WEBVTT 1 00:00:00.091 --> 00:00:01.653 The industry is in a tailspin. 2 00:00:02.174 --> 00:00:05.998 There are fewer opportunities for writers and producers and people who make things. 3 00:00:06.299 --> 00:00:18.293 And I really had the conviction that AI can help you figure out who you are, how you have to survive in this world, and also how to make things that will then promote you and promote your career. 4 00:00:19.014 --> 00:00:26.020 Depending on who you ask, right now is an apocalypse for creative people, or it's like the best time to be alive ever. 5 00:00:26.320 --> 00:00:30.364 I think it's one of the most important questions about AI as a technology. 6 00:00:30.604 --> 00:00:35.148 Can you use it to make good art, or can you only use it to make slop? 7 00:00:35.488 --> 00:00:37.410 That's why I'm extremely excited for this episode. 8 00:00:37.570 --> 00:00:40.612 My co-host, Natalia Quintero, interviewed Fred Graver. 9 00:00:40.752 --> 00:00:45.416 Fred is a four-time Emmy Award-winning filmmaker and technologist who 10 00:00:46.297 --> 00:00:51.121 is using AI all the time to make really great stuff in film. 11 00:00:51.361 --> 00:01:00.769 And he came on to share how he uses it, how it can be used as a creative tool, and the dirty secret in Hollywood that a lot of people use it even if they won't admit to it. 12 00:01:01.649 --> 00:01:03.631 So, Natalia, tell us about the conversation. 13 00:01:04.064 --> 00:01:11.027 It was so cool to talk to Fred and to learn from someone who's been a leader at the top of their industry for so many decades. 14 00:01:11.747 --> 00:01:16.349 And what's interesting is that Fred has actually done this a few times. 15 00:01:16.769 --> 00:01:19.810 He started his career as a Disney Imagineer. 16 00:01:19.830 --> 00:01:21.451 That's the coolest thing. 17 00:01:21.471 --> 00:01:22.832 Kind of the dream job. 18 00:01:24.692 --> 00:01:29.996 And then he built this incredible career at VH1 and MTV. 19 00:01:30.216 --> 00:01:35.400 And then he kind of threw it all away and became the head of TV at Twitter when it was a startup. 20 00:01:35.460 --> 00:01:37.101 And that didn't totally make sense. 21 00:01:37.681 --> 00:01:43.445 But he had this really interesting thesis of wanting to get closer and closer to the audience. 22 00:01:43.825 --> 00:01:45.367 And now he's done it again. 23 00:01:45.547 --> 00:01:47.348 You know, he's basically built what... 24 00:01:47.668 --> 00:01:59.615 will be compound producing and rebuild the production process from the ground up using his decades of expertise, an Emmy award winning artist to do it all over again. 25 00:02:00.035 --> 00:02:02.777 And we all get to use his skills. 26 00:02:03.157 --> 00:02:06.179 So if you want to be a producer, now you can produce like Fred. 27 00:02:07.019 --> 00:02:07.359 I love it. 28 00:02:07.459 --> 00:02:08.380 It's a great episode. 29 00:02:08.880 --> 00:02:09.360 Let's dive in. 30 00:02:22.195 --> 00:02:24.056 So Fred, thank you so much for being with us. 31 00:02:24.516 --> 00:02:31.878 And I want to start by actually just maybe getting a little bit of background on your career and how you got here. 32 00:02:31.898 --> 00:02:35.119 This maybe let's start at the top. 33 00:02:35.359 --> 00:02:37.319 You're a professional writer. 34 00:02:37.680 --> 00:02:37.900 Yes. 35 00:02:38.140 --> 00:02:39.040 How did you get into that? 36 00:02:39.647 --> 00:02:44.274 I came to New York and I could type over 100 words a minute. 37 00:02:44.634 --> 00:02:51.164 And we are back in the day of IBM Selectrics and printers copies and all of that. 38 00:02:51.324 --> 00:02:54.229 And I went to work and, oh, and I knew how to copy edit. 39 00:02:55.414 --> 00:02:57.936 And I went to work at the National Lampoon. 40 00:02:58.496 --> 00:03:04.500 And I was hired there because I could type and I could copy edit and I wouldn't screw up the joke. 41 00:03:05.081 --> 00:03:08.563 So a bad copy editor would go, well, this isn't grammatically correct. 42 00:03:09.003 --> 00:03:10.925 And I would understand, no, that's the joke. 43 00:03:11.900 --> 00:03:17.024 And I began editing stuff at the National Lampoon and writing for the National Lampoon. 44 00:03:17.384 --> 00:03:27.932 So that's, yes, that's how I got started, which for people who are watching, who are in every audience, yes, your technical skills count. 45 00:03:29.133 --> 00:03:37.899 They will often get you in the door, particularly to jump ahead, you know, the whole kind of, you're gonna get replaced by somebody who knows how to use AI. 46 00:03:39.000 --> 00:03:41.122 People who understand what's happening under the hood 47 00:03:42.044 --> 00:03:45.146 are going to be incredibly in demand, I think. 48 00:03:45.587 --> 00:03:53.752 So I went to work from there for David Letterman, for Late Night with David Letterman, and stayed there for a long time. 49 00:03:54.773 --> 00:03:56.534 It was a great gig, and 50 00:03:57.488 --> 00:04:00.469 Then went to work for Norman Lear. 51 00:04:00.949 --> 00:04:03.329 In the 70s, Norman was the king of television. 52 00:04:03.349 --> 00:04:09.210 I think he had four or five hit shows in the top 10 of the Nielsen rankings at one time. 53 00:04:10.030 --> 00:04:15.251 And he also was socially aware and very smart, and he was doing some new things with television. 54 00:04:15.431 --> 00:04:16.731 And I got hired, I worked with him. 55 00:04:16.751 --> 00:04:23.753 It was an incredible learning experience, which then translated into my being a comedy writer, and I worked for Cheers and In Living Color. 56 00:04:24.453 --> 00:04:26.313 And then I heard a modem. 57 00:04:28.213 --> 00:04:34.224 And the first time I heard a modem, which was actually in the 80s, I just went, okay, this changes everything. 58 00:04:35.025 --> 00:04:40.675 Everything's going to be on something on your desk, and you're going to be able to do everything yourself. 59 00:04:43.398 --> 00:04:48.662 And so my first kind of tech gig was I was working with the Disney Imagineers. 60 00:04:49.883 --> 00:04:53.205 And there was this effort back in the day. 61 00:04:53.225 --> 00:04:56.708 It was Mike Ovitz and Jeffrey Katzenberg and those guys. 62 00:04:56.848 --> 00:04:58.169 And they were terrified. 63 00:04:58.289 --> 00:05:00.331 This was like the late 90s, early 2000s. 64 00:05:00.631 --> 00:05:02.152 They were terrified by the Internet. 65 00:05:02.632 --> 00:05:07.296 But, you know, they began putting real bets down on what is this thing going to be. 66 00:05:07.956 --> 00:05:08.237 I was... 67 00:05:09.017 --> 00:05:12.958 there and I was at MTV networks in the kind of early web days. 68 00:05:14.159 --> 00:05:15.699 And then I took another right turn. 69 00:05:16.060 --> 00:05:17.200 I got hired by Twitter. 70 00:05:17.980 --> 00:05:20.381 This would be late 2000s, right? 71 00:05:20.461 --> 00:05:22.302 It's like the second year that Twitter was around. 72 00:05:22.542 --> 00:05:31.985 And the theory was if famous people that people love, celebrities or TV shows, are on Twitter, then people will get on Twitter. 73 00:05:32.125 --> 00:05:34.246 And that's the best advertising we could do. 74 00:05:34.666 --> 00:05:36.427 And so that was my job was to get 75 00:05:37.047 --> 00:05:40.288 the bird and real use cases onto the screen. 76 00:05:41.548 --> 00:05:47.529 Probably best case that ever happened was the famous Ellen DeGeneres selfie. 77 00:05:47.929 --> 00:05:51.630 That was how many million dollar worth of advertising. 78 00:05:51.690 --> 00:06:03.012 I did that and then I was doing some consulting and things and somebody said to me, hey, there's this machine learning thing that's AI 79 00:06:04.332 --> 00:06:10.698 And the engineers don't understand how people talk. 80 00:06:10.879 --> 00:06:15.143 And people don't know what to talk about when they get to the machine. 81 00:06:15.803 --> 00:06:20.127 And so I became a conversation designer at Microsoft. 82 00:06:20.508 --> 00:06:22.910 I was literally scripting bots. 83 00:06:23.451 --> 00:06:24.832 And I was working with the AI teams. 84 00:06:24.992 --> 00:06:25.352 And it was... 85 00:06:26.193 --> 00:06:39.398 It was weirdly fascinating because you're just dealing with what's going to become AI, what's going to become this interface, and trying to figure out what will that conversation be. 86 00:06:40.078 --> 00:06:41.658 Also, it is a conversation. 87 00:06:41.939 --> 00:06:50.922 I think what a lot of people, when they freak out about AI, and they freak out about an enormous box of things instead of all the things that are inside, but they're kind of 88 00:06:52.551 --> 00:06:57.534 they use it as a search engine or it gave me the wrong... and they don't realize, no, it's a relationship. 89 00:06:58.335 --> 00:07:01.917 I was doing that, I was writing, then they moved me to writing presentations. 90 00:07:02.657 --> 00:07:10.201 Microsoft's great kind of secret sauce is they have decades of data about how people use their software. 91 00:07:10.381 --> 00:07:17.906 If you're United Airlines and people are calling in to do whatever they're going to do, can we have 92 00:07:18.986 --> 00:07:43.216 AI bot can we have co-pilot go through everything you we already know from 30 years of you guys using software to help them solve their problem And so I did that for for for quite a while writing these presentations Saying here's what AI is and here's what it does and and Here's your problems and the solution to your problems oddly enough our Microsoft products are 93 00:07:45.352 --> 00:07:45.972 Of course they are. 94 00:07:46.132 --> 00:07:46.972 Of course they were. 95 00:07:47.012 --> 00:08:06.318 What's so interesting to me about your career though is that at every sort of major turning point you kind of, you have turned towards technology as early as joining the Disney Imagineers which at the time was revolutionary and the work that they did has been so revolutionary. 96 00:08:07.747 --> 00:08:19.898 What do you think has been the core insight that you've taken at every turn, like from the Imagineers to, you know, you worked in like the golden era of TV. 97 00:08:19.918 --> 00:08:20.018 Yeah. 98 00:08:20.258 --> 00:08:27.144 But you also turned from that to go to Twitter, which must have been a bold move. 99 00:08:27.164 --> 00:08:29.726 Like there must have been sort of like a key insight that led there. 100 00:08:29.766 --> 00:08:33.129 But like what have been the key insights and those different pivot points? 101 00:08:33.329 --> 00:08:34.091 There's two answers. 102 00:08:34.391 --> 00:08:40.241 One is the real answer, which reveals me to be a terribly shallow person. 103 00:08:40.622 --> 00:08:43.888 Sometimes you look at something and you just go, I want in. 104 00:08:45.312 --> 00:08:51.555 And for me, growing up, I grew up on the south side of Chicago, and I watched way too much television. 105 00:08:51.795 --> 00:08:55.456 And all I thought to myself is, I want to be on the other side of that screen. 106 00:08:55.937 --> 00:08:56.577 I want in. 107 00:08:57.397 --> 00:08:59.398 So it's I want in there. 108 00:08:59.918 --> 00:09:04.400 It's I want in because there's some very funny people doing stuff here. 109 00:09:04.480 --> 00:09:05.541 And I like funny people. 110 00:09:05.561 --> 00:09:06.901 And I get along with them really well. 111 00:09:07.001 --> 00:09:07.562 I want in. 112 00:09:08.559 --> 00:09:19.105 The technology turn for me, again, I hear a modem, and this was in the days of like CompuServe, and there's no browser, there's no, it's just, 113 00:09:20.355 --> 00:09:25.838 But the first thing I joined was the Whole Earth Electronic Link, the well. 114 00:09:26.838 --> 00:09:35.963 And so the Twitter people, it was Chloe Sladden and Dick Costolo, and I met with them and I was like, yeah, these are my people. 115 00:09:36.703 --> 00:09:40.125 And I think they're doing something so new and different. 116 00:09:40.585 --> 00:09:41.966 It'll just be fun. 117 00:09:42.246 --> 00:09:43.267 It'll be really great. 118 00:09:43.887 --> 00:09:46.568 And so there was that. 119 00:09:46.848 --> 00:09:49.009 For me, it's a little harder right now. 120 00:09:49.569 --> 00:09:57.511 My old tribe is not thrilled with the prospect of what this is going to be or what it might mean. 121 00:09:58.412 --> 00:09:59.372 And yet there's a new 122 00:10:00.304 --> 00:10:04.586 group of people, which I want to be very careful, it's not defined by age. 123 00:10:05.066 --> 00:10:07.567 It's defined by a sensibility and a mindset. 124 00:10:08.048 --> 00:10:17.412 There's a new group of people who are storytellers and who are in what I still call show business because they don't think like David Ellison. 125 00:10:17.452 --> 00:10:17.572 They 126 00:10:19.593 --> 00:10:22.735 blanket and a garage and put on something. 127 00:10:23.535 --> 00:10:30.178 And they're beginning to see, okay, wait a minute, there's this thing, oh, what does C Dance do? 128 00:10:30.578 --> 00:10:33.520 Or how can I do this with Google Story Studio? 129 00:10:33.720 --> 00:10:36.341 Or how can I do, oh wait, this is really interesting. 130 00:10:36.821 --> 00:10:39.483 And then they begin kind of getting in and playing with it. 131 00:10:40.243 --> 00:10:47.787 And those people, to me, that's the group where I go like, okay, I want in, or at this point, I really do feel like, okay, 132 00:10:49.411 --> 00:10:51.673 I might be one of the first kind of settlers. 133 00:10:52.214 --> 00:10:59.140 If there was one key insight that you took away from working early on with like the Imagineers, like what might have that been? 134 00:11:00.261 --> 00:11:04.765 The Imagineers were founded by Disney to help him build stuff in the theme parks. 135 00:11:05.486 --> 00:11:14.434 And anytime new technology came around, the Imagineers, a group of the Imagineers would be assigned to figure out what does that mean to us? 136 00:11:16.381 --> 00:11:30.328 The most threatening thing in the late 90s for Disney was that technology was going to give the audience unprecedented access to IP. 137 00:11:31.368 --> 00:11:35.390 That you could suddenly digitize a copyright. 138 00:11:35.430 --> 00:11:41.153 You could take Snow White and the Seven Dwarves and let your imagination run free. 139 00:11:41.413 --> 00:11:42.554 And this is terrifying. 140 00:11:43.545 --> 00:11:54.348 At the same time, you could also see that the audience is going to be in this new place and they're going to want to have access to that stuff. 141 00:11:54.908 --> 00:11:59.889 So the insight was suddenly there was a brand new tension. 142 00:12:00.529 --> 00:12:03.290 There was a new tension between what 143 00:12:04.768 --> 00:12:21.006 the audience and creators would have access to and be able to make versus what you were able to make given the moat around your business and the barbarians were building bridges over the moat. 144 00:12:21.266 --> 00:12:21.627 Hmm. 145 00:12:22.826 --> 00:12:37.635 So fast forwarding to MTV, VH1, and the work that you did there, what do you think might have been your key takeaway then having worked in the golden era of television? 146 00:12:37.735 --> 00:12:37.995 Yeah. 147 00:12:38.015 --> 00:12:41.577 So the key takeaway for me, again, was open it up to the audience. 148 00:12:42.317 --> 00:12:44.779 And that's your survival. 149 00:12:45.079 --> 00:12:45.819 And what does that mean? 150 00:12:46.219 --> 00:12:47.060 So for example... 151 00:12:48.663 --> 00:12:52.244 I created a show called Best Week Ever, and it was doing very well. 152 00:12:52.785 --> 00:12:57.827 And we had created a Tumblr just for the producers to talk to each other. 153 00:12:58.827 --> 00:13:03.989 And we realized suddenly that there were 10 or 15,000 people reading our Tumblr. 154 00:13:06.497 --> 00:13:10.078 And it literally was just like, read the audience signal, see what's going on. 155 00:13:10.318 --> 00:13:11.538 And it was a complete accident. 156 00:13:11.618 --> 00:13:17.600 That's so interesting, because it then sounds like your move to Twitter was kind of like further embracing the audience, right? 157 00:13:17.620 --> 00:13:18.280 Like moving. 158 00:13:18.360 --> 00:13:24.162 It almost seems like at every point in your career, you've taken a step closer to the audience. 159 00:13:24.302 --> 00:13:25.382 I think that's really true. 160 00:13:25.662 --> 00:13:31.024 If you start out as a writer and you want to use AI, I have kind of a few things that I kind of say you should do. 161 00:13:31.764 --> 00:13:36.065 You should create a who I am as a writer document. 162 00:13:36.993 --> 00:13:43.481 And you should throw in your inspirations, what you love, what you read, what you aspire to do, all of that. 163 00:13:43.621 --> 00:13:47.506 And then you also have a document that this is the thing I'm working on. 164 00:13:48.265 --> 00:13:48.485 Right? 165 00:13:48.966 --> 00:13:53.189 And it could be... And by the way, you don't even need to be specific about... 166 00:13:53.249 --> 00:13:54.150 I have this idea. 167 00:13:54.650 --> 00:13:58.173 I have this thing that I really, really, really want to kind of get at. 168 00:13:58.734 --> 00:13:59.614 I want to think about. 169 00:14:00.095 --> 00:14:04.378 In the same kind of way that like Joan Didion once said, I don't know what I think about something until I write about it. 170 00:14:04.818 --> 00:14:06.119 I have this thing I want to explore. 171 00:14:06.220 --> 00:14:07.381 And AI is great at that. 172 00:14:07.961 --> 00:14:10.203 And then you go... You begin... 173 00:14:10.912 --> 00:14:12.894 putting as much as you know about that thing in. 174 00:14:13.394 --> 00:14:20.779 And the first thing you can say to AI is, does this idea and this person that I'm describing have anything to do with each other? 175 00:14:21.800 --> 00:14:22.060 Right? 176 00:14:22.260 --> 00:14:24.702 And then you kind of play around with it. 177 00:14:24.742 --> 00:14:27.684 And the third question is, who's out there? 178 00:14:28.940 --> 00:14:31.902 Is there an audience out there for this thing? 179 00:14:31.962 --> 00:14:33.283 What platforms are they on? 180 00:14:33.343 --> 00:14:34.684 How do they consume media? 181 00:14:35.005 --> 00:14:40.829 What are the questions they care about that relate to the thing I'm writing or even bigger questions that they care about? 182 00:14:40.969 --> 00:14:42.830 Somebody in the Atlantic did this, I think. 183 00:14:43.411 --> 00:14:45.772 They created the focus group for something. 184 00:14:46.033 --> 00:14:47.554 And you could actually literally do that. 185 00:14:47.594 --> 00:14:57.541 Like create a focus group of eight people from a range where there's a guy over here who's the guy you really wish you hadn't invited into the focus group because he's really negative and you have somebody who really loves what you're doing. 186 00:14:58.681 --> 00:15:04.344 and begin working on the idea that you're working on to get feedback. 187 00:15:04.584 --> 00:15:07.125 And so none of that is about AI writing for you. 188 00:15:07.985 --> 00:15:15.068 Everything is about having AI as kind of a virtual writer's room, as a group of intelligences that can help you think through things. 189 00:15:15.773 --> 00:15:19.915 Compound Writing, which is a great plugin that Katie Parrott on our phone bill. 190 00:15:19.935 --> 00:15:22.075 Yeah, give her all the credit she can get. 191 00:15:22.215 --> 00:15:24.716 Which I'm a big fan of Katie, she's fantastic. 192 00:15:24.996 --> 00:15:31.519 And that was built off of Kieran Klassen's Compound Engineering plugin. 193 00:15:32.239 --> 00:15:38.661 And what makes so much sense about those plugins is that they are built by experts in their domain, right? 194 00:15:38.761 --> 00:15:48.783 So the reason why I use compound engineering is because I'm not an engineer, and I trust Kieran as a really brilliant engineer and now actually a whole team that maintains that particular plugin. 195 00:15:49.363 --> 00:15:55.605 And so when I want to build a new product, I want to defer to experts in that field to... 196 00:15:56.525 --> 00:16:03.367 follow the steps that they would take in order to build a product so that I'm doing it safely and following best practices. 197 00:16:03.407 --> 00:16:04.467 The same for writing, right? 198 00:16:04.507 --> 00:16:10.788 Katie has codified at least her approach to writing, the kind of writing that she does. 199 00:16:11.449 --> 00:16:18.070 And one of the things that you've been exploring is how to codify writing in your capacity. 200 00:16:18.090 --> 00:16:19.490 Can you speak a little bit more about that? 201 00:16:19.510 --> 00:16:20.311 Yeah, yeah, yeah. 202 00:16:20.331 --> 00:16:21.091 So think about... 203 00:16:21.980 --> 00:16:31.251 There's a kind of core that exists, and what you pour into that core are two different things. 204 00:16:31.612 --> 00:16:41.363 One is expertise around, so in my case, it's expertise around creating entertainment product. 205 00:16:42.717 --> 00:16:48.922 And there's a fair amount of product design and product management that goes into this. 206 00:16:49.683 --> 00:16:56.209 And you get the expertise of everybody, and so much of this has been published, and so much of this is freely available. 207 00:16:56.669 --> 00:17:06.679 You could go into everybody from David O. Selznick to Francis Coppola to Quentin Tarantino, to all the way down the line. 208 00:17:07.139 --> 00:17:08.579 There's all the expertise. 209 00:17:08.620 --> 00:17:15.662 It goes around writing, it goes around producing, it goes around editing, it goes around marketing and distribution. 210 00:17:15.722 --> 00:17:20.823 There's even people who are just really brilliant, who are just like business operations and legal people. 211 00:17:21.083 --> 00:17:26.425 And there's this, you can put all of that, and I'm building that, where I'm kind of like going in and going like, 212 00:17:28.273 --> 00:17:32.397 You're going to make 1,001 decisions in the next week, probably. 213 00:17:32.877 --> 00:17:37.361 So half of this is, here's the experts, right? 214 00:17:37.481 --> 00:17:40.343 And that AI has ingested all this, thought about it. 215 00:17:40.744 --> 00:17:44.347 And if you're going, okay, the marketing people and the writers are disagreeing. 216 00:17:45.190 --> 00:17:45.410 Right? 217 00:17:46.831 --> 00:17:50.933 The writer thinks these characters are answering these questions and the marketing people say, I can't sell that. 218 00:17:51.713 --> 00:17:51.913 Right? 219 00:17:52.634 --> 00:17:57.536 Well, what are other ways people have solved those issues? 220 00:17:58.577 --> 00:18:10.623 Just to give you advice, just to suggest or kind of like help you mediate that, which by the way, also results in less friction because so much of this in show business gets personal, 221 00:18:11.782 --> 00:18:15.406 People get in a huff, they walk away, they have horrible fights. 222 00:18:15.686 --> 00:18:20.932 The best part about this is when the movie comes out or the TV series premieres, everybody's patting themselves on the back. 223 00:18:21.133 --> 00:18:22.755 Oh, we made it despite this. 224 00:18:23.155 --> 00:18:23.956 I'm like, stop. 225 00:18:24.376 --> 00:18:26.259 Why did you, you didn't have to do it in the first place. 226 00:18:26.719 --> 00:18:29.722 The other half though is also your personal culture. 227 00:18:30.283 --> 00:18:43.466 So if you're a production company or a studio, or you have a way of doing things and thinking about things, and you can put that into the hub of the AI central core as well. 228 00:18:44.306 --> 00:18:49.027 And everybody who's part of the process has already also said, here's a document about who I am. 229 00:18:49.407 --> 00:18:50.667 Here's what my career is all about. 230 00:18:50.687 --> 00:18:52.007 Here's the kind of way I work. 231 00:18:52.567 --> 00:18:56.268 And so now you've got this kind of really great Petri dish 232 00:18:57.462 --> 00:19:01.583 of cultural documents that AI can help you think through. 233 00:19:02.744 --> 00:19:10.966 And the last thing I'll say is, people in show business still make things the way Henry Ford made cars. 234 00:19:11.566 --> 00:19:12.767 There's this assembly line. 235 00:19:13.267 --> 00:19:15.388 And if you say to them, no, make it a circle. 236 00:19:15.708 --> 00:19:25.671 So by the way, the marketing and business people are talking to the writer and the pre-production people now before you start making decisions that get locked in or whatever. 237 00:19:26.925 --> 00:19:30.309 Sometimes the production company has got 10 people, 15 people. 238 00:19:30.750 --> 00:19:33.994 Sometimes it's a writer who at this point is the entire team. 239 00:19:34.314 --> 00:19:34.554 Right. 240 00:19:35.555 --> 00:19:45.207 So what I love about plugins in particular is that they demystify what is otherwise a really complicated process into simple steps, right? 241 00:19:46.368 --> 00:19:48.530 it's the only thing that AI can do really well. 242 00:19:48.570 --> 00:19:50.631 It's like follow directions clearly. 243 00:19:51.712 --> 00:19:56.595 And I find at least AI to be as powerful as the instructions are. 244 00:19:57.215 --> 00:20:08.823 So what I love about, again, like the, the engineering tools that, or the plugins that our team has put out is that when it comes to engineering, I understand that there's basically like three steps to engineering. 245 00:20:09.564 --> 00:20:14.687 And what Kieran has done is just kind of like codified what excellence looks like in those three steps. 246 00:20:14.747 --> 00:20:15.728 First you make a plan and, 247 00:20:16.568 --> 00:20:19.411 There's all kinds of things involved with making a good plan. 248 00:20:19.651 --> 00:20:21.692 So we'll follow the instructions on that. 249 00:20:22.213 --> 00:20:29.339 Then we follow that plan and build something, you know, based on kind of like that PRD or that kind of set of requirements. 250 00:20:29.839 --> 00:20:31.821 And then we'll review that plan and kind of like build on it. 251 00:20:31.841 --> 00:20:33.422 It's kind of like that loop that you discussed. 252 00:20:34.363 --> 00:20:40.568 What in its most basic form do you think is that loop for writing or producing in your case? 253 00:20:40.588 --> 00:20:40.868 Yeah. 254 00:20:41.238 --> 00:20:43.499 We'll break it into three steps. 255 00:20:43.819 --> 00:20:51.563 There's ideation, and we'll call it architecture or design. 256 00:20:52.613 --> 00:20:55.937 It is the thing that produces a script, right? 257 00:20:56.097 --> 00:21:08.771 And that's the first thing, is you get something that is a written document that is the best embodiment of the idea you're trying to work on. 258 00:21:09.392 --> 00:21:12.555 The next phase is production. 259 00:21:13.675 --> 00:21:17.698 And it's everything from how do we prepare to make this thing? 260 00:21:18.539 --> 00:21:28.988 And it's a very similar process whether you're working with a straight-on regular kind of movie or TV series or a hybrid series that uses some AI. 261 00:21:29.108 --> 00:21:36.774 Or even in all, I'm working with a studio now, and we're making an AI animated series. 262 00:21:37.034 --> 00:21:39.717 And the way we're doing it is they've hired a bunch of writers. 263 00:21:40.417 --> 00:21:41.397 to create something. 264 00:21:41.437 --> 00:21:45.798 There's an animation studio in India that's doing the animation. 265 00:21:45.938 --> 00:21:47.239 Some of it's with AI, some of it isn't. 266 00:21:47.559 --> 00:21:50.019 And then there's a group of computer scientists in Zurich. 267 00:21:50.540 --> 00:21:52.180 I don't know why they're in Zurich. 268 00:21:52.380 --> 00:21:53.000 It sounds... 269 00:21:53.200 --> 00:21:59.442 I keep imagining a Goldfinger's lair up in Zurich and they're like, we're watching. 270 00:22:00.122 --> 00:22:04.183 And they're trying to figure out how do you build AI software that would keep the people with it. 271 00:22:05.003 --> 00:22:13.020 So there's this production process that's everything from pre-production considerations to keeping everybody going within the production. 272 00:22:13.040 --> 00:22:16.126 And I would include editing in there as well. 273 00:22:17.533 --> 00:22:18.693 And there's a lot of stuff in there. 274 00:22:18.713 --> 00:22:20.254 There's storyboards. 275 00:22:21.074 --> 00:22:30.258 It's fascinating to me that Martin Scorsese is one of the first people to kind of come out and go, oh, I'm working with this company that's creating storyboards with AI. 276 00:22:30.498 --> 00:22:31.778 And people freaked out about it. 277 00:22:31.798 --> 00:22:33.159 Like, well, what about the storyboard artists? 278 00:22:33.579 --> 00:22:35.520 And I think Scorsese, I don't know. 279 00:22:35.680 --> 00:22:36.420 Scorsese is like... 280 00:22:36.880 --> 00:22:39.403 You're going to have to get another job. 281 00:22:39.463 --> 00:22:40.263 It's the mob. 282 00:22:42.245 --> 00:22:45.448 So there's the production phase, and AI can help with that. 283 00:22:45.769 --> 00:22:49.172 And then there's the how do you reach the audience with this thing phase. 284 00:22:49.212 --> 00:22:51.194 How do you market and distribute this? 285 00:22:51.894 --> 00:22:57.499 What's fascinating to me about having worked with a couple of production studios in Hollywood 286 00:22:58.260 --> 00:23:04.802 is the marketing people are all in on figuring out how to use AI. 287 00:23:05.362 --> 00:23:12.245 To the point where they will create a campaign for something, and they work with companies that are doing it with AI. 288 00:23:13.045 --> 00:23:23.609 They will watch intent to purchase a ticket or ticket purchases based on a trailer, and they will rewrite the trailer almost on an hour by hour basis. 289 00:23:24.029 --> 00:23:25.569 You go, OK, let's A-B test this. 290 00:23:25.609 --> 00:23:26.270 Let's try this. 291 00:23:26.290 --> 00:23:26.950 Let's try that. 292 00:23:28.042 --> 00:23:31.644 And that, again, AI can be part of that process. 293 00:23:31.684 --> 00:23:41.790 What I think is fascinating is, again, you make it a circle so that here's the ideation and design and architecture phase, and it works. 294 00:23:42.170 --> 00:23:47.333 Why not have the marketing people right away going, well, let's create a proxy trailer for that. 295 00:23:47.793 --> 00:23:51.575 Let's see if we can sell tickets to that even though it doesn't exist yet, right? 296 00:23:51.595 --> 00:23:54.057 Which is essentially what show business has always been about. 297 00:23:54.917 --> 00:24:05.703 So those are the three phases, and that's where I think in every step, AI can speak to the people who are doing it and can inform the decisions that everybody's making. 298 00:24:06.264 --> 00:24:07.525 So let me make sure I got that. 299 00:24:07.585 --> 00:24:17.830 So it's ideation, and then it's basically the production, which includes a bunch of things, and then there's the marketing, like the connecting to the audience in a loop. 300 00:24:18.211 --> 00:24:18.391 Yeah. 301 00:24:19.191 --> 00:24:19.992 How do you... 302 00:24:21.024 --> 00:24:22.849 Thank you for, by the way, cutting it down to three. 303 00:24:22.949 --> 00:24:25.695 I've never done it before, and that was very helpful. 304 00:24:25.756 --> 00:24:26.297 Okay. 305 00:24:28.906 --> 00:24:30.387 I mean, this is so helpful to me. 306 00:24:30.407 --> 00:24:41.132 I'm the reason why I enjoy using AI and why I got into using AI is because, you know, Anthropic describes themselves as like AI for problem solvers. 307 00:24:41.172 --> 00:24:44.134 And I think it's such a smart description because I think it's totally right. 308 00:24:44.234 --> 00:24:49.397 You know, like for ambitious people, like you just want to do great work. 309 00:24:50.097 --> 00:24:53.279 And AI at its best, I find, allows me to do great work. 310 00:24:53.299 --> 00:24:53.439 Mm-hmm. 311 00:24:54.199 --> 00:25:09.008 But also when I'm getting creative and doing work that is fun and not in my line of expertise, it's easy to make AI slop, which is where these plugins are such helpful boundaries. 312 00:25:09.028 --> 00:25:16.213 They're sort of like the rails in a bowling alley that keep you from going off into the dark side of AI slop. 313 00:25:16.953 --> 00:25:32.764 How is it that you find these three steps might help someone who is not familiar with using AI in this kind of like writing capacity or actually someone who is an expert in sort of like being a writer or a producer? 314 00:25:32.964 --> 00:25:41.891 How would these three steps help them, you know, build even more beautiful, valuable work that is not AI slop? 315 00:25:41.911 --> 00:25:42.011 Yeah. 316 00:25:42.493 --> 00:25:49.503 The first match you've thrown into the gas can here are the words AI slop, right? 317 00:25:49.984 --> 00:25:54.871 It is the thing that terrifies everyone, right? 318 00:25:55.913 --> 00:25:56.994 And again, I go back to my... 319 00:25:57.555 --> 00:26:02.337 my kind of show business metaphor or framework, right? 320 00:26:02.817 --> 00:26:12.422 If all you need is a sheet and a garage and some chairs to put on a show, why doesn't everybody put on a show? 321 00:26:12.562 --> 00:26:15.443 Well, first thing, not everybody's gonna put on a show. 322 00:26:16.784 --> 00:26:24.147 Annie Leibovitz once said, everybody has a camera that's better than the camera that I used for 30 years in their pocket. 323 00:26:25.072 --> 00:26:29.695 But there are very, very few people who are taking pictures like I take. 324 00:26:30.376 --> 00:26:40.183 So the first step here is in defining who you are and what you're trying to do and thinking deeply about them. 325 00:26:40.423 --> 00:26:45.147 And in fact, I say in all three steps, you want to begin by thinking deeply. 326 00:26:46.027 --> 00:26:51.831 There was a book that was very popular for a long time, and you don't hear much about it, called Flow. 327 00:26:51.851 --> 00:26:51.952 Flow. 328 00:26:53.073 --> 00:26:59.182 And it was all about what it is to be in that state where you're composing or writing or painting or whatever. 329 00:27:00.044 --> 00:27:03.849 And you're just locked in, you're in the zone. 330 00:27:04.190 --> 00:27:08.155 For me, in all three of these steps, if you haven't begun, 331 00:27:09.282 --> 00:27:13.765 By going there first, you're not going to get a good product. 332 00:27:14.465 --> 00:27:27.873 You have to start with uniquely you, what you know, spend a day or two dumping everything you know about what you're going to do onto a page, right? 333 00:27:28.434 --> 00:27:33.297 Do it by hand, do it by whisper, do it by whatever, but empty out. 334 00:27:34.397 --> 00:27:38.560 You will never have that opportunity again once you open AI. 335 00:27:40.177 --> 00:27:48.719 There's one chance to kind of be yourself and confront the idea or confront, how am I going to produce this really, really well? 336 00:27:49.180 --> 00:27:51.040 Or confront, how will we reach the audience? 337 00:27:51.060 --> 00:27:55.461 And all three of these are really important expertise and very, I don't think anybody has all three. 338 00:27:56.862 --> 00:28:01.943 Maybe, no, I don't think anybody, I think people, any single person knows how to find all three people. 339 00:28:02.564 --> 00:28:06.905 But you have to empty out, you have to do this yourself as a human being. 340 00:28:08.048 --> 00:28:11.670 then the nice part is it's a very lonely business. 341 00:28:12.311 --> 00:28:15.292 And for all of us, it's like, well, great, I can throw all this stuff out. 342 00:28:15.312 --> 00:28:16.733 What am I going to do with it once I do it? 343 00:28:17.013 --> 00:28:18.274 Ah, now there's AI. 344 00:28:18.894 --> 00:28:25.158 Now there's Claude or ChatGPT, and we can have this discussion about which one I like or anybody likes. 345 00:28:26.239 --> 00:28:28.540 But you go in and you say, this is who I am. 346 00:28:28.780 --> 00:28:29.881 This is what I'm doing. 347 00:28:30.701 --> 00:28:32.082 Let's talk about this. 348 00:28:33.063 --> 00:28:36.305 One of the things that I love that I actually did very early on 349 00:28:37.705 --> 00:28:42.929 I asked Claude, I'm a writer, I'm a producer, I'm trying to do these things. 350 00:28:43.369 --> 00:28:44.370 How can you help me? 351 00:28:45.111 --> 00:28:46.232 What can you do for me? 352 00:28:46.392 --> 00:28:47.873 Oh, should I be using Claude Code? 353 00:28:48.213 --> 00:28:51.436 Should I be using, back when Cowork was a separate thing, should I be using Cowork? 354 00:28:51.876 --> 00:28:53.037 I read about these skills. 355 00:28:54.318 --> 00:28:58.661 What about, you know, and there'll be people on Substack publishing crazy stuff. 356 00:28:58.981 --> 00:29:01.363 And by the way, you don't need those subscriptions. 357 00:29:01.383 --> 00:29:02.164 You can have everything. 358 00:29:04.006 --> 00:29:07.088 But there's people publishing all this stuff, and I will feed it into Claude. 359 00:29:07.508 --> 00:29:08.769 And I'll say, is this useful? 360 00:29:09.209 --> 00:29:12.231 And Claude might say, no, we're already doing that. 361 00:29:12.431 --> 00:29:13.933 Or yes, I can build that with you. 362 00:29:15.093 --> 00:29:16.935 And so I built a process. 363 00:29:18.122 --> 00:29:26.326 that began with myself and then began with almost conversation or relationship with the AI that helped me do it better. 364 00:29:27.386 --> 00:29:34.049 And so in each of those steps, so if you have, I have many friends who are producers, and that's a tough process. 365 00:29:34.149 --> 00:29:34.509 job. 366 00:29:34.710 --> 00:29:36.551 You are the drill sergeant. 367 00:29:36.671 --> 00:29:40.294 You're making sure things get done faster, better, and cheaper. 368 00:29:42.216 --> 00:29:44.838 And you can keep your schedules. 369 00:29:44.858 --> 00:29:46.580 You can keep people in touch with each other. 370 00:29:46.620 --> 00:29:48.081 You can find tax incentives. 371 00:29:48.121 --> 00:29:50.123 You can find what the union is requiring. 372 00:29:50.143 --> 00:29:52.905 You can find all of these things in one tool. 373 00:29:53.906 --> 00:29:55.328 And the same with marketing. 374 00:29:55.728 --> 00:29:57.389 You've got these tools 375 00:29:58.290 --> 00:30:07.763 Within Claude, specifically Claude and Chat GPT now, it almost feels like everything for everybody, all the time, everywhere. 376 00:30:08.464 --> 00:30:09.545 It's starting to get there. 377 00:30:10.206 --> 00:30:12.069 The question is, can it do... 378 00:30:13.048 --> 00:30:19.491 your version of everything all the time for you when you need to have it done. 379 00:30:20.052 --> 00:30:21.973 And that's where the real skill is. 380 00:30:22.073 --> 00:30:24.674 That's where you kind of go in and go, this is my culture. 381 00:30:24.714 --> 00:30:25.915 This is what I care about. 382 00:30:26.315 --> 00:30:27.596 This is what the priorities are. 383 00:30:27.736 --> 00:30:28.676 That's so interesting. 384 00:30:28.716 --> 00:30:38.441 So it sounds like you're saying before you can really get any value out of these tools, you really need to not touch these tools and really go inward, like really understand. 385 00:30:38.461 --> 00:30:39.902 I really firmly believe that. 386 00:30:39.982 --> 00:30:40.562 And I don't know. 387 00:30:40.582 --> 00:30:40.882 I... 388 00:30:42.023 --> 00:30:46.399 I firmly, firmly believe that it could be that's how I started. 389 00:30:47.495 --> 00:30:57.018 Back in the day, if you wanted to convince somebody to do something, you'd write something and you'd pitch them. 390 00:30:57.518 --> 00:31:01.639 And it was everything that you knew about the thing you were trying to do. 391 00:31:02.059 --> 00:31:07.701 And you would edit that, you'd figure it out, you'd go, I've only got 120 seconds with this executive. 392 00:31:08.201 --> 00:31:10.061 Or it's a five minute meeting. 393 00:31:10.521 --> 00:31:16.023 And the minute they start talking about their weekend in the palms, you're sunk. 394 00:31:16.363 --> 00:31:16.583 Right? 395 00:31:17.044 --> 00:31:18.645 But, oh, now it's down to two minutes. 396 00:31:19.806 --> 00:31:20.927 That's how I began. 397 00:31:21.167 --> 00:31:26.571 And I still believe after a lot of, actually, I'll give you an example. 398 00:31:26.611 --> 00:31:26.751 Yeah. 399 00:31:26.891 --> 00:31:30.414 After a lot of experiments, I still believe that that's the way to do it. 400 00:31:30.854 --> 00:31:44.205 I think I was using compound writing and then I was using Claude Code to try and figure out a way to, in kind of five bullet points, hammer away at any idea I might have. 401 00:31:45.103 --> 00:31:48.790 And so I gave it three or four ideas and it was, we were refining it. 402 00:31:48.810 --> 00:31:49.672 Well, that's not quite right. 403 00:31:50.092 --> 00:31:54.741 And then Claude said, let me come up with an idea and I'll test it against myself. 404 00:31:56.865 --> 00:31:57.907 Oh, okay, Claude. 405 00:31:59.818 --> 00:32:17.661 So the idea it came up with was a series, a web series, that would be like Brooklyn Coffee Shop, but it was a group of 60 and 70-somethings who were running the Facebook page for a senior citizen home. 406 00:32:19.483 --> 00:32:26.154 And it gave all the pros and cons, how it would be produced, who the audience is, how you'd reach them. 407 00:32:26.675 --> 00:32:31.061 And I'm sitting there reading it and going, okay, that's not a terrible idea. 408 00:32:34.021 --> 00:32:34.501 Wow. 409 00:32:35.802 --> 00:32:37.604 And so then I was like, tell me more. 410 00:32:39.225 --> 00:32:40.606 What else do you know about this? 411 00:32:40.886 --> 00:32:42.468 And it was kind of churning things out. 412 00:32:44.309 --> 00:32:48.693 I only feel that the reason I bring it up is I think Claude has learned from me. 413 00:32:49.333 --> 00:32:53.737 And it's kind of like, yeah, let me dump everything I know about the 60 plus audience. 414 00:32:54.477 --> 00:32:56.717 and vertical series like Brooklyn Coffee Shop. 415 00:32:57.318 --> 00:33:01.518 And it made that great, I think, a great Facebook connection, which is where that audience is. 416 00:33:01.778 --> 00:33:02.098 Yeah. 417 00:33:02.319 --> 00:33:04.039 And I kind of like, oh, okay, great. 418 00:33:04.359 --> 00:33:07.920 But I feel like it's because it learned kind of that piece of me. 419 00:33:08.080 --> 00:33:08.600 And who you are. 420 00:33:08.640 --> 00:33:09.660 Yeah, that's interesting. 421 00:33:09.680 --> 00:33:20.562 I mean, one of the sort of like key questions I would certainly have is for like the next generation of producers and writers, how could they have a shot at being the next producer? 422 00:33:20.602 --> 00:33:24.483 Fred Graver because the TV industry doesn't exist in the way that it did. 423 00:33:24.784 --> 00:33:45.591 And it sounds like maybe what you're saying is so long as you give yourself the space to go inwards and build your own compass and understand who you are and what you want and what it is that you're trying to say at any given moment in time, you can have AI work for you in the different capacities that you would usually need a team to. 424 00:33:47.112 --> 00:33:48.712 instead of working for AI, right? 425 00:33:48.873 --> 00:33:51.473 It's like if you have a compass, you could have it work for you. 426 00:33:51.513 --> 00:33:56.655 If you don't have a compass, you're just kind of following Claude in whatever direction it's taking you in. 427 00:33:56.895 --> 00:34:08.498 Which is essentially how you have a career in, you know, the minute you say media, I just keep thinking David Zaslav and people who are just like trying to control the vertical. 428 00:34:08.518 --> 00:34:13.619 You know, to have a career as a storyteller, that's the test. 429 00:34:14.039 --> 00:34:15.600 Do you have something to say? 430 00:34:16.597 --> 00:34:20.859 And what's great about AI tools is it can help you say it. 431 00:34:21.539 --> 00:34:24.000 There are so many ways to help you say it. 432 00:34:26.461 --> 00:34:37.225 But it can't give you, if you ask it to tell you what it is you should be saying, you'll get, it goes to the mean. 433 00:34:37.885 --> 00:34:38.926 It won't do the thing. 434 00:34:39.566 --> 00:34:41.807 And some people don't have that thing. 435 00:34:42.047 --> 00:34:45.068 I hate to be a jerk about it, but some people just don't have it. 436 00:34:45.848 --> 00:34:49.910 People send their kids to my wife. 437 00:34:49.950 --> 00:34:51.330 I met my wife at Letterman. 438 00:34:51.791 --> 00:34:53.091 And so they come all the time. 439 00:34:54.492 --> 00:34:56.332 Once a month, somebody shows up for coffee. 440 00:34:56.792 --> 00:34:58.693 And they're like, I want to be a writer. 441 00:34:59.534 --> 00:35:05.076 And I will always say, and I'm a jerk about it, hey, tell me about the last thing you wrote. 442 00:35:06.276 --> 00:35:13.019 If it was more than two weeks ago or a month ago, I kind of go, is there something else you like to do? 443 00:35:14.492 --> 00:35:15.353 This is hard. 444 00:35:15.934 --> 00:35:16.835 It's lonely. 445 00:35:17.376 --> 00:35:18.918 It's full of rejection. 446 00:35:20.500 --> 00:35:31.113 And if you're not addicted to the thing, if you don't have to do this, and if you don't have something unique that is just burning there, there might be other places where you can go. 447 00:35:33.105 --> 00:35:48.675 Where the hard problem is right now, and this is where I stub my toe all the time, I believe that that is, for the first two pieces of the circle, I think all that is coming into place. 448 00:35:49.989 --> 00:35:52.731 It's the third piece that's really, really hard. 449 00:35:53.571 --> 00:35:57.834 I was really lucky to come up in the second half of the 20th century. 450 00:35:57.974 --> 00:35:59.795 It was, as you said, the golden days. 451 00:36:00.415 --> 00:36:08.200 You got development deals to have an office for a network with the promise that you would write two things. 452 00:36:09.481 --> 00:36:10.641 Oh, my God. 453 00:36:11.702 --> 00:36:15.104 And they flew you first class, and it was glorious. 454 00:36:17.807 --> 00:36:18.988 That's gone. 455 00:36:19.609 --> 00:36:23.131 As Bruce Springsteen would say, boys, those jobs ain't coming back. 456 00:36:24.252 --> 00:36:27.894 And so, I mean, maybe this takes me to your thesis, right? 457 00:36:27.914 --> 00:36:29.315 So you shared your thesis. 458 00:36:29.335 --> 00:36:32.897 We're hosting a conference in a few weeks called Thesis. 459 00:36:33.037 --> 00:36:35.118 It's about what happens after automation. 460 00:36:35.219 --> 00:36:40.342 And you wrote your perspective on like what your thesis is of what happens, what comes next. 461 00:36:41.499 --> 00:36:49.807 Yeah, so essentially I was saying those first two pieces, the new writer's room, will include AI, right? 462 00:36:50.107 --> 00:36:59.316 I remember the first time I walked into a writer's room and they had a Google Doc up on a screen that was capturing what everybody was talking about and helping to write. 463 00:37:00.237 --> 00:37:00.777 They were trying to... 464 00:37:02.198 --> 00:37:04.240 outline a story for an episode. 465 00:37:04.540 --> 00:37:06.462 And I was like, oh my God, that's so great. 466 00:37:06.482 --> 00:37:10.766 Because you were always before then going like, Kevin, did you get that? 467 00:37:11.186 --> 00:37:12.647 Did you read it back to us? 468 00:37:12.887 --> 00:37:13.728 Make sure you got it right. 469 00:37:13.988 --> 00:37:15.810 Now, it was there in front of everybody. 470 00:37:16.550 --> 00:37:25.278 I think in that same way, you will have AI in there and they will be asking questions like, OK, this is the season arc. 471 00:37:25.798 --> 00:37:26.378 We're here. 472 00:37:27.720 --> 00:37:28.841 What happens if... 473 00:37:30.001 --> 00:37:42.706 this character is doing that, or even more to the point, right, so there's a micro example, and there's a macro example, which is, okay, we're looking at successful TikTok series, 474 00:37:43.768 --> 00:37:48.211 that are, we'll go back to, we'll help Claude with his idea. 475 00:37:49.872 --> 00:37:51.693 The audience is 55 and older. 476 00:37:52.473 --> 00:37:56.616 I need to understand the audience, where they are, what their concerns are. 477 00:37:56.996 --> 00:38:00.498 If it was set here, what could I dive into? 478 00:38:00.938 --> 00:38:03.039 I need characters that reflect their problems. 479 00:38:03.320 --> 00:38:04.940 Let's talk about that stuff. 480 00:38:05.601 --> 00:38:06.041 And, and, 481 00:38:07.682 --> 00:38:09.805 There's one that you could do it singularly. 482 00:38:09.905 --> 00:38:15.011 I think it's much better if you've suddenly got three people with different perspectives. 483 00:38:15.771 --> 00:38:21.358 And by the way, another way to remove friction in the production process, everyone can hate the AI. 484 00:38:22.912 --> 00:38:25.553 So it's kind of like, I love what you're saying, Natalia. 485 00:38:26.353 --> 00:38:27.413 Claude's an idiot. 486 00:38:28.874 --> 00:38:30.114 But you're working everything out. 487 00:38:30.194 --> 00:38:35.575 And you also have this, you really have an expansive intelligence that is thinking about things. 488 00:38:35.815 --> 00:38:39.136 If you've set it up right, it's thinking about things that you would not have thought of. 489 00:38:39.776 --> 00:38:41.657 And so that has to be in there. 490 00:38:41.837 --> 00:38:42.957 And you have to be pushing it. 491 00:38:43.097 --> 00:38:45.298 And I think that becomes part of the process. 492 00:38:45.858 --> 00:38:46.939 Yeah. 493 00:38:46.959 --> 00:38:49.000 Do you happen to remember your thesis? 494 00:38:49.080 --> 00:38:52.181 I'm curious if you could maybe state a version of it so we have it. 495 00:38:52.201 --> 00:38:53.762 Yeah, I think I can. 496 00:38:53.922 --> 00:39:02.426 And then, by the way, we could also talk about how I used the dev edit to create the original version, and then Georgia had to keep paring it down to something. 497 00:39:02.446 --> 00:39:09.110 And what she ended up paring it down to was supporting the show business truism, 498 00:39:09.950 --> 00:39:14.712 I may be the only person on the Every Podcast ever who uses the term show business all the time. 499 00:39:15.112 --> 00:39:23.356 I'm realizing, I'm realizing, old man Graver shows up and talks about show business. 500 00:39:26.237 --> 00:39:34.661 So the truism is success is knowing what the audience wants before the audience knows it. 501 00:39:37.973 --> 00:39:39.374 AI is really helpful. 502 00:39:39.494 --> 00:39:41.336 So go back to my old man series. 503 00:39:41.856 --> 00:39:46.499 It could tell me right now a whole bunch of stuff that would slot right into that. 504 00:39:47.860 --> 00:39:56.387 What it wouldn't tell me is the personal quirks that would make these characters endearing. 505 00:39:56.907 --> 00:39:57.127 Right? 506 00:39:58.508 --> 00:40:06.554 For example, the guy who did the voices, who does the voice of Homer Simpson, he improvised. 507 00:40:06.935 --> 00:40:07.215 Don't. 508 00:40:09.645 --> 00:40:10.585 To me, that's the whole story. 509 00:40:10.845 --> 00:40:13.846 It's like, OK, we knew the whole thing, and then he improvised that. 510 00:40:13.926 --> 00:40:16.747 So AI can help you do a bunch of stuff. 511 00:40:16.787 --> 00:40:21.168 And then there's these kind of what is inside of you that nobody else has. 512 00:40:21.788 --> 00:40:25.729 And if there's two or three people or more in a writer's room, it's fascinating. 513 00:40:25.849 --> 00:40:27.130 It's what's inside all of you. 514 00:40:27.230 --> 00:40:28.630 OK, there's a guy over there. 515 00:40:28.670 --> 00:40:29.370 He went to Harvard. 516 00:40:29.570 --> 00:40:30.291 We had to take him. 517 00:40:32.111 --> 00:40:37.294 And then there's a couple of other people, and you've got an improv background, or you've got this background. 518 00:40:38.014 --> 00:40:46.378 Or sometimes, I know I've been in some writers' rooms where you literally brought somebody in who just knows the setting of the show. 519 00:40:46.678 --> 00:40:55.382 I worked on an ill-fated show for the producer Jim Brooks that was about a 14-year-old tennis phenom. 520 00:40:55.442 --> 00:40:56.803 In fact, the show was called Phenom. 521 00:40:57.404 --> 00:41:06.314 And we had in the writer's room somebody who coached teenagers into their tennis championships. 522 00:41:06.454 --> 00:41:11.220 And he was great at kind of going like, oh, I do something horrible there. 523 00:41:11.240 --> 00:41:11.601 Yeah. 524 00:41:13.143 --> 00:41:13.364 Right? 525 00:41:13.564 --> 00:41:14.705 And it was always kind of helpful. 526 00:41:14.945 --> 00:41:20.469 So you've got this writer's room, you've got AI, and that's how it comes together. 527 00:41:20.709 --> 00:41:21.089 Yeah. 528 00:41:21.590 --> 00:41:33.698 I mean, it kind of feels to me like a full circle, you know, that your thesis is essentially you have to know what your audience wants before the audience knows. 529 00:41:33.718 --> 00:41:35.400 And going back to... 530 00:41:37.723 --> 00:41:41.524 being a kid in the South side of Chicago on the other side of the TV. 531 00:41:41.544 --> 00:41:51.667 It's like every step you've taken in your career has been to be on the other side of the television and to close the gap between being the audience and being the producer. 532 00:41:52.167 --> 00:41:56.928 And now it's like we're at the singularity where you are the audience and the producer. 533 00:41:56.948 --> 00:42:04.710 So you have the opportunity to like have that core insight but also to do all of the work to create the show that you think could kind of 534 00:42:05.350 --> 00:42:10.053 produce that whole insight into a story that is entertaining and would connect to a broader audience? 535 00:42:11.635 --> 00:42:14.016 I had never thought of it as the singularity. 536 00:42:15.097 --> 00:42:16.538 It's the Fred singularity. 537 00:42:17.158 --> 00:42:24.003 Yeah, I have a very good friend, Danny Hillis, who studied AI with Marvin Minsky in the 80s. 538 00:42:24.103 --> 00:42:30.948 And yeah, and he constantly jokes that, yeah, that I should just get the brain implant. 539 00:42:30.968 --> 00:42:31.068 Yeah. 540 00:42:31.668 --> 00:42:31.988 now. 541 00:42:34.549 --> 00:42:38.892 Which, by the way, if I can, I will. 542 00:42:39.212 --> 00:42:40.933 That's awesome. 543 00:42:41.213 --> 00:42:42.573 Yeah, I think that's true. 544 00:42:43.074 --> 00:42:45.435 And I think, again, the hard part right now... 545 00:42:46.946 --> 00:42:57.390 is for those of you who are business people and are watching, it's that third piece here where Kevin Kelly has this great thousand true fans theory, right? 546 00:42:57.450 --> 00:42:58.991 You're a musician, you're a photographer, whatever. 547 00:42:59.351 --> 00:43:08.374 If you could just get a thousand people to give you $10 a month for a bunch of stuff that you do online, you can make a living. 548 00:43:09.054 --> 00:43:13.056 And nobody has set that up yet for people who 549 00:43:14.018 --> 00:43:20.026 do what I do who no longer have the wonderful golden age of television. 550 00:43:20.146 --> 00:43:23.130 And I feel like that's what we finally need. 551 00:43:23.150 --> 00:43:23.571 If you're able to, 552 00:43:25.358 --> 00:43:34.368 create a really, really funny animated sitcom, and Fox is only buying one a year, where do the other 50 good ideas go? 553 00:43:34.388 --> 00:43:39.094 And I keep looking, I point as if it's an actual place here. 554 00:43:39.334 --> 00:43:45.041 I keep looking at that third sector and going, somehow we need that to happen. 555 00:43:45.990 --> 00:43:48.172 So you've done that emptying out process. 556 00:43:48.632 --> 00:43:49.453 What happens to us? 557 00:43:49.473 --> 00:43:52.315 You're sitting on your computer, you have all of these notes. 558 00:43:52.655 --> 00:43:53.256 What do you do? 559 00:43:53.856 --> 00:44:02.183 I have a couple of skills that I've built that are my readers, right? 560 00:44:02.423 --> 00:44:04.745 And they're the famous kind of readers over your shoulder. 561 00:44:04.765 --> 00:44:08.207 There's people that I think think like me or would look at this. 562 00:44:08.888 --> 00:44:12.150 I've lived in Claude for a long time, so it's the room I keep going back to. 563 00:44:12.771 --> 00:44:15.153 And I will say, Claude, this is this, 564 00:44:16.345 --> 00:44:18.307 Here's a folder of all my work. 565 00:44:18.327 --> 00:44:19.047 You know what I work on. 566 00:44:19.087 --> 00:44:19.828 This is what I'm going to do. 567 00:44:20.268 --> 00:44:22.670 Let's talk about this. 568 00:44:23.350 --> 00:44:23.611 Right? 569 00:44:23.711 --> 00:44:36.641 And I think it's, and then I'll say what I, you know, because prompting, it's kind of like, what I need back from you is an analysis of, does this sound like me? 570 00:44:36.801 --> 00:44:38.002 Like something I would work on? 571 00:44:39.723 --> 00:44:44.887 Is there an audience for this on some platform somewhere that feels real? 572 00:44:45.768 --> 00:44:45.928 And, 573 00:44:46.395 --> 00:44:54.740 What will I need to get to a finished product, a script, or to get it to screen that I don't have now? 574 00:44:56.211 --> 00:45:00.474 And the reason I asked that last question is, do I really want to go down the road? 575 00:45:00.654 --> 00:45:03.056 I have a lot of, everybody has a lot of ideas, right? 576 00:45:03.076 --> 00:45:10.821 But it's like, do you really want to go down the hard road of writing all of this to get to, to find people to help you make it or whatever? 577 00:45:11.481 --> 00:45:17.065 And I will have a very long conversation with Claude about the reality of that idea. 578 00:45:17.445 --> 00:45:19.186 So let's say we get to it and like, okay, great. 579 00:45:19.246 --> 00:45:19.566 You know what? 580 00:45:19.626 --> 00:45:20.407 I'm working on this. 581 00:45:21.372 --> 00:45:29.555 And I know kind of my steps for subsequent drafts and all of that, where I'm going to need to go from A to B. 582 00:45:31.376 --> 00:45:33.937 And I will tell Claude, here's what we're going to do. 583 00:45:34.077 --> 00:45:35.598 And here's what I think the schedule is. 584 00:45:35.638 --> 00:45:37.038 And here's when I'm going to deliver this. 585 00:45:37.719 --> 00:45:40.640 I have a Claude agent that is my morning check-in. 586 00:45:41.480 --> 00:45:46.926 And he's very unpleasant, and I've asked him to be very unpleasant. 587 00:45:47.626 --> 00:45:56.515 And so when I opened up the computer again that night to say I've got another idea, the first thing in the morning I will get back is, wait a minute. 588 00:45:58.070 --> 00:45:59.371 You're working on this. 589 00:46:00.332 --> 00:46:02.133 You told me you wanted me to talk about this. 590 00:46:02.373 --> 00:46:04.515 And then in the night you came in with this other thing. 591 00:46:04.915 --> 00:46:05.756 What are we doing? 592 00:46:06.496 --> 00:46:06.676 Right? 593 00:46:06.896 --> 00:46:11.680 Because you owe this, you owe your thesis statement to every tomorrow afternoon. 594 00:46:12.741 --> 00:46:12.981 Right? 595 00:46:14.982 --> 00:46:16.263 That's kind of a practical thing. 596 00:46:16.283 --> 00:46:18.165 But then the other thing is like it knows... 597 00:46:19.598 --> 00:46:21.419 It knows what frameworks I use. 598 00:46:21.580 --> 00:46:26.683 I love the Ken Adams because therefore framework. 599 00:46:26.723 --> 00:46:28.865 It's what Pixar uses in a lot of their stuff. 600 00:46:28.905 --> 00:46:31.067 It's the simplest way to do anything. 601 00:46:31.667 --> 00:46:33.448 And I'll go, here's my story. 602 00:46:33.468 --> 00:46:34.469 Here's what I think I'm doing. 603 00:46:35.070 --> 00:46:38.312 Help me figure out if that's doing the work. 604 00:46:38.592 --> 00:46:41.054 Plus, by the way, am I still answering questions from my audience? 605 00:46:41.935 --> 00:46:45.117 So I'm constantly writing myself. 606 00:46:45.697 --> 00:46:47.699 I'm constantly asking questions 607 00:46:48.059 --> 00:46:50.260 Claude, to look at what I've written. 608 00:46:51.941 --> 00:46:54.943 The compound part is it knows what decisions I've already made. 609 00:46:56.023 --> 00:47:00.986 And by the way, you can also say, you've clearly rejected the decision you made three days ago. 610 00:47:01.426 --> 00:47:02.247 Where's that going? 611 00:47:04.170 --> 00:47:05.751 Again, writing is a lonely business. 612 00:47:05.951 --> 00:47:14.435 When you have somebody to talk to about these things, or three people, so I'll go from Claude, and I'll take what I'm doing and work on it. 613 00:47:14.515 --> 00:47:16.055 I'll go back to ChatGPT. 614 00:47:16.636 --> 00:47:24.399 ChatGPT is the film school student who has read every guide, seen everything, does not know what it means. 615 00:47:25.520 --> 00:47:32.203 It's great at, I know you're trying to do this, and here's the, as they say in the Olympics, the compulsories. 616 00:47:33.563 --> 00:47:36.244 Here's what you haven't done or what you are accomplishing. 617 00:47:36.424 --> 00:47:38.946 It's very useful in that sense. 618 00:47:39.146 --> 00:47:43.988 And then I might even go to notebook to say, hey, look at my documents, look at this stuff. 619 00:47:44.368 --> 00:47:46.709 And can I listen to the podcast? 620 00:47:48.910 --> 00:47:50.111 I am so shallow. 621 00:47:51.191 --> 00:47:56.854 I am a sucker for those two podcast hosts going, well, we're talking about something fascinating today. 622 00:47:58.139 --> 00:47:59.941 I love Notebook L on podcasts. 623 00:48:00.001 --> 00:48:03.244 They're so great. 624 00:48:03.565 --> 00:48:05.426 It's the best thing for a writer. 625 00:48:05.827 --> 00:48:07.449 They love what you're working on. 626 00:48:09.811 --> 00:48:12.594 Yeah, so that's roughly kind of like, and it just rinse and repeat. 627 00:48:13.228 --> 00:48:17.511 So do you think we might be able to get access to one or two of those skills? 628 00:48:17.611 --> 00:48:25.517 Because it sounds like such a valuable, I mean, even this process that you've mentioned that Pixar uses, I wouldn't have known that. 629 00:48:25.897 --> 00:48:30.981 But once you've codified it, it's such a helpful sort of shortcut to make sure you're following it. 630 00:48:31.301 --> 00:48:32.462 I teach an online class. 631 00:48:33.643 --> 00:48:35.844 at the Barcelona School of Art and Design. 632 00:48:37.024 --> 00:48:43.947 The reason is that European film schools have embraced AI in a big way that American film schools are very afraid to do. 633 00:48:45.007 --> 00:48:52.130 And I created a skill that helps you create a pitch deck for your idea. 634 00:48:52.980 --> 00:48:53.760 which is kind of great. 635 00:48:54.060 --> 00:49:04.483 And within that skill, there's the Ken Adams, like, okay, can you in one slide do a, here's the normal world, then this happens, because that happens, they do that. 636 00:49:05.043 --> 00:49:09.965 And it basically makes you create the log line and the characters and all that stuff 637 00:49:11.405 --> 00:49:15.468 and it won't let you finish each step. 638 00:49:15.708 --> 00:49:17.989 They won't let you do a step until you've completed the other. 639 00:49:18.009 --> 00:49:19.910 So I'm happy to share that with you. 640 00:49:20.070 --> 00:49:21.351 I am so excited to try it. 641 00:49:21.411 --> 00:49:22.952 Yes, definitely would love that. 642 00:49:23.332 --> 00:49:24.673 And it might not be that good. 643 00:49:24.793 --> 00:49:25.533 You can tell me. 644 00:49:25.733 --> 00:49:26.974 That's... Let me know. 645 00:49:27.234 --> 00:49:28.035 Well, perfect. 646 00:49:28.335 --> 00:49:30.076 I mean, it's so fun, though. 647 00:49:30.196 --> 00:49:38.841 I mean, because then it's like it'll be a second in your brain, you know, to, like, try to follow that process and understand how you might approach. 648 00:49:38.941 --> 00:49:39.061 Yeah. 649 00:49:39.081 --> 00:49:39.181 Yeah. 650 00:49:39.701 --> 00:49:42.927 doing that particular step in production. 651 00:49:42.947 --> 00:49:46.193 So you have a great sub stack. 652 00:49:46.213 --> 00:49:49.559 And speaking of a thousand great fans, I'm a fan of your sub stack. 653 00:49:49.619 --> 00:49:50.581 Where can people find you? 654 00:49:51.070 --> 00:49:51.610 Oh, sure. 655 00:49:51.771 --> 00:49:52.651 It's a subject. 656 00:49:52.711 --> 00:49:55.894 It's called the AI Writers Studio. 657 00:49:56.454 --> 00:49:59.076 Search my name, Fred Graver, and you can't miss it. 658 00:49:59.377 --> 00:49:59.617 Yeah. 659 00:49:59.917 --> 00:50:00.137 Cool. 660 00:50:00.277 --> 00:50:00.698 Terrific. 661 00:50:00.778 --> 00:50:01.278 Thank you. 662 00:50:02.019 --> 00:50:04.681 Hey, thanks for joining us for this conversation with Fred Graver. 663 00:50:04.781 --> 00:50:07.823 If you want to use the skill that Fred shared, it's in the description below. 664 00:50:07.983 --> 00:50:09.344 It's really fun to use. 665 00:50:09.404 --> 00:50:10.485 I highly recommend it. 666 00:50:11.186 --> 00:50:13.027 And you can also find us at every.tube. 667 00:50:13.167 --> 00:50:13.828 We'll see you next time. 668 00:50:14.385 --> 00:50:17.187 If you liked this video, you should check out every.to. 669 00:50:17.308 --> 00:50:21.891 We have a two week free trial for anyone who's watching or listening at the link in the description. 670 00:50:22.632 --> 00:50:25.394 Every is the only subscription you need to stay at the edge of AI. 671 00:50:25.454 --> 00:50:32.420 It is the one subscription that if you're super AI-pilled and you wanna learn how to use it to do the best work of your life, it's got everything you need. 672 00:50:32.520 --> 00:50:35.143 It has education and equipment on the education side, 673 00:50:35.623 --> 00:50:38.665 We review every new model that drops and tell you what's good. 674 00:50:39.066 --> 00:50:47.272 On the equipment side, we have software tools specifically made for you, like an agent that helps you AI pill your whole company, all bundled together under one subscription. 675 00:50:47.752 --> 00:50:50.114 So get two weeks free now at the link in the description.