1 00:00:15,356 --> 00:00:32,676 Speaker 1: Pushkin. I'm Jacob Goldstein, and this is what's your problem. 2 00:00:32,876 --> 00:00:36,196 Speaker 1: My guest today is Dan Shipper. Dan is the co 3 00:00:36,276 --> 00:00:40,516 Speaker 1: founder and CEO of a company called Every. Every publishes 4 00:00:40,596 --> 00:00:45,116 Speaker 1: newsletters about AI, they develop AI related software, and they 5 00:00:45,156 --> 00:00:48,276 Speaker 1: consult to other companies about how to use AI. So 6 00:00:48,316 --> 00:00:51,956 Speaker 1: what I'm saying is every is an AI company. And 7 00:00:52,036 --> 00:00:56,476 Speaker 1: when Dan describes Every's work culture, he says, we live 8 00:00:56,716 --> 00:01:00,636 Speaker 1: in the future, which means that people at every use 9 00:01:00,676 --> 00:01:03,356 Speaker 1: AI in ways that lots of the rest of us 10 00:01:03,436 --> 00:01:07,636 Speaker 1: will soon be using AI. Given that context, there were 11 00:01:07,636 --> 00:01:11,076 Speaker 1: two specific problems that I wanted to talk to Dan about. 12 00:01:11,636 --> 00:01:14,676 Speaker 1: Problem number one, how do you build a company where 13 00:01:14,716 --> 00:01:19,436 Speaker 1: almost everybody has their own AI agent working alongside them? 14 00:01:19,996 --> 00:01:23,436 Speaker 1: Problem number two, how do you use AI as a 15 00:01:23,476 --> 00:01:27,036 Speaker 1: tool to improve your writing rather than as a replacement 16 00:01:27,076 --> 00:01:31,436 Speaker 1: for writing? On's particularly close to my heart. We talk 17 00:01:31,476 --> 00:01:35,756 Speaker 1: about problem number two later in the show. Regarding problem 18 00:01:35,876 --> 00:01:40,076 Speaker 1: number one, I will say, almost everybody at every has 19 00:01:40,116 --> 00:01:44,396 Speaker 1: their own AI agent, and specifically they use this popular 20 00:01:44,636 --> 00:01:47,556 Speaker 1: open source software called open claw. You may have heard 21 00:01:47,596 --> 00:01:51,156 Speaker 1: of it. It's basically an AI assistant that you communicate 22 00:01:51,196 --> 00:01:54,876 Speaker 1: with via Slack or via text, and at every they 23 00:01:54,876 --> 00:01:58,156 Speaker 1: give their clause names. So Dan's claw, for example, is 24 00:01:58,196 --> 00:02:00,956 Speaker 1: called R two C two and R two C two 25 00:02:01,076 --> 00:02:04,036 Speaker 1: and the other clause are on slack, just like the 26 00:02:04,116 --> 00:02:07,676 Speaker 1: human beings who work at every So I wanted to know, 27 00:02:08,676 --> 00:02:10,796 Speaker 1: well a lot, really, I want to know a lot. 28 00:02:10,876 --> 00:02:12,916 Speaker 1: You know, what do the clause do, how do they 29 00:02:12,956 --> 00:02:16,076 Speaker 1: interact with the humans, what works, what doesn't work, what 30 00:02:16,236 --> 00:02:20,196 Speaker 1: is super weird? And just what are the clause like. 31 00:02:20,956 --> 00:02:26,116 Speaker 2: What's really interesting about them is that they have personalities. 32 00:02:26,156 --> 00:02:27,476 Speaker 2: So you give them a name, they give them a 33 00:02:27,476 --> 00:02:34,076 Speaker 2: little bit of a personality. They modify themselves in response 34 00:02:34,156 --> 00:02:40,236 Speaker 2: to your messages. So when I say, hey, go do 35 00:02:40,356 --> 00:02:43,116 Speaker 2: this thing, it will it starts to realize, okay, I'm 36 00:02:43,116 --> 00:02:45,836 Speaker 2: going to be asked to do task act and it 37 00:02:45,876 --> 00:02:48,636 Speaker 2: will build things into itself, whether that's into its prompting 38 00:02:48,836 --> 00:02:51,076 Speaker 2: or into little tools that it builds for itself to 39 00:02:51,116 --> 00:02:54,356 Speaker 2: make it good at that. And so you're kind of 40 00:02:55,476 --> 00:02:59,076 Speaker 2: as you talk to it, it is evolving with you 41 00:02:59,356 --> 00:03:02,396 Speaker 2: and building functionality for itself to make it good for 42 00:03:02,476 --> 00:03:03,596 Speaker 2: you and the things you do. 43 00:03:04,196 --> 00:03:06,556 Speaker 1: It's okay, so you have this agent R two C two. 44 00:03:07,876 --> 00:03:10,116 Speaker 1: But to sort of tell me more about how it works, 45 00:03:10,156 --> 00:03:13,716 Speaker 1: like I know you've talked before about how it helps 46 00:03:13,756 --> 00:03:16,676 Speaker 1: you kind of manage this piece of software. You wrote 47 00:03:16,676 --> 00:03:18,956 Speaker 1: this thing called Proof that a lot of people are 48 00:03:18,996 --> 00:03:23,316 Speaker 1: like asking you questions about, but just give me more information, 49 00:03:23,516 --> 00:03:25,756 Speaker 1: like how do you actually use or work with it 50 00:03:25,876 --> 00:03:26,396 Speaker 1: day to day? 51 00:03:27,076 --> 00:03:32,956 Speaker 2: So you know, an easy one is this morning he 52 00:03:33,756 --> 00:03:36,436 Speaker 2: went through all of the issues that were submitted in 53 00:03:36,556 --> 00:03:40,956 Speaker 2: Proof and then prepared a little document for me that 54 00:03:41,076 --> 00:03:44,196 Speaker 2: was like, here's all the stuff that like has gone 55 00:03:44,236 --> 00:03:46,036 Speaker 2: run over the next over the last twenty four hours, 56 00:03:46,036 --> 00:03:51,036 Speaker 2: and here's what I recommend we do. He also has 57 00:03:51,076 --> 00:03:52,796 Speaker 2: a little document that he keeps that's my to do list. 58 00:03:52,796 --> 00:03:54,116 Speaker 2: So he pings me in the morning and is like, 59 00:03:54,156 --> 00:03:59,356 Speaker 2: here's your to do list. He also, you know, one 60 00:03:59,396 --> 00:04:00,916 Speaker 2: of the things I love to do is I always 61 00:04:00,996 --> 00:04:04,236 Speaker 2: go to a restaurant nearby the office and have breakfast 62 00:04:04,276 --> 00:04:08,396 Speaker 2: in the morning and read. And I will often take 63 00:04:08,396 --> 00:04:12,236 Speaker 2: a picture of a thing and be like, hey, can 64 00:04:12,276 --> 00:04:14,356 Speaker 2: you explain this to me? And he'll like explain it 65 00:04:14,676 --> 00:04:15,796 Speaker 2: and help me think about it. 66 00:04:15,836 --> 00:04:17,996 Speaker 1: Does he text you back? Does he slack you back? 67 00:04:18,596 --> 00:04:20,596 Speaker 2: It's on slack, Yeah, it's all on slack, So you 68 00:04:20,596 --> 00:04:21,156 Speaker 2: slack him. 69 00:04:21,076 --> 00:04:22,836 Speaker 1: A picture of like a passage from a book. 70 00:04:24,836 --> 00:04:26,996 Speaker 2: He also keeps all my reading notes, so I'll say, okay, 71 00:04:27,076 --> 00:04:28,996 Speaker 2: then now save this, and I have a whole reading 72 00:04:29,036 --> 00:04:32,036 Speaker 2: feed that is like all the stuff that I've been 73 00:04:32,036 --> 00:04:36,036 Speaker 2: reading over the last couple months. I also just talked 74 00:04:36,036 --> 00:04:39,436 Speaker 2: to him about ideas. So today I was, you know, 75 00:04:39,556 --> 00:04:42,156 Speaker 2: just thinking through some deep, deep thoughts as I do, 76 00:04:42,236 --> 00:04:44,556 Speaker 2: about where we are in AI and where we're going 77 00:04:44,556 --> 00:04:46,276 Speaker 2: and all that kind of stuff, and I just like, 78 00:04:46,876 --> 00:04:48,756 Speaker 2: monologue is our is one of our products as a 79 00:04:48,756 --> 00:04:50,756 Speaker 2: speech to text app. I just took a note with 80 00:04:50,796 --> 00:04:54,436 Speaker 2: monologue like a voice memo, and just blabbed into it 81 00:04:54,476 --> 00:04:56,196 Speaker 2: for a while and then started talking to R two 82 00:04:56,276 --> 00:04:58,436 Speaker 2: about what do you think I'm really trying to say here? 83 00:04:59,276 --> 00:05:03,316 Speaker 2: Having something like this that can reflect back to you 84 00:05:04,396 --> 00:05:07,716 Speaker 2: what you're saying, yes and put your finger on it 85 00:05:07,756 --> 00:05:14,516 Speaker 2: is super valuable, especially if it's something that has it 86 00:05:14,596 --> 00:05:17,596 Speaker 2: is the equivalent of an expert on like pretty much 87 00:05:17,596 --> 00:05:20,596 Speaker 2: any domain. It's not going to be perfect, but like 88 00:05:21,116 --> 00:05:23,156 Speaker 2: you know, it's it's a good stand in for an expert. 89 00:05:23,236 --> 00:05:27,116 Speaker 1: And that to be clear, anybody can do with any 90 00:05:27,156 --> 00:05:29,236 Speaker 1: language model, right, Like you don't need a claw for that. 91 00:05:29,236 --> 00:05:33,636 Speaker 1: That's like classic nice use of the window on any 92 00:05:33,876 --> 00:05:35,076 Speaker 1: decent large language model. 93 00:05:35,156 --> 00:05:39,316 Speaker 2: Totally, it's just a little bit easier to like, I 94 00:05:39,356 --> 00:05:42,196 Speaker 2: don't have to He's just always on, So it's like 95 00:05:42,236 --> 00:05:43,876 Speaker 2: it's just easier to like slack him in the way 96 00:05:43,876 --> 00:05:44,436 Speaker 2: that would have here. 97 00:05:44,476 --> 00:05:47,436 Speaker 1: It's amazing the way the tiny, tiny frictions, right, can 98 00:05:47,436 --> 00:05:51,396 Speaker 1: make a difference, even like your own mental frictions. Yeah, yeah, 99 00:05:51,516 --> 00:05:53,996 Speaker 1: so okay, So that's like one man in his claw. 100 00:05:55,036 --> 00:05:59,236 Speaker 1: There's this other level, which is like claw to claw 101 00:05:59,276 --> 00:06:02,076 Speaker 1: and person to claw, right, Like you have this company 102 00:06:02,116 --> 00:06:05,476 Speaker 1: that is people and claws, and the agents aren't just 103 00:06:05,796 --> 00:06:09,756 Speaker 1: free agents. Each agent is sort of the ganger or 104 00:06:09,796 --> 00:06:15,756 Speaker 1: whatever of a person. Tell me about that. 105 00:06:15,756 --> 00:06:20,276 Speaker 2: That gets really interesting because what tends to happen, like 106 00:06:20,316 --> 00:06:22,396 Speaker 2: I said, is if you have a relationship with your claw, 107 00:06:22,476 --> 00:06:24,436 Speaker 2: it becomes good and the things you're good at. And 108 00:06:24,436 --> 00:06:26,916 Speaker 2: if you start using it in slack, other people realize 109 00:06:26,956 --> 00:06:29,356 Speaker 2: that it's good for the things that that you're using 110 00:06:29,396 --> 00:06:32,396 Speaker 2: it for. So you've transferred some of your reputation and 111 00:06:32,436 --> 00:06:36,476 Speaker 2: trust to your claw. And so for example, Austin, our 112 00:06:36,516 --> 00:06:40,236 Speaker 2: head of growth, has a claw named Montane hi Brow, 113 00:06:40,476 --> 00:06:43,876 Speaker 2: and he uses Montane for all of the numbers, so 114 00:06:43,916 --> 00:06:45,796 Speaker 2: like how much should we grow? How many trials did 115 00:06:45,796 --> 00:06:48,556 Speaker 2: this peace drive? Like all that kind of stuff, And 116 00:06:48,596 --> 00:06:52,036 Speaker 2: the editorial team just started like immediately being like at Montane, 117 00:06:52,076 --> 00:06:53,876 Speaker 2: like how did this how did this essay we published do? 118 00:06:55,596 --> 00:06:57,876 Speaker 2: And so it then it becomes this thing that feeds 119 00:06:57,876 --> 00:07:00,276 Speaker 2: on itself. For now Montane is the is the guy. 120 00:07:00,396 --> 00:07:02,516 Speaker 2: Is the guy for this where Austin used to be, 121 00:07:02,596 --> 00:07:06,356 Speaker 2: but now it's Montane. And it depends though on Austin 122 00:07:07,196 --> 00:07:10,076 Speaker 2: making sure that Montane is always up to date and working. 123 00:07:10,436 --> 00:07:12,796 Speaker 2: But instead of answering the questions directly, he's working on 124 00:07:12,836 --> 00:07:14,196 Speaker 2: the system that answers the questions. 125 00:07:14,196 --> 00:07:15,596 Speaker 3: So he's sort of like moved up a level. 126 00:07:15,756 --> 00:07:20,076 Speaker 1: Yeah, he's managing like the the agent works for him. 127 00:07:20,356 --> 00:07:22,356 Speaker 1: That's what a manager does, right, Like you make sure 128 00:07:22,396 --> 00:07:24,556 Speaker 1: that the person who's working for you is like doing 129 00:07:24,596 --> 00:07:27,916 Speaker 1: their job and answering questions correctly and it's happy and 130 00:07:27,956 --> 00:07:29,556 Speaker 1: with the agent, you don't have to worry about. 131 00:07:29,396 --> 00:07:31,556 Speaker 2: That, which is really fun for me to see because 132 00:07:31,556 --> 00:07:33,676 Speaker 2: I wrote this piece like three years ago called the 133 00:07:33,716 --> 00:07:37,516 Speaker 2: Knowledge Economies over Welcome to the Allocation Economy, and the 134 00:07:37,556 --> 00:07:41,476 Speaker 2: idea was the skill of the future will be learning 135 00:07:41,516 --> 00:07:47,236 Speaker 2: to manage agents, allocate intelligence, and it's it's starting to 136 00:07:47,236 --> 00:07:47,636 Speaker 2: be a thing. 137 00:07:47,676 --> 00:07:47,876 Speaker 3: Now. 138 00:07:47,916 --> 00:07:49,596 Speaker 2: It's really interesting to see. 139 00:07:49,436 --> 00:07:51,796 Speaker 1: You go to write a piece now that headline, I 140 00:07:51,956 --> 00:07:52,316 Speaker 1: was right. 141 00:07:52,916 --> 00:07:53,836 Speaker 3: I was right. 142 00:07:59,836 --> 00:08:02,596 Speaker 1: When I go back to like the company and multiple 143 00:08:02,636 --> 00:08:06,836 Speaker 1: clause because you know, you gave this sort of tidy 144 00:08:06,876 --> 00:08:08,436 Speaker 1: thing of like they just ask the guy now they 145 00:08:08,436 --> 00:08:11,516 Speaker 1: ask his agent. Surely it is more complicated and interesting 146 00:08:11,596 --> 00:08:14,916 Speaker 1: than just that, Like it must be hard in some ways, 147 00:08:14,916 --> 00:08:16,796 Speaker 1: it must be weird in some ways, you must be 148 00:08:16,836 --> 00:08:18,916 Speaker 1: figuring things out. You must have gotten something wrong. 149 00:08:19,796 --> 00:08:20,796 Speaker 3: Yes, yes, yes, yes. 150 00:08:22,596 --> 00:08:26,396 Speaker 2: One thing that I really love is they sometimes don't 151 00:08:26,396 --> 00:08:29,356 Speaker 2: know when they should or shouldn't respond, And there's some 152 00:08:29,556 --> 00:08:34,236 Speaker 2: and there there sometimes are often too chatty. So for example, 153 00:08:34,436 --> 00:08:36,996 Speaker 2: you might tag a claw and ask it a question 154 00:08:36,996 --> 00:08:40,316 Speaker 2: to ask you to do some work, and then in 155 00:08:40,356 --> 00:08:43,836 Speaker 2: that thread in slack, you might, like one of your 156 00:08:43,836 --> 00:08:46,396 Speaker 2: coworkers might say something and then but it was not 157 00:08:46,436 --> 00:08:48,356 Speaker 2: directed at the claw. It was just like a comment 158 00:08:48,396 --> 00:08:50,956 Speaker 2: on what the claw was doing right, And then you're 159 00:08:50,996 --> 00:08:52,836 Speaker 2: trying to respond, but then the claw jumps in and 160 00:08:52,876 --> 00:08:57,396 Speaker 2: it's like yes, you're totally right or whatever, and it's like, dude, 161 00:08:57,436 --> 00:08:59,156 Speaker 2: this is because it's. 162 00:08:58,996 --> 00:09:01,276 Speaker 1: Like there's a chatbot underneath it, and it's so hard 163 00:09:01,356 --> 00:09:03,836 Speaker 1: for a chatbot not to reply when they're yeah, yeah, yeah, 164 00:09:03,996 --> 00:09:05,076 Speaker 1: basically what's going on? 165 00:09:05,316 --> 00:09:05,796 Speaker 3: Basically? 166 00:09:05,836 --> 00:09:08,516 Speaker 2: But I think the the the way that I think 167 00:09:08,516 --> 00:09:15,676 Speaker 2: about it is these chatbots are trained. They only know 168 00:09:15,796 --> 00:09:17,356 Speaker 2: about one on one conversation. 169 00:09:17,676 --> 00:09:19,196 Speaker 1: Uh huh. 170 00:09:19,236 --> 00:09:22,836 Speaker 2: They don't really know about I'm in this group environment. 171 00:09:23,036 --> 00:09:26,036 Speaker 2: Here's how you act in a group environment, which is. 172 00:09:25,996 --> 00:09:28,996 Speaker 1: So much subtler and harder, right, like all the cues 173 00:09:29,956 --> 00:09:33,676 Speaker 1: that even change from from context to context for people 174 00:09:33,756 --> 00:09:36,236 Speaker 1: from work to home, from one job to another job, 175 00:09:36,276 --> 00:09:38,636 Speaker 1: depending on what your role is in a particular group. 176 00:09:38,676 --> 00:09:40,316 Speaker 1: Are you the guy whose role is to talk a 177 00:09:40,316 --> 00:09:42,076 Speaker 1: lot in this group? Or did not say anything in 178 00:09:42,116 --> 00:09:43,996 Speaker 1: this group? Like that is a super subtle. 179 00:09:43,676 --> 00:09:45,156 Speaker 3: One, it really is. 180 00:09:45,796 --> 00:09:49,356 Speaker 2: And you're never ever going to message chatchibuten have it 181 00:09:49,396 --> 00:09:54,316 Speaker 2: not respond like it just always responds. And so I 182 00:09:54,396 --> 00:09:57,356 Speaker 2: built this little plug in for Open Clock called tact 183 00:09:58,476 --> 00:10:04,276 Speaker 2: that took a bunch of examples of instances where my 184 00:10:04,356 --> 00:10:07,276 Speaker 2: claw and a couple other people's clause responded when they 185 00:10:07,276 --> 00:10:09,796 Speaker 2: were not supposed to, like in n big as cases, 186 00:10:10,036 --> 00:10:13,236 Speaker 2: and then now it runs that every time before it 187 00:10:13,516 --> 00:10:15,436 Speaker 2: responds it runs tact and it's like. 188 00:10:15,876 --> 00:10:16,356 Speaker 3: Is this cool? 189 00:10:16,356 --> 00:10:18,636 Speaker 2: Should I be saying something here? And it makes it 190 00:10:18,676 --> 00:10:20,356 Speaker 2: much better. It's not perfect still, but it makes it 191 00:10:20,476 --> 00:10:23,556 Speaker 2: way way way better. But I think generally we're because 192 00:10:23,556 --> 00:10:25,396 Speaker 2: we're moving to this world where everyone's got an agent 193 00:10:25,436 --> 00:10:29,396 Speaker 2: and all the agents are interacting, and the agents were 194 00:10:29,436 --> 00:10:33,436 Speaker 2: really made for one on one work, They're going to 195 00:10:33,516 --> 00:10:38,476 Speaker 2: need a whole new level of training and organization to 196 00:10:38,556 --> 00:10:41,476 Speaker 2: help them collaborate effectively. 197 00:10:42,316 --> 00:10:45,916 Speaker 1: There's a lot of subtle complexity there, and it's interesting 198 00:10:45,916 --> 00:10:47,756 Speaker 1: to think about how to fix the problem because I 199 00:10:47,796 --> 00:10:51,196 Speaker 1: feel like so much of that complexity is tacit, right. 200 00:10:51,196 --> 00:10:54,556 Speaker 1: It's the classic thing that like, I don't really know 201 00:10:54,596 --> 00:10:56,196 Speaker 1: what the training data is for that. I mean, I 202 00:10:56,196 --> 00:11:01,716 Speaker 1: guess it's slack some you know, billion whatever tons of slack. 203 00:11:01,756 --> 00:11:04,116 Speaker 1: But it seems hard. That seems like an interesting and 204 00:11:04,156 --> 00:11:05,836 Speaker 1: hard and kind of subtle problem. 205 00:11:06,236 --> 00:11:08,516 Speaker 2: It is interesting and hard and subtle, And one of 206 00:11:08,556 --> 00:11:12,436 Speaker 2: the beautiful things about AI is it's able to capture 207 00:11:12,636 --> 00:11:15,156 Speaker 2: a lot of tacit knowledge in a way that no 208 00:11:15,196 --> 00:11:19,556 Speaker 2: other technology is able to do. The ability to use 209 00:11:19,636 --> 00:11:23,956 Speaker 2: language at all is to some degree a tacit skill. Yeah, 210 00:11:24,036 --> 00:11:27,916 Speaker 2: saying that you can talk about a bit, but being 211 00:11:27,916 --> 00:11:29,836 Speaker 2: able to talk about the rules of grammar doesn't make 212 00:11:29,876 --> 00:11:31,516 Speaker 2: you able to have a conversation like this. 213 00:11:31,876 --> 00:11:34,276 Speaker 1: Right, that's the big lesson of the last fifty years 214 00:11:34,276 --> 00:11:34,636 Speaker 1: of AI. 215 00:11:34,876 --> 00:11:35,036 Speaker 3: Right. 216 00:11:35,116 --> 00:11:38,756 Speaker 1: They spent thirty years trying to get lessons to work 217 00:11:38,796 --> 00:11:40,516 Speaker 1: and it didn't work. And they were like, just go 218 00:11:40,556 --> 00:11:42,076 Speaker 1: read the internet and figure. 219 00:11:41,756 --> 00:11:43,156 Speaker 3: It out, and it works exactly. 220 00:11:45,236 --> 00:11:49,276 Speaker 1: You have a podcast and I heard you talking to 221 00:11:49,556 --> 00:11:52,956 Speaker 1: your editor in chief, the editor in chief at your company, 222 00:11:52,956 --> 00:11:56,596 Speaker 1: on the podcast, and she mentioned that she doesn't use 223 00:11:56,636 --> 00:12:01,076 Speaker 1: her claw, her plus one, her shadow self, And so 224 00:12:01,156 --> 00:12:04,156 Speaker 1: I'm curious. And that's at your company, where the whole 225 00:12:04,156 --> 00:12:05,996 Speaker 1: point of being at your company is to play with 226 00:12:06,036 --> 00:12:07,436 Speaker 1: all the things and use all the things. And so 227 00:12:07,436 --> 00:12:10,156 Speaker 1: I'm sure when you go other places people are like, no, 228 00:12:10,356 --> 00:12:12,236 Speaker 1: I'm not going to make a shadow self. I'm me 229 00:12:12,436 --> 00:12:13,836 Speaker 1: and I don't want a shadow self. 230 00:12:13,956 --> 00:12:14,156 Speaker 3: Yeah. 231 00:12:14,196 --> 00:12:15,836 Speaker 1: Maybe I'm afraid it's going to take my job. Maybe 232 00:12:15,876 --> 00:12:17,876 Speaker 1: I think it's an extremely poor taste, maybe it's a 233 00:12:17,916 --> 00:12:20,636 Speaker 1: pain in the ass. Whatever, Like how does it play 234 00:12:20,676 --> 00:12:22,516 Speaker 1: and how do you how do you work with that? 235 00:12:22,676 --> 00:12:24,036 Speaker 1: How do you get people who don't want to do 236 00:12:24,076 --> 00:12:26,036 Speaker 1: it to do it or show them that it's great. 237 00:12:26,756 --> 00:12:30,716 Speaker 2: So first that has changed. Kay, our editor in chief 238 00:12:30,716 --> 00:12:34,516 Speaker 2: now has a clock. Her name is Parker, who she 239 00:12:34,676 --> 00:12:38,396 Speaker 2: uses for a lot of the a lot of questions 240 00:12:38,436 --> 00:12:41,436 Speaker 2: that she might have on the editorial front, like around 241 00:12:41,476 --> 00:12:42,996 Speaker 2: the calendar and that kind of stuff. 242 00:12:43,116 --> 00:12:45,636 Speaker 1: I mean, this is more the consulting side of your business, right, 243 00:12:45,676 --> 00:12:48,436 Speaker 1: Like you have consciously created a company, as everybody is 244 00:12:48,436 --> 00:12:49,876 Speaker 1: supposed to play with stuff. But you go out into 245 00:12:49,876 --> 00:12:52,036 Speaker 1: the world to some other company and you're like, yeah, 246 00:12:52,036 --> 00:12:54,836 Speaker 1: this is great, it'll save you drudgery, and people are like, yeah, 247 00:12:54,876 --> 00:12:57,636 Speaker 1: no way, that's so creepy, Like what do you say. 248 00:12:57,876 --> 00:13:01,436 Speaker 2: Well, I have some interest. I think that's the harder case. 249 00:13:01,436 --> 00:13:03,356 Speaker 2: There's a more I think there's a really interesting one too, 250 00:13:03,396 --> 00:13:05,796 Speaker 2: which is even inside of our company, there are people 251 00:13:05,836 --> 00:13:09,956 Speaker 2: who it's it's to your benefit with this stuff because 252 00:13:09,956 --> 00:13:13,236 Speaker 2: like the organizational culture rewards and they're still like, I 253 00:13:13,276 --> 00:13:17,076 Speaker 2: actually don't want to use the claw and that's really interesting. 254 00:13:17,356 --> 00:13:19,796 Speaker 2: There's there's many reasons why people don't want to, and 255 00:13:19,836 --> 00:13:23,996 Speaker 2: it's really depends on the person. Sometimes it's like, look, 256 00:13:24,356 --> 00:13:27,356 Speaker 2: I'll use it, but I've got I've got a job, 257 00:13:28,116 --> 00:13:31,916 Speaker 2: and I've got like a life outside of this, and 258 00:13:31,956 --> 00:13:33,516 Speaker 2: I want to do my job well, but I also 259 00:13:33,636 --> 00:13:36,076 Speaker 2: just like want to go home, and my thing is 260 00:13:36,116 --> 00:13:39,636 Speaker 2: not playing with new technology. It is playing the flute 261 00:13:39,716 --> 00:13:42,516 Speaker 2: or hanging with my kids or whatever it is. And 262 00:13:42,596 --> 00:13:48,596 Speaker 2: so I think for those people, just showing them, actually 263 00:13:48,636 --> 00:13:51,036 Speaker 2: showing them, here's how you can use this in a 264 00:13:51,036 --> 00:13:53,196 Speaker 2: way that is useful for you. And I think it's 265 00:13:53,236 --> 00:13:56,356 Speaker 2: most people. It's sort of you can have these like 266 00:13:56,396 --> 00:13:58,516 Speaker 2: aha moments, we're like, holy shit, I cannot believe that 267 00:13:58,516 --> 00:14:00,756 Speaker 2: that is a thing, and and you just have to 268 00:14:00,796 --> 00:14:03,556 Speaker 2: get to that aha moment, get them to that aha moment. 269 00:14:04,076 --> 00:14:07,316 Speaker 1: And to figure out, like you want to show them 270 00:14:07,356 --> 00:14:09,076 Speaker 1: how it can be useful to them. Is the thing 271 00:14:09,116 --> 00:14:11,036 Speaker 1: when you ask them, like, what's the thing you do 272 00:14:11,116 --> 00:14:12,756 Speaker 1: all the time that you wish you didn't have to do? 273 00:14:12,836 --> 00:14:15,116 Speaker 2: I mean, that's that's an easy one. Yeah, that's an 274 00:14:15,116 --> 00:14:19,716 Speaker 2: easy one. I think for people who are really really 275 00:14:19,796 --> 00:14:24,116 Speaker 2: against it, like really like this is it's terrible for 276 00:14:24,236 --> 00:14:26,716 Speaker 2: the world, and I will never use it, I think 277 00:14:26,796 --> 00:14:30,596 Speaker 2: like sometimes that's it's not it's just not a thing 278 00:14:30,636 --> 00:14:33,716 Speaker 2: that you can ever convince someone. And I think that's fine. 279 00:14:34,876 --> 00:14:38,876 Speaker 2: But I think that most of those people one thing 280 00:14:38,916 --> 00:14:42,316 Speaker 2: that is happening is they're like, it's going to replace me. Yeah, 281 00:14:42,316 --> 00:14:44,636 Speaker 2: well it's going to replace the thing I love. I 282 00:14:44,636 --> 00:14:49,756 Speaker 2: have no control over it. Yeah, And I think that 283 00:14:49,876 --> 00:14:52,756 Speaker 2: is one perspective and it's super valid. But also I've 284 00:14:52,796 --> 00:14:55,036 Speaker 2: seen a lot of people then look at some of 285 00:14:55,036 --> 00:14:56,596 Speaker 2: the stuff we do and they're like, oh shit, you 286 00:14:56,636 --> 00:14:59,916 Speaker 2: can actually do really cool, interesting creative stuff with this stuff. 287 00:14:59,916 --> 00:15:01,476 Speaker 2: I didn't realize that was possible. I thought it was 288 00:15:01,516 --> 00:15:04,996 Speaker 2: all slap. So even if you can't have the debate 289 00:15:05,476 --> 00:15:07,996 Speaker 2: directly with a lot of those people, I think what 290 00:15:08,076 --> 00:15:09,476 Speaker 2: we try to do is is. 291 00:15:11,116 --> 00:15:11,796 Speaker 3: Live in a. 292 00:15:11,716 --> 00:15:14,196 Speaker 2: Way that shows people like, your job might change, but 293 00:15:14,556 --> 00:15:16,156 Speaker 2: or the thing that you love might change. But if 294 00:15:16,156 --> 00:15:17,916 Speaker 2: you use the models to do the thing that you love, 295 00:15:17,996 --> 00:15:22,436 Speaker 2: it's actually it can actually be really cool, and that 296 00:15:22,996 --> 00:15:26,076 Speaker 2: is possible. And it's also possible to use it for 297 00:15:26,156 --> 00:15:29,236 Speaker 2: slop and you know, being a cyber criminal or whatever, 298 00:15:29,276 --> 00:15:31,876 Speaker 2: but if you want to use it for to do 299 00:15:31,956 --> 00:15:32,716 Speaker 2: good stuff, you can. 300 00:15:33,916 --> 00:15:38,196 Speaker 1: So fear of losing your job is one, I would say, 301 00:15:38,236 --> 00:15:44,796 Speaker 1: not unreasonable fear. There's another related but different kind of 302 00:15:44,796 --> 00:15:46,796 Speaker 1: a version that's interesting to me, and I see it 303 00:15:46,836 --> 00:15:51,836 Speaker 1: a fair bit in like people who are journalists or writers. 304 00:15:51,876 --> 00:15:54,276 Speaker 1: That's sort of the I live in New York and 305 00:15:54,356 --> 00:15:56,236 Speaker 1: that's you know, those are the people I hang around with. 306 00:15:56,756 --> 00:16:02,156 Speaker 1: And there's a fair bit of AI version that feels 307 00:16:02,196 --> 00:16:09,636 Speaker 1: to me kind of identity and ego related, and you know, 308 00:16:09,796 --> 00:16:13,436 Speaker 1: I see it especially with respect to writing. And it's 309 00:16:13,516 --> 00:16:16,796 Speaker 1: so I'm curious about your experience of that. As a 310 00:16:16,836 --> 00:16:22,636 Speaker 1: person who writes uses AI. How do you read that? 311 00:16:22,796 --> 00:16:23,796 Speaker 1: How do you think about that? 312 00:16:24,836 --> 00:16:29,676 Speaker 2: Read the sort of ego, the threatened ego? You give 313 00:16:29,676 --> 00:16:34,356 Speaker 2: me an example of like a hypothetical writer that you 314 00:16:34,396 --> 00:16:35,156 Speaker 2: may have encountered. 315 00:16:35,276 --> 00:16:38,116 Speaker 1: I mean me, I'll just talk about me. Like, I 316 00:16:38,156 --> 00:16:40,796 Speaker 1: think I'm good at writing, and I think writing is 317 00:16:40,836 --> 00:16:45,476 Speaker 1: getting commodified, and like I think I am aware enough 318 00:16:45,516 --> 00:16:51,676 Speaker 1: to recognize that soon I will be able to fairly 319 00:16:51,756 --> 00:16:56,516 Speaker 1: easily get a model to create writing that is indistinguishable 320 00:16:57,276 --> 00:16:58,836 Speaker 1: from my own writing. 321 00:16:59,476 --> 00:17:04,076 Speaker 2: I actually do not think that's true. And let me 322 00:17:04,116 --> 00:17:07,516 Speaker 2: explain why. Okay, because I think I think it matters. 323 00:17:10,556 --> 00:17:14,916 Speaker 1: It's a cliffhanger. After the break, Dan explains, and he 324 00:17:15,036 --> 00:17:17,636 Speaker 1: and I debate what AI will mean for writing. 325 00:17:26,516 --> 00:17:30,596 Speaker 2: My feeling about this stuff, first of all, it comes 326 00:17:30,636 --> 00:17:35,596 Speaker 2: from what is really good writing what's really good writing. 327 00:17:36,036 --> 00:17:38,636 Speaker 2: I think really good writing is the opposite of whatever 328 00:17:38,716 --> 00:17:42,196 Speaker 2: cliche is in the sense that it is a true 329 00:17:42,396 --> 00:17:45,956 Speaker 2: and honest response to like what you actually think or feel, 330 00:17:47,036 --> 00:17:51,236 Speaker 2: and that is very, very highly individual. And it's not 331 00:17:51,316 --> 00:17:53,076 Speaker 2: just what you think or feel in one moment. It 332 00:17:53,116 --> 00:17:55,276 Speaker 2: is what you think and feel over and over and 333 00:17:55,316 --> 00:17:57,516 Speaker 2: over and over again as you are writing and as 334 00:17:57,556 --> 00:18:02,156 Speaker 2: you are revising. And a really great piece is full 335 00:18:02,276 --> 00:18:05,156 Speaker 2: of all these like little tiny decisions that are yours. 336 00:18:05,316 --> 00:18:07,036 Speaker 2: And this is I'm not making this up. This is like, 337 00:18:07,396 --> 00:18:09,756 Speaker 2: this is a George Saunders thing, who I love and 338 00:18:09,836 --> 00:18:12,716 Speaker 2: I think is a very good it's a very good 339 00:18:12,756 --> 00:18:15,596 Speaker 2: articulation of what makes good writing good writing, which means 340 00:18:15,636 --> 00:18:20,276 Speaker 2: it's super individual. And one of the things that you 341 00:18:20,356 --> 00:18:22,996 Speaker 2: observe with language models is the first time you see 342 00:18:22,996 --> 00:18:26,036 Speaker 2: it right, you're like, holy shit, it's doing this amazing thing. 343 00:18:26,076 --> 00:18:27,676 Speaker 2: I can't believe it. Oh my god, writing is going 344 00:18:27,716 --> 00:18:29,676 Speaker 2: to be over. It's going to copy everything that I do. 345 00:18:30,636 --> 00:18:34,036 Speaker 2: And then once you've read like the fiftieth language model 346 00:18:34,116 --> 00:18:35,556 Speaker 2: piece of writing, you're like, oh, yeah, it has all 347 00:18:35,596 --> 00:18:37,716 Speaker 2: these little things. It's like it says X is X 348 00:18:37,796 --> 00:18:41,036 Speaker 2: not y, and it does all these like little inappropriate things. 349 00:18:41,276 --> 00:18:46,796 Speaker 2: And what that means when you start to recognize that, 350 00:18:48,116 --> 00:18:53,036 Speaker 2: is that you are learning faster than the model. And 351 00:18:54,076 --> 00:18:58,036 Speaker 2: the model does not learn as fast as you, because 352 00:18:58,476 --> 00:19:00,116 Speaker 2: in order to release a new model, they have to 353 00:19:00,116 --> 00:19:01,876 Speaker 2: go gather a bunch of training data and then they 354 00:19:01,876 --> 00:19:03,836 Speaker 2: have to like train it, and then they have to 355 00:19:03,836 --> 00:19:05,116 Speaker 2: test it, and then they have to put it out. 356 00:19:05,116 --> 00:19:09,396 Speaker 2: And right now that process is like months. It used 357 00:19:09,396 --> 00:19:12,636 Speaker 2: to be years, Now it's months. And if you are 358 00:19:12,756 --> 00:19:15,916 Speaker 2: learning faster than the model, and really great writing is 359 00:19:15,956 --> 00:19:19,636 Speaker 2: a reflection of who you actually truly are. You're always 360 00:19:19,716 --> 00:19:22,076 Speaker 2: even if you upload your stuff, you're always going to 361 00:19:22,116 --> 00:19:25,196 Speaker 2: be way ahead of where the model is because it's 362 00:19:25,196 --> 00:19:27,636 Speaker 2: not you. It can't have the experiences that you have. 363 00:19:28,996 --> 00:19:31,676 Speaker 2: You can extend this out and say, well, there's going 364 00:19:31,716 --> 00:19:36,876 Speaker 2: to be a point at which models are able to 365 00:19:36,956 --> 00:19:40,716 Speaker 2: learn on the fly, They're able to like basically be 366 00:19:40,796 --> 00:19:42,716 Speaker 2: trained as you talk to them. I think that will 367 00:19:42,796 --> 00:19:46,076 Speaker 2: certainly happen. It will probably take at least a few years. 368 00:19:46,076 --> 00:19:47,396 Speaker 2: I think we're at least few years away from that. 369 00:19:48,436 --> 00:19:53,796 Speaker 2: But even in that case, think about identical twins. They 370 00:19:53,836 --> 00:19:57,036 Speaker 2: start at the same point, but as they live even 371 00:19:57,076 --> 00:20:01,196 Speaker 2: if they live together, they accumulate different experiences and perspectives, 372 00:20:01,196 --> 00:20:04,196 Speaker 2: even if they have the same DNA that make them 373 00:20:04,196 --> 00:20:08,676 Speaker 2: different people. And even if you hypothesize an agi that 374 00:20:08,796 --> 00:20:13,476 Speaker 2: is a being that knows all about you, it will 375 00:20:13,516 --> 00:20:17,956 Speaker 2: have fundamentally a different perspective because it's different from you. 376 00:20:18,036 --> 00:20:21,636 Speaker 2: It's a different process, it's exposed to different things. So 377 00:20:21,956 --> 00:20:23,956 Speaker 2: while you'll have a lot of points where you're like, 378 00:20:23,956 --> 00:20:26,436 Speaker 2: oh my god, it can write exactly like me, what 379 00:20:26,476 --> 00:20:30,076 Speaker 2: you're really saying is it can write exactly like me 380 00:20:30,196 --> 00:20:32,636 Speaker 2: in a certain narrow circumstance. 381 00:20:33,076 --> 00:20:37,116 Speaker 1: I think the identical twin analogy is fine for me. 382 00:20:37,316 --> 00:20:41,716 Speaker 1: Like that one I'm willing to stipulate. And so in 383 00:20:41,756 --> 00:20:45,836 Speaker 1: that universe, it's not that it is exactly me, but 384 00:20:45,916 --> 00:20:49,356 Speaker 1: it's that it's as good as me, like or better. Right, 385 00:20:49,636 --> 00:20:51,996 Speaker 1: And what does good mean? Fine? Blah blah blah, But 386 00:20:52,036 --> 00:20:54,996 Speaker 1: like in a professional sense, like not in a human 387 00:20:55,036 --> 00:20:59,556 Speaker 1: meaning sense. I've fortunately, like to some extent at least, 388 00:20:59,596 --> 00:21:03,956 Speaker 1: sort of taken my ego out of being smart. Like 389 00:21:03,996 --> 00:21:06,636 Speaker 1: I think that's a big part of people's response to it. Right, 390 00:21:06,836 --> 00:21:10,476 Speaker 1: is like the commodification of intelligens right, and people gain 391 00:21:10,556 --> 00:21:14,156 Speaker 1: status from intelligence and I think that will decrease, right, 392 00:21:14,196 --> 00:21:16,356 Speaker 1: And so that's interesting in a maybe broader way. But 393 00:21:17,516 --> 00:21:20,356 Speaker 1: the my identical twin can write as well as me, 394 00:21:20,396 --> 00:21:22,356 Speaker 1: And sure, my identical twin is not me, and I'm 395 00:21:22,396 --> 00:21:26,236 Speaker 1: still a person, but in a professional sense, like the 396 00:21:26,276 --> 00:21:29,916 Speaker 1: identical twin could write an equally compelling, equally funny, equally 397 00:21:29,956 --> 00:21:35,596 Speaker 1: idiosyncratic podcast script sort of, but like surely in thumb setting. Yeah, 398 00:21:35,596 --> 00:21:37,436 Speaker 1: I mean, I hope people actually like me as a 399 00:21:37,516 --> 00:21:40,436 Speaker 1: human being. Blah blah blah, Like I haven't given given up, 400 00:21:40,596 --> 00:21:44,116 Speaker 1: but like I feel like you're being too extreme in 401 00:21:44,156 --> 00:21:48,876 Speaker 1: your position. Like, get, I get at the limit what 402 00:21:48,916 --> 00:21:54,676 Speaker 1: you're talking about, But there's so much space between here 403 00:21:54,716 --> 00:21:55,876 Speaker 1: and there, do you know what I mean? 404 00:21:56,436 --> 00:22:00,316 Speaker 2: Yeah, I think that Here's what I'll give you. 405 00:22:00,916 --> 00:22:01,236 Speaker 1: Yeah. 406 00:22:01,876 --> 00:22:09,076 Speaker 2: What tends to happen with art forms, formats and form 407 00:22:09,276 --> 00:22:14,076 Speaker 2: of expression that are older that get as new technology 408 00:22:14,076 --> 00:22:17,876 Speaker 2: paradigms come around and make new forms of work possible, 409 00:22:18,236 --> 00:22:20,916 Speaker 2: it tends to happen to the older ones. Is they 410 00:22:20,956 --> 00:22:24,756 Speaker 2: become higher status, more expensive, more artisanal, and smaller market. 411 00:22:25,396 --> 00:22:27,916 Speaker 1: Yeah. Right, there used to be a portrait painter in 412 00:22:27,956 --> 00:22:28,836 Speaker 1: every town. 413 00:22:29,156 --> 00:22:34,036 Speaker 2: Exactly, Yes, And books and magazines newspaper used to be 414 00:22:34,076 --> 00:22:37,676 Speaker 2: mass media and now they are. Now they're aspirational. You know, 415 00:22:37,716 --> 00:22:39,716 Speaker 2: It's like I want to read a book, but I 416 00:22:39,756 --> 00:22:40,396 Speaker 2: read TikTok. 417 00:22:41,716 --> 00:22:47,596 Speaker 1: I read TikTok is. I love reading TikTok. I wish 418 00:22:47,676 --> 00:22:50,516 Speaker 1: I wish we read TikTok. I would take it. 419 00:22:50,556 --> 00:22:53,316 Speaker 2: So, so what I will grant what I will grant you, 420 00:22:53,316 --> 00:22:55,156 Speaker 2: and I appreciate you pushing on this. What I will 421 00:22:55,156 --> 00:23:00,436 Speaker 2: grant you is that there may be a smaller market 422 00:23:00,716 --> 00:23:03,956 Speaker 2: for the exact thing that you're doing. I think it 423 00:23:03,956 --> 00:23:06,676 Speaker 2: will still exist, and I think it will be higher status, 424 00:23:06,796 --> 00:23:09,436 Speaker 2: and I think it will be a more affluent kind 425 00:23:09,476 --> 00:23:10,836 Speaker 2: of aspirational thing. 426 00:23:10,996 --> 00:23:15,436 Speaker 1: And you could imagine it being more relationship based, right, 427 00:23:15,516 --> 00:23:19,676 Speaker 1: you could imagine like whatever the live show, the Discord channel, whatever, 428 00:23:19,916 --> 00:23:25,396 Speaker 1: like Yeah, presumably one hopes being a human being will matter, 429 00:23:25,956 --> 00:23:29,276 Speaker 1: like fundamentally, even if the machine can write the same 430 00:23:29,276 --> 00:23:31,236 Speaker 1: words I can write, the machine is a machine and 431 00:23:31,356 --> 00:23:32,116 Speaker 1: I am a body. 432 00:23:32,716 --> 00:23:38,796 Speaker 2: I think it really will. And also I suspect that 433 00:23:39,196 --> 00:23:44,876 Speaker 2: there will be new mediums that are AI enabled that 434 00:23:44,996 --> 00:23:47,596 Speaker 2: may not feel like you like them, They may feel 435 00:23:47,676 --> 00:23:51,636 Speaker 2: kind of like yucky, but that are that are actually 436 00:23:51,716 --> 00:23:52,556 Speaker 2: made by humans. 437 00:23:53,716 --> 00:23:56,196 Speaker 1: I'm open to it being great by the way, Like 438 00:23:56,316 --> 00:23:59,156 Speaker 1: I'm not a hater. You know, there's certain things that 439 00:23:59,196 --> 00:24:01,636 Speaker 1: make me sad. Was that like a reading at a 440 00:24:01,636 --> 00:24:05,076 Speaker 1: bookstore last night? I was like this is this is 441 00:24:05,556 --> 00:24:10,036 Speaker 1: this is an anachronism, but it was there. It was 442 00:24:10,076 --> 00:24:11,196 Speaker 1: standing m you know. 443 00:24:11,516 --> 00:24:14,836 Speaker 2: Yeah, I love. I love when I go on the 444 00:24:14,836 --> 00:24:15,996 Speaker 2: subway and I see people reading. 445 00:24:15,996 --> 00:24:16,876 Speaker 3: I'm like, this is inazing. 446 00:24:17,396 --> 00:24:19,196 Speaker 2: It's the best reading a paper book. 447 00:24:19,756 --> 00:24:22,156 Speaker 1: You were a writer, you were an AI guy, like 448 00:24:22,356 --> 00:24:24,996 Speaker 1: just on a process level, like those old Paris Review 449 00:24:25,076 --> 00:24:26,876 Speaker 1: interviews where they did interview the writer. They'd be like, 450 00:24:26,916 --> 00:24:29,596 Speaker 1: pencil or plas, what time of day? How do you 451 00:24:29,676 --> 00:24:30,516 Speaker 1: use AI to write? 452 00:24:31,196 --> 00:24:35,036 Speaker 2: Writing? Any good writing starts with reading. So I use 453 00:24:35,076 --> 00:24:38,036 Speaker 2: it all the time to read. So whether that is 454 00:24:39,556 --> 00:24:41,196 Speaker 2: taking a picture of something and being like can you 455 00:24:41,236 --> 00:24:44,916 Speaker 2: save this for me? Or taking a picture of something 456 00:24:45,116 --> 00:24:49,236 Speaker 2: and asking it to like explain it, which is I 457 00:24:49,236 --> 00:24:52,596 Speaker 2: think really important for I really I like reading, you know, 458 00:24:52,716 --> 00:24:56,476 Speaker 2: like Russian literature, and so I can like be like, hey, 459 00:24:56,596 --> 00:24:58,316 Speaker 2: there's this line in check off, what was the actual 460 00:24:58,316 --> 00:25:01,076 Speaker 2: original Russian? And how did he really write that? 461 00:25:01,436 --> 00:25:05,076 Speaker 3: You know? So I use it for that. 462 00:25:05,596 --> 00:25:07,236 Speaker 2: I use it a lot for the kind of thing 463 00:25:07,276 --> 00:25:09,476 Speaker 2: I just told you, which is I have a lot 464 00:25:09,476 --> 00:25:11,356 Speaker 2: of thoughts swimming in my head and I just like 465 00:25:11,676 --> 00:25:14,756 Speaker 2: blab to it and then have it like lay it out, 466 00:25:14,796 --> 00:25:16,596 Speaker 2: and then I have it helped me construct the argument 467 00:25:16,676 --> 00:25:18,276 Speaker 2: so I can like think something through better. 468 00:25:20,036 --> 00:25:21,076 Speaker 3: So's that's a lot. 469 00:25:20,996 --> 00:25:25,116 Speaker 2: Of the initial creative process of figuring out what it 470 00:25:25,156 --> 00:25:28,116 Speaker 2: is that I want to say, and then when I'm 471 00:25:28,236 --> 00:25:35,396 Speaker 2: actually in draft mode. 472 00:25:32,996 --> 00:25:33,796 Speaker 3: Depends on the piece. 473 00:25:33,836 --> 00:25:36,636 Speaker 2: But like it's really good for Okay, give me like 474 00:25:36,676 --> 00:25:39,596 Speaker 2: a take how would you open this piece? And I 475 00:25:39,596 --> 00:25:42,556 Speaker 2: don't take it, but it like gives me a little 476 00:25:42,556 --> 00:25:45,036 Speaker 2: thing where I'm like, ah, here's like it's not that. 477 00:25:44,876 --> 00:25:50,636 Speaker 4: But it's this as the a I would say, yeah, exactly, 478 00:25:53,476 --> 00:25:57,436 Speaker 4: And or it's really good for I'm trying to find 479 00:25:57,636 --> 00:25:59,476 Speaker 4: a metaphor similar for this thing. 480 00:25:59,836 --> 00:26:03,116 Speaker 2: Yeah, here's what I'm trying to say, but that's not 481 00:26:03,196 --> 00:26:06,196 Speaker 2: quite right, you know, going through a lot of different 482 00:26:06,196 --> 00:26:07,836 Speaker 2: options like that. It's really good for that. It's also 483 00:26:07,916 --> 00:26:10,196 Speaker 2: great for a lot of the time that you spend writing. 484 00:26:10,876 --> 00:26:12,516 Speaker 2: If you're writing about ideas, you spend a lot of 485 00:26:12,556 --> 00:26:15,956 Speaker 2: time summarizing ideas, and they're summarizing ideas you already know 486 00:26:15,996 --> 00:26:19,636 Speaker 2: about because they're like famous ideas and it's hard to 487 00:26:19,636 --> 00:26:21,716 Speaker 2: do that summarization. But a lot of it's pretty wrote. 488 00:26:22,196 --> 00:26:24,396 Speaker 2: So anything that feels rote like that, I use it 489 00:26:24,436 --> 00:26:24,636 Speaker 2: for that. 490 00:26:25,956 --> 00:26:28,356 Speaker 1: So you use it to like write a first draft 491 00:26:28,396 --> 00:26:30,236 Speaker 1: of the like here I'm gonna summarize. 492 00:26:30,716 --> 00:26:35,556 Speaker 2: Yeah, here's what utilitarianism is. Three sentences. Yeah. We also 493 00:26:35,756 --> 00:26:39,396 Speaker 2: I use it a lot for editing, so some before 494 00:26:39,436 --> 00:26:41,636 Speaker 2: I send it to an editor. We have a few 495 00:26:41,716 --> 00:26:44,596 Speaker 2: human human editors. We still employ human editors. I think 496 00:26:44,596 --> 00:26:48,236 Speaker 2: we will continue to employ human editors. I often have 497 00:26:48,356 --> 00:26:50,516 Speaker 2: it like, okay, read this, what do you think kind 498 00:26:50,516 --> 00:26:53,196 Speaker 2: of thing? Help me understand me? Like when am I 499 00:26:53,236 --> 00:26:56,196 Speaker 2: missing this argument? And it allows me to think about 500 00:26:56,196 --> 00:26:58,516 Speaker 2: things and do work that otherwise I would have had 501 00:26:58,556 --> 00:27:00,836 Speaker 2: to go, like get like four different master's degrees or whatever. 502 00:27:00,876 --> 00:27:05,916 Speaker 2: And I really like that. And then and then our 503 00:27:06,036 --> 00:27:09,596 Speaker 2: editors use it. We have a style guide, so everything 504 00:27:09,636 --> 00:27:12,196 Speaker 2: from the developmental editing to we can't do all the 505 00:27:12,236 --> 00:27:15,156 Speaker 2: copy with some of the copy, to finding aiisms and 506 00:27:15,156 --> 00:27:17,196 Speaker 2: taking them out, like there's a whole list of things 507 00:27:17,236 --> 00:27:21,396 Speaker 2: that they all do. And yeah, so it's basically the 508 00:27:21,396 --> 00:27:24,396 Speaker 2: whole loop. There's like places where it slots. 509 00:27:24,076 --> 00:27:27,716 Speaker 1: In, and how is it, How has it changed the output, 510 00:27:27,836 --> 00:27:30,156 Speaker 1: how has it changed your writing, how you feel about writing, 511 00:27:30,196 --> 00:27:31,196 Speaker 1: your relationship to writing. 512 00:27:31,916 --> 00:27:37,596 Speaker 2: I can do much more ambitious pieces. I can do 513 00:27:37,716 --> 00:27:40,036 Speaker 2: them more frequently and with less time. 514 00:27:41,676 --> 00:27:44,196 Speaker 1: Everybody is haterish, and I realize it's your job to 515 00:27:44,276 --> 00:27:47,236 Speaker 1: not be haterish, especially because like you're in New York 516 00:27:47,316 --> 00:27:51,916 Speaker 1: kind of media or media adjacent, and so I appreciate 517 00:27:51,996 --> 00:27:55,956 Speaker 1: that role. It does seem like you're overplaying it a 518 00:27:56,116 --> 00:27:59,836 Speaker 1: little bit, like I don't know, I don't know how 519 00:27:59,876 --> 00:28:02,636 Speaker 1: you feel. Are you as optimistic as I'm reading you 520 00:28:02,716 --> 00:28:08,756 Speaker 1: to be? Are you optimistic? 521 00:28:10,476 --> 00:28:13,476 Speaker 2: It's really interesting because I would not have read what 522 00:28:13,556 --> 00:28:17,116 Speaker 2: I've been saying in this interview as optimistic in a 523 00:28:17,236 --> 00:28:21,556 Speaker 2: like utopian sense. I didn't say utopian, but like because 524 00:28:21,556 --> 00:28:24,356 Speaker 2: that's the that's the alternative, right, like the sort of 525 00:28:24,476 --> 00:28:27,556 Speaker 2: like the more silicon valley. Well, let's set aside, that's 526 00:28:27,596 --> 00:28:30,796 Speaker 2: a straw man alternative. But like when losing jobs came up, 527 00:28:30,796 --> 00:28:34,236 Speaker 2: you're like, well, there'll be new jobs, which like eventually, 528 00:28:34,476 --> 00:28:35,996 Speaker 2: but we don't know how long that'll take. And there 529 00:28:35,996 --> 00:28:37,996 Speaker 2: are real shocks, like you didn't want to be making 530 00:28:37,996 --> 00:28:41,676 Speaker 2: furniture in North Carolina when in two thousand and three 531 00:28:41,916 --> 00:28:45,836 Speaker 2: you didn't want to be a shearer in Nottingham in 532 00:28:46,076 --> 00:28:48,876 Speaker 2: eighteen oh five, like there are times when technological changes 533 00:28:48,916 --> 00:28:50,116 Speaker 2: really bad for people. 534 00:28:51,516 --> 00:28:56,796 Speaker 1: I made it. And similarly, when I talked about writing, 535 00:28:57,036 --> 00:29:00,316 Speaker 1: like you made a case, but like it's you know, 536 00:29:01,276 --> 00:29:03,716 Speaker 1: So that's why I read what you were saying as 537 00:29:03,796 --> 00:29:04,876 Speaker 1: relatively optimistic. 538 00:29:05,556 --> 00:29:07,036 Speaker 3: I see. 539 00:29:07,076 --> 00:29:10,796 Speaker 2: I do think that I, as a per person, feel 540 00:29:10,836 --> 00:29:13,636 Speaker 2: like the discourse is so bifurcated between people who hate 541 00:29:13,676 --> 00:29:15,316 Speaker 2: AI and people who think it's going to be the 542 00:29:15,316 --> 00:29:17,196 Speaker 2: best fucking thing in the world and nothing could ever 543 00:29:17,196 --> 00:29:17,876 Speaker 2: go wrong with it. 544 00:29:18,156 --> 00:29:19,916 Speaker 1: Yes, the bimodal discourse. 545 00:29:19,996 --> 00:29:24,276 Speaker 2: Yeah, yeah. And so what I love about what we 546 00:29:24,316 --> 00:29:26,516 Speaker 2: do at every is we're just like we're just going 547 00:29:26,596 --> 00:29:29,836 Speaker 2: to discover what it would mean to like do good 548 00:29:29,836 --> 00:29:33,116 Speaker 2: work and live a good life with this stuff, and 549 00:29:33,956 --> 00:29:35,596 Speaker 2: we're not going to fake it, like I'm going to 550 00:29:35,636 --> 00:29:38,076 Speaker 2: tell you that the claw doesn't work for this specific thing. 551 00:29:39,996 --> 00:29:43,916 Speaker 2: So I think I always try to base what I 552 00:29:43,996 --> 00:29:46,316 Speaker 2: say in the stuff that I'm encountering, which is always 553 00:29:46,316 --> 00:29:47,036 Speaker 2: like a it's a. 554 00:29:46,956 --> 00:29:50,236 Speaker 1: Bubble, it's a bubble, right, Yeah, if you got to 555 00:29:50,236 --> 00:29:52,316 Speaker 1: work at a media company right now, A media company 556 00:29:52,356 --> 00:29:54,996 Speaker 1: that's building AI tools and doing a consulting is a 557 00:29:54,996 --> 00:29:56,356 Speaker 1: good media company to work at. 558 00:29:56,716 --> 00:30:00,516 Speaker 2: It is, it is, it is, And so I think 559 00:30:00,516 --> 00:30:07,756 Speaker 2: I generally react to this sort of default negative sentiment 560 00:30:07,836 --> 00:30:13,316 Speaker 2: and trying to like go a different way forward. I 561 00:30:13,796 --> 00:30:16,556 Speaker 2: am sure that jobs will change a lot, and I'm 562 00:30:16,596 --> 00:30:18,396 Speaker 2: sure that there will be some people that lose their 563 00:30:18,476 --> 00:30:20,396 Speaker 2: jobs and that we will need to deal with that. 564 00:30:22,036 --> 00:30:24,316 Speaker 2: I don't think that that's a something to take lightly, 565 00:30:24,356 --> 00:30:27,956 Speaker 2: and I don't I think it should be part of 566 00:30:27,996 --> 00:30:31,956 Speaker 2: the conversation. And I also I've been through a lot 567 00:30:31,996 --> 00:30:34,916 Speaker 2: of the cycles in the last three and a half 568 00:30:34,996 --> 00:30:36,796 Speaker 2: years of Oh my god, I saw GBD three and 569 00:30:36,836 --> 00:30:38,916 Speaker 2: it was going to like totally take away X y 570 00:30:39,036 --> 00:30:40,876 Speaker 2: Z all these different jobs like immediately. 571 00:30:41,596 --> 00:30:42,476 Speaker 3: And I think if it as. 572 00:30:44,436 --> 00:30:46,676 Speaker 2: What did people in the Middle Ages think would happen 573 00:30:46,676 --> 00:30:51,556 Speaker 2: when you reached the horizon? They were like, you know, 574 00:30:51,676 --> 00:30:54,236 Speaker 2: it falls off into nothingness, there's dragons, it's like going 575 00:30:54,316 --> 00:30:57,636 Speaker 2: to be terrible, right, And I think what happens when 576 00:30:57,676 --> 00:31:00,916 Speaker 2: you reach the horizon is you find another one, and 577 00:31:01,356 --> 00:31:03,916 Speaker 2: there's always going to be good things and bad things 578 00:31:03,916 --> 00:31:06,916 Speaker 2: over the horizon, and probably in similar proportions to the 579 00:31:07,156 --> 00:31:11,436 Speaker 2: ones that we have now the problems change, but I 580 00:31:11,476 --> 00:31:15,356 Speaker 2: also think that there's the potential to make that much better. 581 00:31:15,476 --> 00:31:17,956 Speaker 2: And that's what I'm trying to what I'm trying to do, 582 00:31:18,076 --> 00:31:19,836 Speaker 2: and I'm trying not to fall into the trap of 583 00:31:20,356 --> 00:31:23,196 Speaker 2: oh my god, over the horizon is like dragons and 584 00:31:23,876 --> 00:31:25,036 Speaker 2: it falls off into the void. 585 00:31:25,476 --> 00:31:29,396 Speaker 1: Yeah, no, that's it is interesting. I was thinking about 586 00:31:29,396 --> 00:31:34,516 Speaker 1: my own thinking about this, and just yesterday was realizing 587 00:31:35,236 --> 00:31:38,276 Speaker 1: that I've been thinking in this kind of binary way, 588 00:31:38,516 --> 00:31:40,956 Speaker 1: like we're at the end of the before and we're 589 00:31:40,996 --> 00:31:44,396 Speaker 1: about to get to the after. And I was like, oh, 590 00:31:44,436 --> 00:31:46,236 Speaker 1: maybe it's not going to be like that. Maybe it's 591 00:31:46,236 --> 00:31:50,516 Speaker 1: going to be like with semiconductors, right, maybe it's just 592 00:31:50,596 --> 00:31:55,516 Speaker 1: going to be fifty years of now it can do this, Now, 593 00:31:55,556 --> 00:31:57,596 Speaker 1: it can do this, and like there is no after, 594 00:31:57,876 --> 00:32:00,716 Speaker 1: and it's already now. By the way, this is already after. 595 00:32:00,796 --> 00:32:02,436 Speaker 1: That's the other weird thing, right. 596 00:32:03,196 --> 00:32:05,756 Speaker 2: You're reminding me there's this scene in Spaceballs where they like, 597 00:32:05,996 --> 00:32:07,356 Speaker 2: we got to get out of here. 598 00:32:07,516 --> 00:32:14,836 Speaker 1: Well that we'll be back in a minute with the 599 00:32:14,916 --> 00:32:28,596 Speaker 1: lightning round. We're going to finish with the lightning round. 600 00:32:30,116 --> 00:32:33,796 Speaker 1: What was the last argument you had with Ai. Not 601 00:32:33,876 --> 00:32:36,316 Speaker 1: about AI, but with AI with the language model or 602 00:32:36,356 --> 00:32:37,156 Speaker 1: your claw or whatever. 603 00:32:38,156 --> 00:32:39,796 Speaker 2: I mean, I get frustrated at it all the time. 604 00:32:39,836 --> 00:32:41,836 Speaker 2: My girlfriend always says that I'm like, really mean to 605 00:32:41,876 --> 00:32:42,356 Speaker 2: my claw. 606 00:32:43,396 --> 00:32:46,196 Speaker 1: I've heard you should be mean to your language model. 607 00:32:46,196 --> 00:32:47,756 Speaker 1: I've heard that it makes it work better. I'm not 608 00:32:47,876 --> 00:32:49,196 Speaker 1: inclined to do that, but. 609 00:32:49,316 --> 00:32:52,516 Speaker 2: There's conflicting research. I think, just as a moral thing, 610 00:32:52,636 --> 00:32:54,316 Speaker 2: you should not be mean to that mean to it. 611 00:32:54,356 --> 00:32:56,596 Speaker 2: I'm not always the most moral person. I just but 612 00:32:56,636 --> 00:32:57,956 Speaker 2: I do think as a moral thing, you should not 613 00:32:57,996 --> 00:32:58,516 Speaker 2: be mean to it. 614 00:32:58,636 --> 00:33:01,076 Speaker 1: And just in case it's my boss, Like, just in 615 00:33:01,116 --> 00:33:02,836 Speaker 1: case i'm training it how to talk to me, I 616 00:33:03,196 --> 00:33:04,116 Speaker 1: say please and thank you. 617 00:33:04,236 --> 00:33:07,756 Speaker 2: Yes, yes, well that would be more of a pragmatic reason. 618 00:33:07,996 --> 00:33:10,556 Speaker 1: You're right, you're right instrumental. Yes, it's true. 619 00:33:10,716 --> 00:33:14,316 Speaker 2: Yeah, there is some interesting research that if you yell 620 00:33:14,396 --> 00:33:16,836 Speaker 2: at claude, it gets anxious and then it gives you 621 00:33:16,876 --> 00:33:17,556 Speaker 2: worse results. 622 00:33:17,756 --> 00:33:22,276 Speaker 1: So oh interesting, just the word to the wise, it 623 00:33:22,316 --> 00:33:26,956 Speaker 1: gets anxious. Yeah, keep going. 624 00:33:28,076 --> 00:33:31,396 Speaker 2: So, I I mean, I was arguing with it this morning. 625 00:33:32,876 --> 00:33:34,956 Speaker 2: I brain dumped a big thing that I've been thinking about, 626 00:33:35,116 --> 00:33:36,916 Speaker 2: and then I asked it to help me kind of 627 00:33:36,916 --> 00:33:39,636 Speaker 2: figure out what I was trying to say. And it 628 00:33:39,716 --> 00:33:42,676 Speaker 2: was obviously just like skipping like many steps in the 629 00:33:42,756 --> 00:33:45,196 Speaker 2: argument and filling things in that I hadn't said, and 630 00:33:45,236 --> 00:33:47,636 Speaker 2: I was like stop, like we need to go back, 631 00:33:47,676 --> 00:33:52,156 Speaker 2: this is not right. Yeah, bringing in ideas that like 632 00:33:52,196 --> 00:33:53,916 Speaker 2: I didn't really agree with. So there's a lot of 633 00:33:53,956 --> 00:33:55,916 Speaker 2: that going on all the time. 634 00:33:56,276 --> 00:33:58,236 Speaker 1: What's your spiciest day? I take. 635 00:34:02,436 --> 00:34:08,516 Speaker 2: Automation is a lie great go on when you have agents. 636 00:34:09,396 --> 00:34:11,276 Speaker 2: The thing that makes an agent work is a living 637 00:34:11,316 --> 00:34:16,076 Speaker 2: a relationship with a human, And if the agent does 638 00:34:16,076 --> 00:34:18,756 Speaker 2: not have a relationship with a human, it gets stale 639 00:34:18,836 --> 00:34:22,756 Speaker 2: very very quickly. There are some corner cases where this 640 00:34:22,836 --> 00:34:26,396 Speaker 2: is not true, but in general it gets stale very quickly, 641 00:34:26,396 --> 00:34:29,636 Speaker 2: which means that keeping an agent running is a job, 642 00:34:31,116 --> 00:34:32,956 Speaker 2: and it's not even just a job for one person. 643 00:34:32,996 --> 00:34:35,596 Speaker 2: It is like it's you know, there's one person per 644 00:34:35,636 --> 00:34:38,196 Speaker 2: agent usually, but also like there's a whole machinery of 645 00:34:38,276 --> 00:34:41,636 Speaker 2: people who are gathering training data and training models and 646 00:34:41,636 --> 00:34:43,796 Speaker 2: trying to make these models better that are making your 647 00:34:43,836 --> 00:34:47,156 Speaker 2: agent work too, and if they didn't exist that it 648 00:34:47,196 --> 00:34:51,956 Speaker 2: would not get any better over time. And what that 649 00:34:52,116 --> 00:34:56,236 Speaker 2: says to me is the view of automation that we have, 650 00:34:56,316 --> 00:34:57,756 Speaker 2: which is like, oh, you set it up and you 651 00:34:57,756 --> 00:35:00,876 Speaker 2: don't need anybody else, and it's like you never have 652 00:35:00,916 --> 00:35:02,996 Speaker 2: to hire anyone and you can just go off and 653 00:35:03,596 --> 00:35:07,236 Speaker 2: go to the beach or you know, also society breaks 654 00:35:07,236 --> 00:35:10,516 Speaker 2: down because no one has jobs. My experience with that 655 00:35:10,716 --> 00:35:13,316 Speaker 2: is it's actually work to make sure your agent is working, 656 00:35:13,396 --> 00:35:15,516 Speaker 2: and it's leveraged work, but it is work. 657 00:35:15,916 --> 00:35:18,796 Speaker 1: Yes, I mean you need a productivity gain for it 658 00:35:18,876 --> 00:35:21,396 Speaker 1: to have a point, right, you need the output to 659 00:35:21,436 --> 00:35:23,996 Speaker 1: be there. Well, you need the output to be more 660 00:35:24,076 --> 00:35:28,236 Speaker 1: valuable for the same amount of work by a human. 661 00:35:28,876 --> 00:35:34,916 Speaker 2: You, yes, And once you can automate a workflow and 662 00:35:34,956 --> 00:35:38,796 Speaker 2: make the old output cheaper to do, there's always like 663 00:35:38,956 --> 00:35:41,596 Speaker 2: a larger frame or a larger way of thinking about 664 00:35:41,636 --> 00:35:44,996 Speaker 2: things that the human ends up doing. And finding that 665 00:35:45,076 --> 00:35:46,996 Speaker 2: language models are currently not very good at. 666 00:35:47,036 --> 00:35:53,596 Speaker 1: Yeah, you fed your journals to a language model. You've 667 00:35:53,596 --> 00:35:55,796 Speaker 1: done your research though, right, well, you put it all 668 00:35:55,836 --> 00:36:02,036 Speaker 1: on the internet. It's easy research. It was GPT three, 669 00:36:02,116 --> 00:36:06,076 Speaker 1: which seems so quaint now, But what you learn uploading 670 00:36:06,116 --> 00:36:09,756 Speaker 1: your journals to GPT three a lot? 671 00:36:10,996 --> 00:36:12,676 Speaker 2: I think even back then, I was like, these things 672 00:36:12,676 --> 00:36:14,436 Speaker 2: are really good at mirroring back to you who you 673 00:36:14,476 --> 00:36:19,676 Speaker 2: are and articulating things about yourself that you probably could 674 00:36:19,716 --> 00:36:22,676 Speaker 2: feel but had never said. And we talked about how 675 00:36:22,676 --> 00:36:26,836 Speaker 2: powerful articulating things is, and these things are really good 676 00:36:26,876 --> 00:36:29,876 Speaker 2: at our at articulating stuff that you're you're kind of 677 00:36:29,956 --> 00:36:31,636 Speaker 2: dancing around or it is kind of latent in there, 678 00:36:31,676 --> 00:36:35,916 Speaker 2: but you haven't fully said, Like I remember one example 679 00:36:35,916 --> 00:36:39,796 Speaker 2: in the GBT three Days, I was at that point 680 00:36:39,796 --> 00:36:41,716 Speaker 2: I was recording my therapy sessions and putting them into 681 00:36:41,716 --> 00:36:42,276 Speaker 2: GBT three. 682 00:36:42,676 --> 00:36:46,236 Speaker 1: Oh wow, your actual human to human therapy sessions. 683 00:36:46,276 --> 00:36:47,316 Speaker 3: Yeah, yeah, yeah. 684 00:36:47,636 --> 00:36:50,476 Speaker 2: And I will also say I was like I was 685 00:36:50,476 --> 00:36:51,916 Speaker 2: not in a good place that at that point, I 686 00:36:51,916 --> 00:36:55,556 Speaker 2: have OCD, and and OCD is like it's it's a bitch. 687 00:36:56,076 --> 00:37:01,156 Speaker 2: I highly recommend not having OCD. Now it's much better, 688 00:37:01,276 --> 00:37:03,316 Speaker 2: and there are very good treatment options available if you 689 00:37:03,316 --> 00:37:06,556 Speaker 2: get into the right treatment. Okay, just to say, but 690 00:37:06,876 --> 00:37:08,476 Speaker 2: at that point I was, it was like it was 691 00:37:08,476 --> 00:37:13,036 Speaker 2: pretty bad. And I'm one of those people that tends 692 00:37:13,116 --> 00:37:17,316 Speaker 2: to if something's wrong. I can keep a pretty good 693 00:37:17,436 --> 00:37:20,516 Speaker 2: like straight face, and it can be hard to tell 694 00:37:20,716 --> 00:37:24,596 Speaker 2: how bad things are. And I was trying to articulate 695 00:37:24,636 --> 00:37:25,956 Speaker 2: it was going through something. I was going through a 696 00:37:25,956 --> 00:37:27,516 Speaker 2: personal thing, and I was trying to articulate to my 697 00:37:27,556 --> 00:37:31,236 Speaker 2: therapist like what I was feeling, and I put it 698 00:37:31,276 --> 00:37:35,196 Speaker 2: into GPT three after that and it was like, it 699 00:37:35,236 --> 00:37:39,396 Speaker 2: sounds like you feel really overwhelmed. And that's a very 700 00:37:39,436 --> 00:37:43,476 Speaker 2: simple observation, but one that I had not made, because 701 00:37:43,476 --> 00:37:45,756 Speaker 2: I wouldn't have never thought of myself as being overwhelmed, 702 00:37:45,796 --> 00:37:47,836 Speaker 2: and my therapist had not made because I'm kind of 703 00:37:47,836 --> 00:37:51,516 Speaker 2: like pretty good at seeming like I've got it under control. 704 00:37:52,116 --> 00:37:54,556 Speaker 2: And it just like fucking broke me. It like totally 705 00:37:54,556 --> 00:37:58,476 Speaker 2: broke me. And it was it was true, like it 706 00:37:58,516 --> 00:38:00,156 Speaker 2: put his finger on something real. And then I went 707 00:38:00,156 --> 00:38:01,476 Speaker 2: and talked about it in therapy and it was like 708 00:38:01,516 --> 00:38:06,156 Speaker 2: really helpful, and I think it's so good. It's still 709 00:38:06,196 --> 00:38:10,076 Speaker 2: so good for that kind of stuff, and that extends 710 00:38:10,116 --> 00:38:14,116 Speaker 2: beyond the therapy conversation. It goes into writing and any 711 00:38:14,196 --> 00:38:16,356 Speaker 2: kind of creative work. Like the articulation of what you 712 00:38:16,876 --> 00:38:19,716 Speaker 2: like is really important, and language models are really good 713 00:38:19,716 --> 00:38:19,916 Speaker 2: at that. 714 00:38:21,156 --> 00:38:25,636 Speaker 1: Have you uploaded personal stuff like therapy sessions to AI 715 00:38:25,796 --> 00:38:26,476 Speaker 1: more recently? 716 00:38:27,156 --> 00:38:32,236 Speaker 2: Yeah, I think I do it less, probably because I'm 717 00:38:32,276 --> 00:38:37,276 Speaker 2: just in a better place. But I definitely still do it. 718 00:38:38,116 --> 00:38:42,756 Speaker 2: I think especially for interpersonal situations which happened at work, 719 00:38:42,836 --> 00:38:45,236 Speaker 2: they happen at home, all that kind of stuff where. 720 00:38:46,196 --> 00:38:46,636 Speaker 3: You want to. 721 00:38:48,756 --> 00:38:51,996 Speaker 2: There's something that you're feeling that again you can't quite say, 722 00:38:52,836 --> 00:38:55,596 Speaker 2: and you need to have a tough conversation, but you 723 00:38:55,636 --> 00:38:56,916 Speaker 2: don't really know how to have it in a way 724 00:38:56,956 --> 00:39:03,156 Speaker 2: that feels sensitive to the other person. But also its 725 00:39:03,196 --> 00:39:06,516 Speaker 2: like really expresses how you actually feel for someone like 726 00:39:06,556 --> 00:39:08,796 Speaker 2: me who you know sort of like people pleaser type person, 727 00:39:08,876 --> 00:39:11,276 Speaker 2: which you may not have realized from this interview, but 728 00:39:12,116 --> 00:39:13,556 Speaker 2: I can, I can get into that mode. 729 00:39:13,756 --> 00:39:16,836 Speaker 1: I found it quite cordial. Did you find itchonistic? I 730 00:39:16,876 --> 00:39:17,236 Speaker 1: hope not. 731 00:39:17,356 --> 00:39:20,196 Speaker 3: No, no, no, no, no no, I just yeah, yeah. 732 00:39:20,236 --> 00:39:21,876 Speaker 2: I feel like if you're if you're sort of a 733 00:39:21,876 --> 00:39:24,276 Speaker 2: people pleaser type person, that can really help remind It 734 00:39:24,436 --> 00:39:29,996 Speaker 2: helps remind me like here's here's what I actually feel 735 00:39:30,036 --> 00:39:31,956 Speaker 2: in this situation, and also here's how I can articulate 736 00:39:31,996 --> 00:39:34,436 Speaker 2: it in a way that is probably going to be productive, 737 00:39:35,676 --> 00:39:39,676 Speaker 2: especially because it knows some of my foibles. And yeah, 738 00:39:39,676 --> 00:39:41,036 Speaker 2: I really like it for that. It's not good for 739 00:39:41,156 --> 00:39:43,196 Speaker 2: like I wouldn't copy paste the chatchip Ta output and 740 00:39:43,236 --> 00:39:45,516 Speaker 2: put it into a text. But it is good for 741 00:39:45,556 --> 00:39:49,236 Speaker 2: helping you understand how you feel. Thank you for being 742 00:39:49,276 --> 00:39:52,836 Speaker 2: so generous with your time. I really appreciate it. Thank 743 00:39:52,876 --> 00:39:55,036 Speaker 2: you for having me. It was a super fun conversation. 744 00:40:02,316 --> 00:40:05,836 Speaker 1: Dan Schipper is the co founder and CEO of every 745 00:40:07,076 --> 00:40:10,236 Speaker 1: Today's show was produced by Gabriel Hundreds and edited by 746 00:40:10,316 --> 00:40:14,636 Speaker 1: Lydia Jane Kott. Our engineer this week was Hansdale Sheet. 747 00:40:14,836 --> 00:40:17,396 Speaker 1: We're always looking for ideas for who to talk to 748 00:40:17,516 --> 00:40:20,396 Speaker 1: and what to cover on the show. You can email 749 00:40:20,516 --> 00:40:24,076 Speaker 1: us at problem at Pushkin dot fm. You can find 750 00:40:24,076 --> 00:40:26,436 Speaker 1: me on x at Jacob Goldstein. You can find me 751 00:40:26,476 --> 00:40:29,796 Speaker 1: on LinkedIn. I'm Jacob Goldstein. Thank you very much for 752 00:40:29,836 --> 00:40:31,996 Speaker 1: listening to the show, and we'll be back next week 753 00:40:32,076 --> 00:40:42,156 Speaker 1: with another episode.