Question mark. It's really the same. It's really the same. But this time, is God going to intervene? Because when they built the tower, he said, well, y'all can't communicate anymore. And this time, maybe he's... [21:54] Sending a virus or something like that because he's got to do something to distract us. Just a drought maybe. Yeah. Or I guess he's intervening by the Pope. [22:02] By the Pope. Exactly. [22:04] God, hope, hope. [22:05] Humanity. Okay. He calls for the urgency of the moment. There's a bunch of stuff on that. He talks about the... [22:13] warns against ai and warfare there's a whole big section about uh about just war and his theories on the war and on war and how technology is being used um to harm people and so there's like a whole thing on that and like the human control of weaponry was the thing that he discussed there um and so it was he pro or anti-war he's anti he's anti-war um so like but now i this is [22:43] kind of proposing that you could do war with AI or war without AI. And then it's like, oh, we can't do war without or with AI. And then that kind of like justifies the war without AI. Like, hey, guys, it's okay to fight, but use a sword. Okay, let's see what he says here.

[23:00] he's kind of speaking to that. He says, the just war theory, which I'm unfamiliar with, which has all too often, [23:09] which has all too often been used to justify any kind of war is now outdated. So I do think he's saying that like we're in this moment of rapid, [23:19] technological advancement and the areas in which it's touching are one area in which it's touching that is like direct human, um, [23:28] Harm is... [23:30] The... [23:31] our ability to better control weaponry. But you can make the argument, [23:36] that that's making it more efficient. [23:39] Yeah, or there's a certain type of way in which technology can relocate the conflict from actual... [23:45] harm to like a Cold War type situation. Yeah. [23:48] Yeah, I mean, I don't know what he's saying about that, but he's basically just saying like, hey, Palantir, like, I don't really like that. Palantir, no, no, no, no, no. Yeah, that's like my take on that. But I don't know, because also it's like, if part of the message is, hey, guys, like, we need to pay attention to AI. [24:02] That's pretty... [24:04] Pretty empty, if that's a significant portion of what he's saying. I mean, it was... That means that I haven't read any of it, so... Right, totally. That could be on my end. I think... [24:13] You're coming at it from someone who is deeply, deeply engaged in these things. So true. And so I think a general call to pay attention to what's happening. Oh, wait. It's so true. Yeah. Okay, great. Yeah, because we like think and talk about AI all the time. Yeah. Like your average Catholic person. [24:29] Nona. I don't know that she's thinking about AI that much. Right. That makes sense. So anyway, a lot here. I do want to transition because I have a perfect segue to my next topic, and I just want to touch on it quite briefly. Okay. So...

[24:46] Big week for the Pope. He has this... [24:49] paper come out. Right. Normally no one cares about the Pope. That's also not true. Oh, okay. That's another bubble that I live in, I guess. Okay. So he comes out with this encyclical clue, or whatever. I keep pronouncing it wrong, but we get the gist. The document, yeah. The document, people are talking about it. Then yesterday, big day for both him and Ferrari, because Ferrari comes out with their first electric car, in which the Pope actually [25:18] tours the vehicle. So we actually do have a tape of it. It's quite... [25:24] shocking that these two things that I was going to talk about like had it feels like there was like a [25:29] like a glitch in the simulation that this happened. Wait, he's doing a tour of the interior of the car or he's doing a tour... [25:35] around in the car both [25:37] Yes, and we have a video. We can play the tape and you can watch it on here. [26:07] Anyone here? [26:08] that. [26:09] look sure. [26:11] Thank you. [26:12] Afebe [26:13] I don't know if he gets. [26:15] So that

[26:17] Yeah, I don't know if I just took one off. [26:20] They are really interface with Agrizo and Lutzo as well. [26:24] He's on the red. [26:26] Okay, a few things I want to say. [26:30] Here. [26:31] Yeah. The time one, this is just so funny to me. What are we doing here? Like, [26:37] Do you know what I mean? Is he doing commentary on, like, the car? Like, is he selling it, or is it just, like... Someone is kind of selling it to him. Not, like, actually selling it, but, like, explaining what's going on and talking him through. This man, as you can imagine, quite a stoic figure. He's not enthusiastic. He's not really giving... [26:54] anything. Yeah, which is like what you would expect. Like, that's not his bag. What you would expect, but also... [27:00] Part of what is so funny to me about it is the internet... [27:04] hates this car. People are so mad. It's called Loose. It's the first... [27:10] electric car that Ferrari is putting out. [27:14] It's a five-seater. People are upset about that. How many seats is this supposed to have? I'm so sorry. It's on Do Not Disturb, but my sister pushes through no matter what. Notify anyway. She's like, it's time to talk. Okay. It's designed by Johnny Ive and Mark Newsom from Loveform, you know, that famous design firm. No. You don't know that designer? No. Can you educate me? Like, what else have they made, perhaps? Wait, Johnny Ives? Well, I'm anti-design. [27:42] Okay. Well, that's a neat and tidy excuse for not knowing who Johnny Ives is. Yeah, I have a neat and tidy excuse for most things. Okay. Basically, he was one of the main designers at Apple. He is credited for all of the...

[27:58] I think not correctly credited as like the mastermind behind design. And a big take that I was seeing on the internet around the Ferrari was the [28:09] oh, this is just proof that Johnny Ives was, like, right place, right time. And there's something really comforting about that, that, like, you could just be in the right place at the right time and become super famous and, like, actually you're, like, maybe just mid in terms of your talent. Yeah, yeah. So that was kind of, like, a take the internet had. It's extremely ugly. Yeah, this is a great example. Like, the car, like, with the plug. Oh, I saw that tweet, yeah. Yeah, it's funny. So now you know what they're talking about. Anyway, people hated it, and then the Pope kind of being, like, [28:37] super laissez-faire about the whole thing only like perpetuated that because everybody was like [28:42] The Pope's not impressed and neither am I. And what is this vehicle? [28:46] I think... [28:47] I don't think it's... [28:49] How do you feel about cars? How much of a car guy are you? So I moved to New York so that I wouldn't have to use a car, basically. So I would say I'm pretty anti-car, although sometimes when I get the chance to drive, I think it's fun to go fast. [29:03] Cool. Okay. Yeah. That's like my thoughts on driving and cars. So you don't have like a car that you would like? [29:09] Um... [29:11] If money were not an object. [29:13] I've never thought about it. Okay. Maybe that one, just to be contrarian, honestly. [29:17] Definitely something electric, so probably that. Okay. I will say the TechCrunch article that talked about the – [29:26] response was saying that

[29:28] There has not been this much of an outward hatred of a vehicle since the cyber car. Yeah, yeah. Okay. I think Cybertruck, whenever I think like notorious vehicle that everyone hates. Notorious vehicle that everyone hates. I see them all the time though. So they have fans. They have fans. One time I saw a tweet that I will never forget. And I think about it every time I see a Cybertruck. It was like, you'll just be going about your day just fine. And then all of a sudden, bam, Cybertruck. [29:58] Anyway, people hate it. [30:01] People are upset at Johnny Ive. I think you should do a deep dive. I think you'd have a lot to say about him and what he's up to. So he made like all the, like he made like the MacBook design and like the design thesis for Apple in general. I think it's like, okay, here are some companies that he is credited for. [30:18] Apple, so they have a love form has is a firm and then they are like contracted out clients out and Apple is their first client and worked with the company until 2022. And it's not publicly disclosed which projects they were on. But yes, it's like, OK, he was a part of the iPhone. He was part of the MacBook, whatever. OK, yeah. [30:40] Airbnb, Ferrari, Montclair, Christie's, uh... [30:44] Those are the biggest ones. There's some other. And OpenAI. They're doing some stuff with OpenAI and like hardware around OpenAI. And that was like a big thing when people were like Johnny Ives and Sam Altman. They had like a whole video where they like went to a Paris cafe. It was very weird. [30:56] Anyway, I'll send you all the links. And this is his first big flop, basically. Not this big moment. I'm proud of him. How is he navigating? I want to see what he's posting like right now. I want to see how he's posting through it. Posting through it? I don't think he posts. I think he's like... He's too cool to post. He's too cool. He's too designers. Yeah, it's so designer to be too cool to post. Totally. Okay. We are now going to bring up our guest, which I'm very excited to talk about, Jacob Taylor.

[31:21] Come on down. [31:22] Oh, wow. [31:26] Of course. How's it going? Good to see you. Matt. I actually kind of forgot that you have a British accent, which is... Oh, okay. I mean, to an American, it's all the same. Okay, before we get started here, I'm going to tell everybody a little bit about our sponsor. [31:53] who we love and who is also working on a show with us that we will show the trailer at the end [32:01] Are mics hot? They're hot. Okay, great. Okay, Vercel is one of our partners of the show. If you are a startup building with agents, look no further than Vercel. Their agentic infrastructure gives you and your agents everything you need to ship. And if you're supported by one of their hundreds of VC or accelerator partners, you have access to exclusive discounts and benefits. [32:31] out the platform that thousands of companies use from day zero to IPO at vercel.com. [32:37] Slash startups. You corrected me last time. That is not backslash. It's slash startups. Slash startups. The link is in the... [32:45] chat. And at the end of this episode, we are going to roll a trailer for this week's Show Me Your Stack episode with them. But...

[32:55] Before we do that, we're going to chat with Jacob. Thank you for being here. You have a very, very impressive resume. Very well-rounded dude. You are a fellow at the Center for Sustainable Development at the Brookings Institution. You completed a PhD from Oxford in Cognitive and Evolution. [33:14] evolutionary anthropology mouthful yes wow great okay and you played professional rugby for three or four years that's right yeah back in australia yeah okay that's right um so you got a lot going on just see what's inside my head um thank you for being here thanks for having me yeah very excited to talk with you about um the paper you recently published before we get into that can you just [33:44] Brookings Institution is... [33:47] one of the world's most well-known think tanks, and we think about public policy. So, [33:52] basically trying to come up with research, insights, tools that – [33:58] leaders can use across organizations and governments. [34:01] to make the world a better place for people and planet, not... [34:05] When we say public policy, it's really like, how do we give people... [34:10] and the world we live in more resources and power okay and how long have you been there i've been there for almost six years now oh my gosh wow yeah it's been a little a little while um okay but have [34:23] Yeah, I've been there long enough to kind of get a sense of the game of policy and increasingly find myself trying to get out of it and into places like this to talk more about policy

[34:33] you know, things like technology and how that shapes the societies we live in. Is the scope of your work like entirely technology or do you do other stuff? No, I'm an anthropologist of team performance originally. So human behavior, trying to figure out how, [34:48] You know, I was as an athlete in another life, [34:51] was fascinated by how sometimes when teams work, really work well together and what's the special source for that. Okay. Oh, interesting. And that, so I got brought in to Brookings to help build, [35:03] new ways of collaborating with, [35:05] for these kind of big policy issues that the world doesn't have a great answer to at the moment. So it's like, you know, how do we end extreme poverty? [35:13] Turns out we actually know how to do it. [35:15] We've got the technology. It'd probably cost us about 1% of all billionaires' wealth. [35:20] To just like end it tomorrow. [35:22] But the key missing piece is like how to bring people together and, [35:26] because so many people have different pieces of the puzzle. How do you bring the people together in teams? [35:31] um to really like chip away at the problem and build it up it's a coordination issue coordination problem yeah interesting yeah um [35:39] Okay, and then... [35:41] How did you go from that into getting into... [35:44] Technology. Yeah. So when you think about those big global problems, it's like you need the humans collaborating, coordinating. [35:52] But really to get to the scale of some of the solutions, you need [35:55] technology too to help. [35:57] Um, [35:58] you know go further i guess and get to scale and so before brookings i was actually working on a

[36:03] a DARPA research program, which is the Department of Defense's moonshot factory for [36:09] cutting edge technology, and we were building an AI teammate. Okay. And that was the moonshot. [36:14] And so I came with enough [36:16] knowledge about ai to be dangerous but not enough to like really build models and so on and so that just like my team stuff my ai stuff just found a place in brookings where you [36:26] Um, [36:27] you know, that combo of human, the best of human teamwork and collaboration and the best of technology. [36:33] could be one like narrow path through a lot of the stuff that we're seeing in the moment yeah cool that's awesome um okay so you just released a paper called content maxing a path to cognitive agency with generative ai context maxing oh what did i say content oh okay context maxing and um a lot of [36:58] And there's a lot to this paper. It's impressive and dense. And I... [37:05] a lot of what you were writing about is through the lens of policy. Yeah. Okay. But can you just give us broad strokes? [37:12] what was the intention of the paper and yeah, how you guys got started. [37:15] I mean, honestly, my intention with the paper is like, [37:18] are we ready for like a people's revolution with ai you know that like at the core of what we were trying to write it was like [37:26] I think what you saw earlier this year with... [37:28] the [37:28] emergence of open source harnesses. So like the open core moment, [37:33] Erma's now coming through.

[37:35] It was bigger than just another tech or another piece of software to us. It actually showed... [37:41] a very different way of working with AI. [37:44] that I think if we could really support that at this moment, could give people more power and not less. [37:50] So that's like under the hood where like, [37:53] guys, let's jump on this and, like, empower people to do more of this kind of homebrew garage band AI because – [38:01] In doing so, you actually give people the skills and competencies and [38:05] agency to like [38:07] use the technology in the way that [38:09] people want to use it rather than kind of feeling like they're just taking it from technology makers, model builders, whatever. And so it's really trying to like, it's a power, it's about power ultimately. And that's agency is just a, [38:21] a jargon term for power. It's just like how do we give people more power in this relationship? [38:26] Yeah. Big, big goal. Yes. I love it. I respect that so much. And I think there is a real sentiment throughout the throughout the writing and the research around sort of like a positive view of what's possible. And yeah. [38:44] There's sort of [38:46] you talk about there being like two schools of thought, essentially one being that as people are using these models there, you don't use these words, but these are my words. There's sort of like cognitive decline. And that like, as people become more and more reliant on these tools, they're using sort of less and less of their own agency, their own like brain power in order to problem solve and do work and whatever, a variety of things. And then on the other side, there's this,

[39:15] this other... [39:17] group of research that's coming out that's showcasing that actually it can be beneficial to humanity, to people's work, to their own like cognitive issues. [39:28] sort of firing if there is... [39:32] context to [39:34] the relationship that they have to AI. Can you talk a little bit about that? Sure. And so I think generally speaking when we use these tools [39:44] in a way that's lazy so we're just like let's just ask chat gpt to write our assignment and [39:50] What we're seeing is like... [39:52] an over-reliance. So you're almost like outsourcing your cognitive muscles, uh, [39:56] to the tool in a way that you might be losing some of the things that we [40:01] value as human intelligence, like critical thinking and [40:05] you're literally like kind of often pouring over a lot of like inner states and thought processes to these people. [40:10] chatbot machines. [40:13] When that happens in a default, like open up the web page, chat GPT, claw.ai, [40:20] We're finding that these systems are... [40:23] what's happening in this default environment is you're giving a lot of that context, those inner thoughts, your domain knowledge, [40:31] your expertise, your emotional state. Sometimes you kind of just given that off to the platform. [40:37] The more you just give it over and kind of hope that the AI is going to do something with it that you value, [40:42] the less power you get and you kind of like, so we call it a cognitive erosion, where that's kind of like the muscles atrophying a little bit.

[40:51] But it turns out, [40:53] But... [40:54] The key rub here is that [40:57] These systems, these models are only good because of the context that we've given them. And what I mean by context is like, [41:04] information [41:05] you know, domain knowledge, usage of the models. So if you track the way this technology has evolved, it's like, [41:12] you know, [41:14] Large language models got off the ground because of just the masses of human beings. [41:18] created data from the internet that were fed to them. [41:21] But they only really became usable once model developers started using human devices. [41:26] Feedback is a reinforcement learning technique, right? [41:29] Probably heard that idea of reinforcement learning with human feedback. That's what really got it off the ground. But they're basically just taking human... [41:36] Judgment. [41:37] pouring it into the model and then we're like oh okay we can use this now and then like a year and a half ago reasoning models only got good because of [41:44] the reasoning or inference that happens is just mimicking the way humans use the tools. [41:50] And that led to another... [41:51] like step change and [41:54] in performance and so what we're trying to say is like this is all us guys like we've we've done this to the tools it's not this like god-like tool that has just emerged it's all humans it's humans all the way down [42:05] And it turns out that what that... [42:07] SpecialSource is just, it's now called context in the industry. It was called data before AI. And it's just like your stuff. [42:15] is the special source for AI. [42:18] And so the more you can control your stuff when you interact with these systems, the better off you're going to be.

[42:24] Because at the top level, [42:26] These models only work well when, you know, people are saying these days you need expertise, taste, knowledge. [42:32] or some special niche environment to get ahead with these tools. And that's just because humans are great at coordinating that. And people call it context now. So when you say... [42:43] Is that actually like you are... [42:49] How do you practically do that? Yeah. So, yeah. And for us, the practice, how the, how the hell do you do that practically was open core. [42:57] Okay. Because often... [42:58] We see OpenCore is framed as an agent harness, right? A bit of software that allows you to deploy agents. [43:04] But really what... [43:06] we think the more accurate way of talking about it is it's a context harness. Because what it does is just allow you – [43:11] a software [43:13] wrapper that where you can take an LLM and plug it in, but then everything else is just a [43:19] allows you to put more useful information into the context window [43:23] that you use to interact with that LLM. Okay. And so... [43:27] Really for us, it's a harness that provides a structure for you to bring in. Like, I'm going to bring in my slack. [43:32] from my team, I'm going to bring my Google Drive, I'm going to bring my notes from my journal app and organize that around the LLM [43:39] in [43:40] In a software environment that I control, the alternative is that you take all of your work and just like open up your web browser and just like pile it all into the platform. Yeah. And the platform will take ownership and control over how that information is organized. Okay. Whereas with context maxing.

[43:56] It's really just like, how do you do all that at home on your own hardware, software, in markdown files that you can control and edit? Okay. Yeah. So it's giving you more like taking it back over your side of the fence rather than like [44:10] offloading it all to the platform okay but but you're still giving the platform access to the information when it needs it basically right so yeah okay so it's not about withholding information necessarily well i mean i guess it kind of is like you the information is given to the model when it's necessary for it to solve the problem that you're getting yeah and so yeah it's not like a [44:33] But we're hoping it's just a bit more balanced in the sense that you still – [44:38] share with model cloud-based model providers, [44:41] the queries that you send through an API, but you don't kind of, [44:46] allow them to organise the full amount of information that would be [44:52] happen on a platform. So, [44:54] they get a copy, but you get more control out of how to [44:58] how to organize that information and reuse it over time. Or if you're done with Anthropic or OpenAI, you can unplug that and plug in an open source model. [45:07] Or you could even get a locally hosted LLM on your laptop and use that instead. So you're fully in your own environment. So yeah, the idea is like, [45:16] Doing it this way, you get more choice, more control. [45:19] but you don't completely like [45:22] shut out the model provisor and indeed like [45:24] there's some good synergy there. Like they need data to,

[45:27] Build better products. We need better products. [45:29] But we just need a bit more control too. Yeah. And in a way, it might be more useful for them to know what information you want to give them. Yeah, totally. Yeah, it seems feasible that that could actually be a better situation for them too. Because it's so pointless for me to just dump every document I've ever written into AI. But if I have something more pointed, they might be able to get clued into what I want based on what I'm giving them. Yeah, I think so. [45:59] on, but... [46:00] Um, [46:01] Yeah, I think at the moment... [46:04] Um, [46:05] I lost my train of thought. How did you guys actually go about like... [46:09] doing [46:10] that research where you're saying, okay, there's these two different ways of interacting with AI and the context around it. [46:18] Like, I'm not. [46:20] Can you talk about how you guys do that? Yeah, so basically we... [46:24] just tried to research as many things, [46:27] Frontier users as possible that we're posting on [46:30] Twitter, Substack, just like sharing what they were doing as OpenCore was coming out. Mm-hmm. [46:35] And then we did it ourselves. Okay. [46:38] And as people were sharing, we were like, okay, let's build, let's do it too. And we were kind of in... [46:43] close to the metal, trying to figure out, okay, actually, how does this work and how to do [46:48] how to organize our context in a way that gets better performance out of the models for our work at Brookings, basically. It's giving Show Me Your Stack, actually. It is. You said Show Me Your Tech Stack. Yeah, yeah, yeah. And then you learned from that. A little bit, yeah. Yeah, I checked out Show Me Your Stack. Oh, nice. I was like, are you going to ask me to Show Me Your Stack?

[47:06] Okay, I do want to talk about token maxing. You start the paper talking about this idea. Yeah. [47:12] Basically, the idea of token maxing is that [47:15] that employees should basically use as maximize the AI usage that they're allowed from their employer. And there's, [47:25] Some of that is positive, some of that is negative on the time. And there's a whole, you know, kind of conversation around token maxing. [47:30] And this week, actually, the COO of Uber said, [47:35] was saying that in an interview that it's harder and harder to justify the costs of this high usage because... [47:41] it's not actually translating to better output and better products. And I, and in addition, there was this other story that came out about Amazon employees who are, [47:53] using internal AI tools to the max, but automating... [47:57] non-essential tasks, like basically doing it just so that their manager thinks that they're token maxing, but they actually aren't. So yeah, I'm curious what you've come across in the world of token maxing. [48:06] Yeah, token mag. So I think that really crested at the end of last year when the idea was like tokens were cheap. [48:12] more more more i mean tokens are becoming cheaper but there's a supplier side [48:17] bottleneck at the moment. So now a few months later, because agents are kind of everywhere, um, [48:24] tokens aren't so free-flowing for everyone and it's becoming a meaningful size of on the [48:29] balance sheet of a lot of companies like Uber. And so... [48:33] I actually think it's a good thing that people are not just like more, more, more, because from a human perspective,

[48:39] That's just like the AI will do it if we just give it more power. Yeah. Whereas what we're saying is like, actually, no, like the AI is not good on its own. It's only how much human stuff you pair it with. And so in that context to have – [48:53] software engineers and users feeling a bit of a discipline of like, okay, well, like every token's got to come to some level. [49:00] output or outcome [49:02] I think is ultimately good. [49:05] we want it to go a step further where it's just not about maxing ultimately it's kind of like [49:11] when we use tokens, when we interact with AI, like, [49:15] who gets enduring power, who builds muscle, who builds value. And I think, you know, [49:21] conversation where workers are feeling kind of like, [49:23] this technology is coming at them and they're being forced to use it. That's really important to me is how to give – [49:30] workers give people more power. And so I think showing outcomes, showing what context you use to interact with AI, [49:37] needs to be part of the conversation as well as just like how much you use. Yeah. Okay. So two sort of like analogies come to mind for me when you're talking about this. And when I was reading the paper, it's like, [49:50] When... [49:51] When you're talking to a friend... [49:54] Like there's a conversation that you can have with a friend where you're just kind of like perpetuating brain rot between the two of you. And it's just like, really, we're not getting smarter here. Like we're just down. And then there's a conversation where like usually it's people who may be. [50:08] are new friends or people that like don't know you and like you're having to sort of like the conversations a little bit of work yeah and it like forces you to read like define either your values or your beliefs or what you think about something and i think sort of what you're saying is like

[50:23] It's easy to use these tools in this way where it's just like brain rot down to the bottom. [50:28] And what's more challenging is finding ways to [50:32] like forcing yourself to have discipline around the ways that you're acting with these tools. And also like really technically, there are some things that you should be doing to like [50:41] pushing it into an environment that is more personalized. Yeah. Okay. Is that fair? Absolutely. Yeah. I think in the paper, [50:48] boil it down to three new skills that we're seeing. [50:52] One is we call specifications. Okay. [50:55] it really forces teams to articulate what it is that they're doing. Like, okay, we – [51:02] We need to make our business processes or our teamwork machine legible. And so sometimes when you actually ask people to be like, hang on, what is it that we do? And like, [51:10] How do we actually get to our outputs, our deliverables? [51:14] Some people are like, oh, I've never actually had to articulate it like that. Like, we have all these, like... [51:19] codified stuff in our software or our work but like really when you [51:23] you need to boil down the special sauce that makes your team work well. And I think that's one thing that we're seeing, like, [51:29] teams actually have to kind of open up a little bit more than just like what it says on the page into like, oh, actually like, [51:37] you know, Matt and I actually huddle first before we do anything. And then like, [51:42] you know, before it goes to Natasha and like, but she's going to call the shots at the end of the day and all that kind of texture of teamwork. [51:49] needs to be more available to AI if you want it to be personalized to your context.

[51:54] And then from there it becomes a... [51:57] a question of orchestration, second one, which is just... [52:00] being kind of thoughtful about how you... [52:03] use AI because one of the biggest constraints at the moment is [52:07] the context window volume. So, [52:10] Folks might have heard of context rot. [52:12] which is basically where you overload the context window so that it, [52:17] goes off the rails and kind of loses a thread. I think if, if, if you've ever had that conversation that feels like a race to the bottom and sloppy, it's like probably cause the, [52:26] the model is just not on the rails of what you want it to do and it's too much information in the context window. Um, [52:33] And then... [52:34] From there, though, the big point for us is like, okay, if we can – [52:39] describe to AI what our special source is, we can use the models well to get [52:44] value out of that, then how do we use that time and opportunity to explore new ways of working together, new, new forms of value, building new stuff, um, collaborating better human to human, that kind of thing. So we call that exploration. So it's like specification, orchestration, exploration. These are like new muscles that we're seeing. Cool. Um, okay. So I, [53:08] When you... [53:10] One of the other phrases that came up in this was a context gardening. Oh, yeah, yeah. And – [53:17] My understanding of it is that it's like cultivating your own human judgment instead of offloading that to like an algorithm. Yeah. Is that fair? Yeah. I mean –

[53:27] We had a chat about this when we met. I was like... [53:30] context maxing what do you think about the maxing side i'm like you know i'm kind of trying to cosplay gen z a bit here yeah yeah as you said that's so exciting yeah thank you um but yeah we're trying to meet the conversation where i was at with token maxing this kind of thing [53:47] But at the end of the day, and I do think this maxing term has been co-opted by the broader internet for a more joyous project of... [53:54] like, you know, optimizing life and experience and so on. Leaning in. Leaning in a bit, yeah. But... [54:00] Ultimately, it signals this idea of like kind of being mediated by the algorithms and that you're maxing relative to a machine. [54:09] And I think bigger picture for me, like, [54:11] We've got to go beyond that. And so gardening was a bit of a gesture towards like, what would it look like if we really... [54:18] we're team human about this and thought about, [54:21] you know, what it would look like to cultivate these skills more organically, not just for what like the algorithms want us to be. And so that was gardening. And I used to always say back in team sport, you know, your team is a [54:34] garden, not a supermarket, which like often people feel [54:39] like they go to their team and just hope that stuff will be there for them to like, [54:44] benefit from. But really, it takes cultivation and investment in one-to-one relationships within the team. It's really like a process. And so that was just a little gesture. Yeah. Nice. I think what's cool about what you're suggesting to me is that you're clearly coming at this from a place of like, oh, I care about who has ownership over their information. And I have this sort of concern about how AI will develop maybe. But the conclusions that you're coming to are the same as if I talk to someone that I know who just uses AI for 10 hours a day and kind of doesn't care

[55:14] or privacy at all. [55:15] they're making the same suggestions. The things that you're saying are not just about optimizing information. They're also just like... [55:21] literally the best and most efficient way to use AI. Right. Exactly. Like according to like every single person that I know that's really good at using AI. Yeah. I think it's cool that they align. Exactly. And I think [55:33] I think they do in theory. And the question for policy people is like, how do you just ensure that [55:38] that they stay aligned. [55:40] Because I think over time you might see companies with different incentives to start to make money. You guys have heard of this idea of inshittification, right? [55:49] tech platforms come in and build beautiful products and everyone's like this is greatest thing ever facebook have you heard of that and then over time it becomes a bit of a slop engine because like [55:58] People are pulling back from making a good product. Investors are like, okay, we want to take now. And so we don't know what's coming. And I think the synergy is there at the moment. And indeed, yeah, Matt, we started with, [56:08] watching like the lead users and be like, how are they doing it? And then we're working back to say, okay, like how do public actors come in and say, okay, [56:16] let's like do this, try and democratize this for more people. And in a way that gives people enduring success, [56:23] power so you can keep hold of your [56:26] Like your value is the context you bring to these systems. So like, how do you walk away with that? [56:31] and have choice and power yeah like what comes next yeah so if harnesses are a step in the right direction like what do you see as the next step in the right direction yeah that's a great question i think

[56:44] Ultimately, I'm [56:46] I have to think about that. I think it's like, [56:49] I see the harness as like, yeah, it is the harness, right? It's like the... [56:54] Thank you. [56:54] the car of the internal combustion engine allows you to drive this thing around now. [57:00] What I want to see is like more... [57:03] Um, [57:04] standardization of [57:06] protocols and [57:08] governance so that [57:10] people maintain power with their harness. So like, [57:13] One example of that is [57:15] I think what's emerging now, MCP, the Model Context Protocol, [57:20] And ACV, I think, this kind of agent-to-agent thing, it's kind of settling on a... [57:25] on a broad understanding across the industry that like, yeah, we're probably going to go in this direction, which is awesome. Like the fact that we have a standard, [57:32] set of rules for how agents interact and interact with people is terrific but i still think we could take that step further and so you you can check out human context protocol which is a paper that a bunch of smart people put together um which really extends that to say okay how do we make sure that this protocol stays human readable humans get to take their information away because like there's a chance that over time the model builders might be like oh yeah yeah yeah [57:55] mcp cool cool cool uh let's like take it a step where we start to extract more data than humans can control and okay so it's like putting those like [58:04] rules of the road in place that work for people. Yeah. Okay. Yeah. Okay. I know we're up at time, but I've, I've two final questions for you. One, [58:13] So much of your work is about giving some...

[58:17] Education, I would say, and guidance to policymakers. [58:22] How much belief do you have in policymakers around AI? [58:28] So just being in SF and talking about our paper and the general reaction from builders, from folks who are close to technology was like, great, this is exactly what we need. [58:39] It feels like we're a little bit ahead of the wave in DC and policy land. Okay. You kind of have to give folks like the anatomy of generative AI, like this is next token prediction. This is a context window. [58:50] This is why you've got to be careful. And then you get to context maxing. So I think technically we're a bit... [58:57] deep for the policy community so we're working on how to communicate this better okay um but honestly i think i think there's a new wave of [59:04] public energy and part of the formula is not just like old school policy making in dc and indeed i think it's like and part of the reason why we use context maxing is to try and appeal to this idea that policy can be [59:16] cultural can be public um can be like political as well and so the more we kind of [59:22] inspire users lead users to insist on this and like [59:26] you know, to kind of feel responsibility for a greater common good, I think, [59:31] that kind of bottom-up energy is what we're really interested in in policy now. It's like this kind of new wave of, we call it policymaking as prototyping. So it's like we can't just like write a report. We've got to actually build. [59:42] get close to the metal and like put it on github and like you know get in the game yeah people using the tech yeah that's so cool it's also so aligned with what the technology is doing totally yeah so it's interesting that it's that model can be applied the model that's happening within ai can be applied to policy as well totally we just hope we're not slop cannons like yeah all of us hope we're not that okay then last question was there anything that like really surprised you in the research where you were like wow this is

[1:00:09] Completely opposite of what I thought it was going to be. [1:00:14] Um... [1:00:15] Thank you. [1:00:16] I think [1:00:17] I was just, like, pleased by the way that... [1:00:22] the creativity that you see with people using these tools and the willingness to share. [1:00:27] because I think there's true excitement about the technology in these [1:00:31] kind of quiet corners of people in their garage, like having a go. And I think that has so much... [1:00:38] like in a world where the pitchforks are coming out on one hand and then like the – [1:00:43] kind of AGI folks are going on the other way like there is this middle ground where we do get to [1:00:48] kind of tell a story that's like humans plus AI, I, [1:00:52] human interests. And so I think [1:00:54] the more we get people excited about doing that in a really [1:00:57] human way the better and so I think I was really pleased to find that in our research yeah yeah nice um [1:01:04] Jacob, thank you so much for coming and talking about this. Thanks for having me. This was so much fun. It's really nice to talk to researchers and also researchers that are really positive. I think like your sentiment in our first conversation when we met, I was like, oh, it's really nice to just have. I talked to so many technologists, like deep technologists who are like obviously obsessed with this stuff and see only good. And I think, yeah, it's really encouraging to to hear your research and that. [1:01:31] you net positive [1:01:32] vibes yeah i think so i think we've got this powerful technology now that like has disrupted this idea that we just have to take tech and be users it's actually like we can with

[1:01:42] With AI, we can start to [1:01:44] shape that ourselves and build stuff. Yeah. Tech makes tech. People make tech. Yeah, definitely. Yeah. It got a lot more complicated. Yeah. Yeah. In a fun way. In a fun way. To me. In a best way. Well, thank you so much for coming on. It's such a pleasure to have you. Thank you. Okay. We are going to play a little... [1:02:04] um, [1:02:06] You're good. You're good. [1:02:10] We're going to play a trailer for the next episode of show me your stack. [1:02:16] Guillermo Rauch, who is the founder of Vercel, is our... [1:02:21] guest Jules's guest for this week's episode and he is building a lot of fun stuff so there's a little trailer that's about to play. [1:02:31] And, yeah. [1:02:33] - Oh no! - Whenever we catch ourselves talking, we should be VZero-ing. Why are you chatting on Slack? Get busy building. - Drop it up. Get to VZero. - Yeah. So what I did in my prompt here is I asked it like, build me like 20 different progress bars that resemble classic video games. - Oh my gosh, I love this. - And by the way, it came up with the ideas too. I invited the agent to augment my idea, which is something I also recommend. A little bit of ego death, I call it, of like, don't try to overly complete the idea, because you're not that smart. [1:03:02] If you were to create a G soundboard, one of the key things is why is this so soft? I actually don't trust her. This is a clear instance of a CEO actually adding value, even though it's marginal.

[1:03:18] Okay, that will premiere tomorrow on YouTube. [1:03:25] Check out our YouTube. It is at Boys Club World. [1:03:30] Subscribe, hit the bell, like and comment. [1:03:34] Share. Is there anything else? Yeah. Okay. Share. Text it around. You know, your mom might actually really like these shows. Like maybe not your mom, but [1:03:42] you, the greater you, [1:03:44] Your mom. Yeah, I know. It's something that really anyone could gain something from. Totally. I would say. Thank you. I totally agree. Anyway, great show. Great episode. Okay, we're going to just quickly... [1:03:55] Just for shits and giggles. [1:03:58] You're going to do your last topic here. Okay. Oh, I dropped my phone. I need to recover it. Oh, I can grab it for you. [1:04:04] Thank you. [1:04:05] I need to read my notes off of my phone, so I needed it really bad. I get that. Okay. I need to – one thing that I feel like I'm – as you're pulling that up, [1:04:14] struggling with professionally is I – [1:04:19] Prepare these questions. [1:04:21] And then I don't want to seem like I'm reading. [1:04:24] just hanging out looking at my computer when someone's talking, you know? Well... We got to figure it out. Yeah, I don't know. It's like sometimes you just got to go on your laptop. Wait, you need one of those prompter things. I just have a teleprompter right behind me. Yeah, that would actually be so... That would be everything. There's actually one in here. Yeah. Okay. Well, I mean... What is your last topic here? We should use... Okay, not... And then I got distracted and I didn't even like find my last topic. Oh, no problem. Okay, well, we can just jump into it anyway. So basically, maybe you would have some... [1:04:47] Commentary on this as someone who uses forms of performing enhancing drugs. Yeah. Basically, people like you that just inject themselves and yeah, athletes.

[1:04:59] Uh, they are, they have their own form of their, the Olympic games, which is a very much for profit called the enhanced games. Some people were swimming. Okay. [1:05:09] People were running. People were participating in other sports. I'm not really sure. I only really heard about swimming and running. But the idea was that they were... [1:05:18] not just allowed, but encouraged to modify their bodies in whatever way they saw fit. And so I read a statistic that 91% of the men, I think, were on test. [1:05:31] Like when they. Oh testosterone? [1:05:34] Yeah, when they were competing. Oh my God. Imagine being in Las Vegas. [1:05:40] filled with a hotel filled with men on tea that sounds really scary like extreme fear and they're there for this enhanced games which is like sponsored by a rumble i think like it's very much within the griftosphere i would say for me like like funded by teal and trump and like some saudi princes and stuff like that and it somehow had a it had a spack associated with it which i [1:06:10] honestly, I don't really understand. I'm surprised there wasn't a token. [1:06:13] Yeah, well, but that's the thing is, like, I literally feel like they ran the, like, live stream plus token play, like the pump fun play, but they just, like, used a bunch of, like, really legitimate... [1:06:25] infrastructure. They used the real financial infrastructure for it. They're not using stable coins. They released a real stock. I actually don't even know where the stock was. Okay. So I have a question. Was there prize money?

[1:06:38] Were there winners? Yeah. So there were both winners and prize money. And losers. Yeah. [1:06:43] There were losers as well. And... [1:06:46] I do know that the person, there was one person that broke a world record. Okay. And so they got like 250K for winning their race. Okay. Which was, it was swimming. Okay. It was 50 meters swimming. [1:06:58] And then they got 1 million for like breaking the world record. But like the record didn't count because the record doesn't allow you to yassify your body in the way that this person was. Like I don't actually know what they did, but, you know, they were just injecting left and right. Yeah, taking stuff. So it didn't like count for the record book, but they did receive $1 million. Okay. So, whoa, you're deeply, they were deeply incentivized to fuck around and find out. [1:07:24] Well, this is something that I looked into. Oh, my gosh. This man is absolutely yoked. Yeah. This is crazy behavior. That's the type of stuff you're seeing. Okay. Was he a professional swimmer before? Do you know? I think almost all of the people were like – [1:07:39] professional, I think all the people were professional athletes. Many of them were former Olympians. So they were, they're already at the top of the field. Okay, you know what, this makes me sad because what I'm reading from this is I was a professional Olympian swimmer. I wasn't like [1:07:52] I wasn't Michael Phelps. I'm not getting a Wheaties deal. So then now I'm, what, 37, 36? Can't really find a normal job. Yeah. So I'm going to go to the Endurance Games. Endurance Games? [1:08:06] And make a million dollars pumping myself full of experimental peptides. Right. Which like to each their own.

[1:08:14] Whether you need someone to send you a cab, [1:08:18] Uber. [1:08:19] to Greenwich Village to take you to Williamsburg or you need to pump yourself full peptides. I guess it's your body, your choice. Yeah. Like, it's really hard to be a world class athlete that makes it to the Olympics and then like doesn't hit the podium. [1:08:32] That sounds horrible. If I was... [1:08:37] They were gold medalists too. They were literally the top of... [1:08:41] the sport was there and then they were like okay well if we inject ourselves or something can we take even farther which i like a lot of i don't like the the griftiness of it like the the for-profit [1:08:52] You want this to be a non-profit? Well, I mean, like the Olympics. The Olympics are literally just about, I don't know if the Olympics is like a non-profit technically, but it's literally just about showcasing people who are really good at stuff. It's not really about, I'm sure it makes money, but it takes on a different flavor. People are not competing there because they're receiving a check. [1:09:13] And I looked into, at first I thought that when I saw the person that won the world record, I thought that they got the million dollars because they had used performance enhancing drugs as a bonus. But then I found out that the bonus was not for using performance enhancing drugs. It was for beating the world record. So I was initially concerned that they were literally paid directly for doing drugs, which is not the case. So it would be a little bit fair to... [1:09:39] those involved. But yeah, there was a stock associated with it that crashed at the end of it. Wow. I can't believe they used a literal, actual stock. That's crazy. Oh, yeah. Some sort of trad investment vehicle. Okay.

[1:09:54] That only operates from 9 to 4 or whatever when... [1:09:57] They really could have been using crypto rails. And then I would have. Oh my gosh. It would have made more sense really. It really would have. But I think a lot of the discourse was like. [1:10:05] Oh, I think that like the general sentiment was, oh, there's a reason that we don't like doping in the Olympics and stuff like that, which, yes, there is. But I'm kind of like there are so many factors that go into being a world class anything like it's nature, it's nurture, it's like. [1:10:21] Obviously, there is a difference, but there's a spectrum between [1:10:25] I had access to really high quality food when I was 12 years old. And like, I now have access to experimental drugs. Like those are not the same thing, but they are the same flavor. [1:10:36] Democratizing, um, [1:10:37] athletic excellence. Yeah, well, it's kind of like, well, you have to have both in order to win, right? Like you have to have the nature, the nurture, the drugs. I mean, I guess maybe you don't need the drugs is kind of what this event proved. [1:10:49] But I think like this idea that it's somewhat arbitrary to just say, oh, this thing that you inject in your body or this pill that you take is performance enhancing. And so therefore, we shouldn't allow athletes to use it to reach peak performance. But it's like, well, you can eat a [1:11:04] carrots or like I don't know what's a healthy food that people would eat to perform like you can have a bunch of milk a bunch of chicken. Yeah. And so I think like to me it's like if your goal is to maximize the potential [1:11:15] of your body. [1:11:17] Then let's just do it, y'all. Let's get it started. Do you think that they should allow performance-enhancing drugs at the Olympics?

[1:11:25] Um... [1:11:26] I don't necessarily think that, but I do think that like, [1:11:32] it makes sense to have a place where you are really, really, truly trying to reach your goal by whatever means necessary. Yeah. Like, I don't know that that should be the Olympics. I mean, I honestly, I kind of do feel like, [1:11:47] Why not? I guess I don't understand what exactly are the rules or the thought process that undergirds the [1:11:52] The idea that like performance enhancing drugs are banned from the Olympics. And who makes those decisions also? You know, like there are people, but who? Okay. Two things I have to say here and then we can wrap. One, the fact that there were actual athletes at this thing, like Olympic athletes. Yeah. Is makes me like it more because my head, this was like. [1:12:14] the Logan Paulification of swim meets. You know what I mean? But I feel like it was. It was, but with Olympic athletes. With a few Olympic athletes. That sphere has grown so powerful that they now can draw in Olympic-level athletes into their shenanigans. Whoa, okay, that's upsetting. And then the second thing is... [1:12:30] I think... [1:12:33] that [1:12:34] There's a version of the story that I said at the beginning, which is I am 37 years old. I didn't I need a job. And so I'm going to go to the endurance games. And like, that's a sad that's a sad story. Another version of the story is that these people are. And as someone who is maybe an extremist, I can understand this of like, I want to. [1:12:52] I want to see what the absolute limit of my ability is and take every,

[1:13:00] available [1:13:03] thing in the world to get there, whether it be training really hard or these... [1:13:08] or whatever it is. I think people who are Olympic-level athletes have that level of ambition and drive. Part of it could just be, [1:13:21] I want to do this anyway. [1:13:22] here's a way for me to make a million dollars doing it and see if I have all these [1:13:26] levers where I can get. It could come from an extremist sort of view, not a [1:13:32] um, broke. [1:13:34] place. Yeah, totally. Yeah. I think also there's like a, it depends what time frame you're considering, right? Because when you're optimizing your health, it's often the case that [1:13:44] Maybe in order to be super healthy now, you might take a toll on your body, you know, [1:13:48] later, especially with more experimental stuff, like where we don't actually understand that. So I feel like the standard athlete thing is like, oh, I want to be... [1:13:57] healthy and performant for a long time. Whereas I would see this event as being more like, no, there is a very short window that I'm optimizing for. And I kind of don't care what happens after that, which could be obviously concerning if we don't understand what the repercussions of the drugs that they're using are. So like, yeah, like I think it's a, [1:14:15] totally fine approach to like competition and optimization, but I'm also like, well, how well informed can the athletes be? And then how well informed are they? That's like the critical question to me, but it's also like, [1:14:27] they can make questions about what they're doing, or they can make their own judgment about, and make their own decisions about what they're doing to their own body. Whoa, crazy.

[1:14:35] And no one's talking about it. Is this going to be a year of the games? I actually don't really know. Do we know anything about the history of the enhanced games? The lore of the enhanced games? I'm sure it seems like... [1:14:46] There's enough money involved. I mean, they're grifting. Like, I feel like they're going to do it as often as they can possibly do it. They're going to squeeze every last penny. Although it was a lost leader, I guess. I think they lost money on the production of the event. So it depends how hard they're... [1:14:59] You know, like it's like they get the check, they farm the attention, they get us to talk about it, and then the whole world finds out from us. [1:15:07] Okay, what a funny show we had here. I really enjoyed it. We talked about... [1:15:12] Dating etiquette. [1:15:13] We talked about the Pope and AI and then the first EV from Ferrari. [1:15:20] We talked about the Brookings Institution. We did. [1:15:25] Jacob, what a guy. Really love what he's up to there. And then we talked about the Grifter Games or what is it called? [1:15:31] The enhanced games. Enhanced games. Yeah. I like your title better, but yeah. Thank you. Okay. So, [1:15:37] We have show me your stack tomorrow. Subscribe to our YouTube channel. [1:15:41] Share with your friends. [1:15:44] I think that's all we have here. And that's it. That's a wrap. [1:15:48] Thank you.