Meeting Title: Uttam <> Andrew: AI Animation Copilot Date: 2025-05-05 Meeting participants: Uttam Kumaran, Andrew Duplessie, Adamabdulhamid
WEBVTT
1 00:00:32.130 ⇒ 00:00:33.270 Andrew Duplessie: No.
2 00:00:33.270 ⇒ 00:00:34.079 adamabdulhamid: How you doing.
3 00:00:34.610 ⇒ 00:00:36.040 Andrew Duplessie: What hat is that dog.
4 00:00:37.462 ⇒ 00:00:40.489 adamabdulhamid: It’s a like a wedding gift hat.
5 00:00:40.620 ⇒ 00:00:42.820 Andrew Duplessie: Wedding we went to in Mexico.
6 00:00:42.820 ⇒ 00:00:43.840 Andrew Duplessie: Oh, really.
7 00:00:43.840 ⇒ 00:00:51.248 adamabdulhamid: Yeah, like says it’s like the resort logo, and like the says Beach Club. And then their initials.
8 00:00:51.970 ⇒ 00:00:52.919 Andrew Duplessie: What’s the hotel called.
9 00:00:53.686 ⇒ 00:00:57.899 adamabdulhamid: It actually doesn’t that it was like the one of the.
10 00:00:58.070 ⇒ 00:01:01.949 adamabdulhamid: It was like that one of the Hyatt’s or Marriott’s, like in Puerto Vallarta or something.
11 00:01:02.810 ⇒ 00:01:06.930 adamabdulhamid: but it the hat just says Beach Club, and it says their initials in 2025.
12 00:01:07.360 ⇒ 00:01:08.450 Andrew Duplessie: Pretty fresh.
13 00:01:08.450 ⇒ 00:01:09.529 adamabdulhamid: Yeah, it’s kind of sick.
14 00:01:09.750 ⇒ 00:01:12.910 Andrew Duplessie: What? What’s been going on? Dude? How’s Miami Beach?
15 00:01:13.460 ⇒ 00:01:37.689 adamabdulhamid: Good not not too much kind of just onboarding to the new job started about a month ago. So like, still, just kind of getting getting up to speed, figuring out what’s going on. And and you know what we need to work on. Yeah, it’s been good so far. I feel like the I I was up in Sunnyvale. So it’s a London based company, and then Us. Headquarters is in Sunnyvale.
16 00:01:37.790 ⇒ 00:02:06.300 adamabdulhamid: So I was up there for the 1st week, and then, you know, been remote, you know, since then I’m going to London. It like right after Phoebe’s wedding, actually like early July for a week. So the the remote onboarding is a little bit tough. I just have like tons of questions, you know, and it’s like, I want to ask people. And like, I think also the time zone makes it a little tough like I want to ask them. But then it’s like, Oh, this person’s in London. And so it’s already like, I gotta wait 24 h, basically to get an answer. But it’s all. It’s all good. It’s it’s been pretty good.
17 00:02:06.300 ⇒ 00:02:08.830 Andrew Duplessie: I do the same. That’s why I kinda like.
18 00:02:09.070 ⇒ 00:02:18.053 Andrew Duplessie: I work like 2 blocks. I work like a morning block, and then I take like the afternoon off, and then I’ll work like a night block and get all my questions in whatever for the London people.
19 00:02:18.610 ⇒ 00:02:20.420 Andrew Duplessie: A bunch of bullshit, but.
20 00:02:20.570 ⇒ 00:02:21.015 adamabdulhamid: Yeah.
21 00:02:21.460 ⇒ 00:02:24.239 Andrew Duplessie: That’s probably the worst part about it, right? Is like, when you wake up.
22 00:02:25.100 ⇒ 00:02:26.880 adamabdulhamid: It’s already like everything, all the responses.
23 00:02:26.880 ⇒ 00:02:31.849 Andrew Duplessie: Messages. Things are melting down. You just gotta make sure you get your meditating like in before.
24 00:02:31.850 ⇒ 00:02:35.368 adamabdulhamid: Yeah. So you don’t just wake up distressed.
25 00:02:35.930 ⇒ 00:02:36.550 Andrew Duplessie: Truly, though.
26 00:02:37.160 ⇒ 00:02:37.800 adamabdulhamid: How you doing.
27 00:02:38.673 ⇒ 00:02:41.710 Andrew Duplessie: Good dude, I mean business as usual.
28 00:02:42.282 ⇒ 00:02:50.730 Andrew Duplessie: The guy on the call today is my friend. Super smart has his own company based in the Us.
29 00:02:51.040 ⇒ 00:02:56.889 Andrew Duplessie: And is just building really cool stuff like he’s working with different companies to just
30 00:02:57.170 ⇒ 00:03:01.570 Andrew Duplessie: add AI, you know, X and
31 00:03:03.110 ⇒ 00:03:05.720 Andrew Duplessie: I got to talking with him over the past couple of weeks.
32 00:03:06.400 ⇒ 00:03:14.139 Andrew Duplessie: cause I’m usually anti outsourcer, right? Like we’ve talked about this. But we have the capital, and this guy is ready to build.
33 00:03:14.280 ⇒ 00:03:17.889 Andrew Duplessie: and he’s a great builder, and so
34 00:03:18.000 ⇒ 00:03:25.719 Andrew Duplessie: I just think it’s a a good match to be honest. And so this would. This would be building like the prototype that we discussed, which is like
35 00:03:26.680 ⇒ 00:03:35.829 Andrew Duplessie: the AI generator and then below it is like the editor, editor and a producer hops in, and the producer can build what could be like a short episode of something.
36 00:03:36.704 ⇒ 00:03:38.877 adamabdulhamid: Yeah, yeah, I saw
37 00:03:40.330 ⇒ 00:03:42.440 adamabdulhamid: There’s been.
38 00:03:42.970 ⇒ 00:03:45.867 adamabdulhamid: I saw one extra product announcement,
39 00:03:46.530 ⇒ 00:03:56.135 adamabdulhamid: from I don’t know the name of the company. Let me see, I can find it again. Actually, I’ll send you the Twitter post that like looks. I think we can use it as like inspiration. It’s it’s more like
40 00:03:56.600 ⇒ 00:03:58.830 Andrew Duplessie: This is the voice, one the Titanic.
41 00:03:59.860 ⇒ 00:04:04.260 adamabdulhamid: No, it wasn’t the voice one. Let me here, let me just text this to you right now.
42 00:04:04.260 ⇒ 00:04:05.880 Andrew Duplessie: The Titanic one’s wild.
43 00:04:07.302 ⇒ 00:04:15.106 adamabdulhamid: I just sent you the Twitter post. But or let me oh, is this like
44 00:04:18.490 ⇒ 00:04:19.499 Andrew Duplessie: Not one thing.
45 00:04:20.329 ⇒ 00:04:21.839 adamabdulhamid: Oh, they made like.
46 00:04:23.600 ⇒ 00:04:25.190 Andrew Duplessie: I don’t know how they did that shit.
47 00:04:25.740 ⇒ 00:04:30.980 Andrew Duplessie: I also recently saw a really accurate lip dub like a crazy, accurate lip dub.
48 00:04:31.360 ⇒ 00:04:31.840 Andrew Duplessie: I don’t!
49 00:04:32.860 ⇒ 00:04:33.610 Andrew Duplessie: Oh, here it is!
50 00:04:34.510 ⇒ 00:04:36.570 Andrew Duplessie: This lift up is like insane.
51 00:04:39.280 ⇒ 00:04:44.559 adamabdulhamid: But anyway, yeah, the the one I sent you is like, I think this is like, this is kind of what I was talking about, where they’re.
52 00:04:44.560 ⇒ 00:04:45.340 Andrew Duplessie: Yeah.
53 00:04:45.340 ⇒ 00:04:49.940 adamabdulhamid: They’re trying to do like AI Photoshop, you know, like you just say, like, Oh, blend! Blend these scenes, whatever
54 00:04:50.350 ⇒ 00:04:55.400 adamabdulhamid: so like. But again, I think like I still feel.
55 00:04:55.400 ⇒ 00:04:57.419 Andrew Duplessie: You’re gonna mass distribution, right?
56 00:04:57.810 ⇒ 00:04:58.670 adamabdulhamid: What do you say?
57 00:04:58.800 ⇒ 00:05:03.639 Andrew Duplessie: Like they’re trying to sign up. Editors pay pay per month. Kind of situation.
58 00:05:03.640 ⇒ 00:05:11.160 adamabdulhamid: Yeah, I assume they’re just like their business is like Photoshop, right? Like they just won’t be able to pay for access to the tool. And then you upload your videos or you.
59 00:05:11.160 ⇒ 00:05:11.620 Andrew Duplessie: Yeah.
60 00:05:11.620 ⇒ 00:05:27.850 adamabdulhamid: You whether they’re real or AI generated, or whatever. And then you just ask the editor like, you know you like like you would in Imovie or Photoshop, or whatever you like. You highlight the clip, and you’re like, Hey, change this bit, or like, you know, I want you to blend blend these scenes together better or something.
61 00:05:28.197 ⇒ 00:05:36.269 adamabdulhamid: But anyway, yeah, I think, like all you know, I think since we last talked, there’s just been, by the way is this guy supposed to join a specific time, or is it like.
62 00:05:37.250 ⇒ 00:05:38.176 Andrew Duplessie: He’s late.
63 00:05:38.640 ⇒ 00:05:44.589 adamabdulhamid: Okay, theoretically. Now, okay, cool. But yeah, like, I think, like, there, you know.
64 00:05:44.910 ⇒ 00:06:00.679 adamabdulhamid: the the video gen models are still getting better, which is great, I think. I think there’s still going to be a need for like like there was that like 3 day period where I felt like my whole twitter. Timeline was just like people doing video like shorts or whatever. And then it kind of died. But I think I think
65 00:06:01.130 ⇒ 00:06:25.419 adamabdulhamid: I think that’s definitely still going to be. There is like, How do you just like organize the content like, have the different image generation options? There’s a bunch of these different like image to video, text, to video image and text to video, how do you just like orchestrate them? So you can like try different options, you know. See the results, like all in one kind of tool and and sort of, you know. Organize your the whole, the whole process. So I I think that’s like a great sort of tool to start with.
66 00:06:25.803 ⇒ 00:06:41.259 adamabdulhamid: And you know it’ll just be like plugging in all the different Apis for for the different you know, for runway for Google studio that has vo. 2, probably openai like plug plugging in the Apis altogether. So you know, you can just pick and pick and play sort of.
67 00:06:42.420 ⇒ 00:06:42.760 Andrew Duplessie: So.
68 00:06:42.760 ⇒ 00:06:45.529 Uttam Kumaran: Hi, sorry for being late. I
69 00:06:46.180 ⇒ 00:06:49.370 Uttam Kumaran: just in a very, very interesting AI meeting.
70 00:06:49.740 ⇒ 00:06:50.420 Andrew Duplessie: I love that.
71 00:06:50.420 ⇒ 00:06:54.741 Uttam Kumaran: A/C. Guy showed up, and then I like just like lost track of time.
72 00:06:55.507 ⇒ 00:06:57.170 Andrew Duplessie: You’re good dude, you’re good. How you doing.
73 00:06:57.170 ⇒ 00:06:58.070 Uttam Kumaran: How are you?
74 00:06:58.760 ⇒ 00:07:03.879 Andrew Duplessie: I’m good. I’m good man. So, Tom, this is Adam Adam. This is Tom. I worked with.
75 00:07:03.880 ⇒ 00:07:04.500 Uttam Kumaran: It’s Adam.
76 00:07:04.500 ⇒ 00:07:10.010 Andrew Duplessie: While at flow code. But he was one of the bright spots.
77 00:07:10.544 ⇒ 00:07:18.005 Andrew Duplessie: Adam. Adam, I met a long time ago, now, like 8 years ago, I guess, and he
78 00:07:18.900 ⇒ 00:07:30.419 Andrew Duplessie: knew Sarah well from college, went to Stanford, and then he worked at cruise. The the AI car startup. And then, Adam, actually, what? Where are you at now? Just so. I know.
79 00:07:30.420 ⇒ 00:07:59.949 adamabdulhamid: Yeah. So I worked at cruise for a long time like working building some of the core AI models core driving models for the self driving car. And then I just left about a month ago and joined a new self driving company called Wave. It’s a London based company, but they’re kind of doing end to end sort of, you know. Ml, modeling approach. So it’s like, you know, camera images in and like direct like control, you know. Turn the steering wheel and accelerate the kind of output.
80 00:08:00.321 ⇒ 00:08:02.439 adamabdulhamid: So kind of kind of like Tesla for self driving.
81 00:08:02.440 ⇒ 00:08:03.980 Uttam Kumaran: Trying to do the Tesla model. Yeah.
82 00:08:03.980 ⇒ 00:08:11.240 adamabdulhamid: Yeah, exactly. So yeah, I just joined. Joined there a month ago. And yeah, sort of work on on some of the core core AI systems. There.
83 00:08:11.840 ⇒ 00:08:12.159 Andrew Duplessie: And
84 00:08:12.480 ⇒ 00:08:14.620 Uttam Kumaran: How is it? How is it being in that industry right now?
85 00:08:15.315 ⇒ 00:08:38.080 adamabdulhamid: It’s good. I think there’s like tons of momentum on the like sort of embodied AI like, you know, robotics industry. Right now, I think lots of lots of funding, and, like the, you know, figure like all these other like home robot companies as well. So I think the industry is excited about about like, you know, vision vision like the vision language action models like combining basically
86 00:08:38.080 ⇒ 00:08:59.369 adamabdulhamid: large Llm style systems and robotic control based systems. So cool to see, you know, cool to see the sort of external interest. I think this company, specifically is is trying to sell the tech to other companies like, you know, they want to go to Toyota and be like, Hey, you want your car to drive itself, like, you know. Throw. Throw a tech in there.
87 00:08:59.734 ⇒ 00:09:13.390 adamabdulhamid: So I think it’ll be like those companies are probably gonna be slow and a little bit just like you know behind behind, in terms of like understanding what they even want and and what capabilities there are. But I think it’s a exciting time for sure.
88 00:09:13.640 ⇒ 00:09:15.719 Uttam Kumaran: Nice. Yeah, that’s dope.
89 00:09:16.210 ⇒ 00:09:16.835 Andrew Duplessie: And
90 00:09:17.570 ⇒ 00:09:32.009 Andrew Duplessie: and Adam, Utam started, basically successful business after flow code building, AI platforms and software and different sort of stuff for for all types of businesses, right? I don’t know. It’s the kind of range with you.
91 00:09:32.570 ⇒ 00:09:46.920 Uttam Kumaran: Yeah, it’s range. So my background’s in like data engineering. So I worked as a data engineer work like built the data team at flow code and then led product at a data startup out of after flow code. And then.
92 00:09:47.544 ⇒ 00:10:00.429 Uttam Kumaran: like, yeah, I was here. I’m here in Austin. I quit that job. And then just started just a services company around primarily implementing data analytics. So we build analytics infrastructure. But I was using AI really heavily like, probably
93 00:10:00.830 ⇒ 00:10:11.316 Uttam Kumaran: like one of the core reasons like I was able to do this business was just being able to do use AI. So I was. I could like, push out hiring and get a lot done, and
94 00:10:11.600 ⇒ 00:10:12.100 adamabdulhamid: Yeah.
95 00:10:12.100 ⇒ 00:10:32.299 Uttam Kumaran: So I just learned a ton about building practical like AI systems for business owners. And then, now, I sell practical AI systems to business owners. So we implement a lot of our focus is on revenue growth. So it’s not a lot on like cost cutting stuff mostly on like developing internal sort of copilots for employees.
96 00:10:32.360 ⇒ 00:10:42.889 Uttam Kumaran: determining like developing like retrieval systems. You know, plugging into like go to market motions and using AI,
97 00:10:44.520 ⇒ 00:10:56.939 Uttam Kumaran: we are Service Company, so I use as much off the shelf as I can. But a lot of this world there’s like not much off everything off the shelf like isn’t so great. Some stuff that’s been here for at least 2 years is getting better. But.
98 00:10:57.491 ⇒ 00:11:07.969 Uttam Kumaran: I don’t know. There’s there’s still so much to go. So we’ve been able to have a lot of success in the last few months, just like going to market with just being able to help build some of these.
99 00:11:08.240 ⇒ 00:11:19.172 Uttam Kumaran: you know, whether it’s like proof of concepts or or Demos, or whether it’s actually like deploying AI agents to employees so they can Access company information. They can take actions that would have previously taken them hours.
100 00:11:19.937 ⇒ 00:11:23.822 Uttam Kumaran: You know, so so kind of everything around that, but all really focused on
101 00:11:24.210 ⇒ 00:11:28.565 Uttam Kumaran: partnering with people that are growing so like kind of totally skipping the
102 00:11:29.130 ⇒ 00:11:32.630 Uttam Kumaran: bring us in. And you fire like 30 people. Stuff like that like, I just don’t.
103 00:11:32.630 ⇒ 00:11:33.260 adamabdulhamid: Yeah.
104 00:11:33.260 ⇒ 00:11:38.550 Uttam Kumaran: Not not interested in that narrative like at all. Most of what I think a lot about is like, how do we
105 00:11:38.700 ⇒ 00:11:41.330 Uttam Kumaran: make people 20 or 30% more efficient.
106 00:11:41.972 ⇒ 00:11:50.109 Uttam Kumaran: And that’ll speak for itself. Like people, your employees will be happier, more money will come in. You can avoid like a couple of future hires.
107 00:11:50.240 ⇒ 00:12:00.589 Uttam Kumaran: and like I did it. So I have really like no like I don’t know. I I do it in my business, so it’s very easy for me to like, have an honest conversation with Ceos, and stuff about like.
108 00:12:00.750 ⇒ 00:12:05.039 Uttam Kumaran: what’s like vaporware. What’s like? Sort of like thought leadership versus.
109 00:12:05.470 ⇒ 00:12:09.300 Uttam Kumaran: Oh, yeah, you shouldn’t hire me. Just go like implement, this thing really quick. And like.
110 00:12:09.520 ⇒ 00:12:11.839 Uttam Kumaran: so, yeah, so it’s worked out for us.
111 00:12:12.450 ⇒ 00:12:13.350 adamabdulhamid: Super cool.
112 00:12:13.350 ⇒ 00:12:17.679 Uttam Kumaran: Yeah. So I mean this, this project is dope. I don’t know if
113 00:12:18.072 ⇒ 00:12:21.710 Uttam Kumaran: Andrew, if you were able to share this with Adam, but I’ll send this, and I just sent.
114 00:12:21.710 ⇒ 00:12:22.179 Andrew Duplessie: Oh, yeah.
115 00:12:22.180 ⇒ 00:12:22.910 Uttam Kumaran: Chat.
116 00:12:23.030 ⇒ 00:12:23.800 Uttam Kumaran: Yeah.
117 00:12:23.800 ⇒ 00:12:25.310 Andrew Duplessie: I didn’t share with him, but.
118 00:12:25.310 ⇒ 00:12:27.199 Uttam Kumaran: I’ll I’ll just add you to the
119 00:12:27.800 ⇒ 00:12:28.480 Andrew Duplessie: Perfect.
120 00:12:28.730 ⇒ 00:12:32.099 Uttam Kumaran: Yeah, I’ll add you here. And we just basically
121 00:12:33.180 ⇒ 00:12:35.980 Uttam Kumaran: just wrote up sort of what
122 00:12:36.620 ⇒ 00:12:43.340 Uttam Kumaran: context Andrew gave to me. You should have access. Now, yeah.
123 00:12:43.340 ⇒ 00:12:43.880 adamabdulhamid: Good.
124 00:12:44.700 ⇒ 00:12:45.020 Andrew Duplessie: Got it.
125 00:12:45.020 ⇒ 00:12:50.659 Uttam Kumaran: But yeah, I mean one. I’m I’m pumped dude. I’m like, I’m like waiting for you to go do something for like.
126 00:12:50.810 ⇒ 00:12:51.220 Andrew Duplessie: Yeah.
127 00:12:51.220 ⇒ 00:12:53.580 Uttam Kumaran: Long time. I mean, you do do a lot of things, but.
128 00:12:53.850 ⇒ 00:12:55.279 Andrew Duplessie: No, this is really cool.
129 00:12:55.280 ⇒ 00:12:57.089 Andrew Duplessie: For a long time. We’ve been talking about this for a long time.
130 00:12:57.090 ⇒ 00:12:59.139 Uttam Kumaran: This is a really cool, cool product.
131 00:12:59.490 ⇒ 00:13:04.770 Uttam Kumaran: I also think, overall your thought process of like not
132 00:13:04.870 ⇒ 00:13:09.899 Uttam Kumaran: I mean, I don’t know what again. I only know what you told me. But if you’re thinking about
133 00:13:10.640 ⇒ 00:13:11.900 Uttam Kumaran: raising
134 00:13:12.290 ⇒ 00:13:27.729 Uttam Kumaran: and using this technology internally and sort of changing the cost model of a production studio. I feel like that’s much more compelling and much lower competition than like trying to build this into a product that you have to go sell to production studios.
135 00:13:28.280 ⇒ 00:13:33.120 Uttam Kumaran: It can be way less polished. You can iterate way faster.
136 00:13:33.300 ⇒ 00:13:37.130 Uttam Kumaran: and you can make sure that your team adopts it from the top down.
137 00:13:39.080 ⇒ 00:13:53.639 Uttam Kumaran: you’re of course you’re you’re starting a production company. But like, I don’t know, I think this is the way to go. I think about this a lot in a couple of under other industries like legal insurance, where I have people talking to me about like, hey? Why don’t we go buy a law firm
138 00:13:53.810 ⇒ 00:14:01.430 Uttam Kumaran: and basically just implement Harvey like, got it and then flip it like.
139 00:14:01.550 ⇒ 00:14:02.080 Andrew Duplessie: Yeah.
140 00:14:02.080 ⇒ 00:14:07.439 Uttam Kumaran: You just sort of have AI in the DNA of everything, from the go to market motion to customer interaction to
141 00:14:07.710 ⇒ 00:14:09.010 Uttam Kumaran: sales. So
142 00:14:09.110 ⇒ 00:14:12.620 Uttam Kumaran: I don’t know. I think it’s a really cool idea. I don’t think I’ve seen anyone think about this.
143 00:14:13.220 ⇒ 00:14:16.670 Uttam Kumaran: I just don’t think the talent exists in that industry
144 00:14:17.360 ⇒ 00:14:22.760 Uttam Kumaran: with people who are able to start new firms, but start them like AI native.
145 00:14:23.100 ⇒ 00:14:27.139 Uttam Kumaran: It’s gonna be a lot of legacy firms adopting runway Ml, or stuff.
146 00:14:27.560 ⇒ 00:14:29.950 adamabdulhamid: Yeah. And that’s gonna take a long time.
147 00:14:31.690 ⇒ 00:14:37.720 Uttam Kumaran: And yeah, I don’t know. I think it’s compelling, Eddie. So we just wrote, I just wrote up a lot of from from our sort of text back and forth.
148 00:14:38.060 ⇒ 00:14:38.570 Uttam Kumaran: Yes.
149 00:14:38.580 ⇒ 00:14:50.359 Andrew Duplessie: This is perfect man like I think I think for me that’s exactly it like. And, Adam, I think we’re aligned on this because we’ve talked about it a lot, but essentially like the outcome that I’m looking for is one.
150 00:14:50.960 ⇒ 00:14:54.719 Andrew Duplessie: Well, I’ll just give you like the the lay of land, Tom.
151 00:14:54.910 ⇒ 00:15:05.480 Andrew Duplessie: So I have a couple of investors that I’ve circled for this like, pre see round out call, which I’m gonna close and build this this.
152 00:15:05.690 ⇒ 00:15:33.430 Andrew Duplessie: It’s gonna serve as sort of our proof of concept what comes out of it like, can we create the bluey, the South Park, the bob’s burgers? Can we create an animated show that’s like legible to social audiences. And and you hit that you hit it on the head where this is. Gonna stay like 100 internal. A producer is actually gonna go into our our, this Ui, and they’re gonna be able to create a show. And so like the end outcome is that is like Adam and I are able to drop 3 producers in this.
153 00:15:33.430 ⇒ 00:15:49.943 Andrew Duplessie: and we give them sort of like a loose structure, like what we want them to accomplish. They deliver a story, and they begin to produce this. This could even be like 5, 10 second clips right on an Instagram, because I don’t. I don’t think as much as I’d like it to be. I’m still not seeing anything of quality yet.
154 00:15:50.210 ⇒ 00:15:50.760 Uttam Kumaran: Yeah.
155 00:15:50.760 ⇒ 00:15:59.952 Andrew Duplessie: Which is concerning but that’s that’s it. We’re empowering creators, but on a quality and really like defined internal scale.
156 00:16:00.910 ⇒ 00:16:06.200 Andrew Duplessie: and and that’s really so like, I think, for this, like, there’s 2 things that are important to me is like.
157 00:16:06.370 ⇒ 00:16:08.760 Andrew Duplessie: I think the ui has to look great.
158 00:16:09.470 ⇒ 00:16:26.269 Andrew Duplessie: So I don’t know if you have someone, or we bring in Ivana or someone like that. I think the Us. To look really showy and really good, because all these investors that I’m working with are not super technical. And then the other side is you know, we just gotta kind of test and learn. I think
159 00:16:26.420 ⇒ 00:16:28.962 Andrew Duplessie: I don’t know if I’m missing stuff, Adam.
160 00:16:29.546 ⇒ 00:16:39.923 adamabdulhamid: Yeah. No, that that sounds great, I think. Let me just reiterate, I guess what I we had talked about before, Andrew, just to like, you know, clarify what? My, what’s on my mind. So yeah, I think,
161 00:16:40.650 ⇒ 00:16:43.089 adamabdulhamid: I totally agree. I think there’s like
162 00:16:43.210 ⇒ 00:16:48.529 adamabdulhamid: we are talking about 2 main side, like sort of outputs of of this business. One is
163 00:16:48.720 ⇒ 00:16:58.499 adamabdulhamid: the actual like tech tool, right? And I think then, like, if you want to go raise money and build that and sell it to other companies. It’s like, it’s a full blown software business, right? Like you’re you’re building a tech startup
164 00:16:58.962 ⇒ 00:17:06.559 adamabdulhamid: and it’s like, sure you can call it a wrapper on runway and open, you know, Sora and vo, 2. And all these like, you know, Api based models
165 00:17:06.670 ⇒ 00:17:15.359 adamabdulhamid: that I think is hard. It’s just like, you know, there’s gonna be more competition there. There’s like hard harder to get it off the ground right, so that that was not the goal.
166 00:17:15.369 ⇒ 00:17:16.099 Uttam Kumaran: That! Don’t.
167 00:17:16.099 ⇒ 00:17:20.059 adamabdulhamid: Yeah, the the goal was really, say, like, build like a really
168 00:17:20.209 ⇒ 00:17:25.749 adamabdulhamid: scrappy version of this for print, like, build a content studio. That is just like AI supercharged.
169 00:17:26.254 ⇒ 00:17:47.889 adamabdulhamid: You know, build some tool that can be used internally to help producers ourselves, etc. Build even like Andrew mentioned proof of concept clips like short clips, little animated things try to understand, like what the models are capable of, what they’re good at, what stories they can. Sort of animate in a in a coherent sense, what styles they work for that kind of stuff. And so
170 00:17:48.129 ⇒ 00:17:56.659 adamabdulhamid: the tool workflow that I was imagining is kind of like, you know, sort of Photoshop ask. But even like, I think the core functionality we really need is like
171 00:17:57.379 ⇒ 00:18:14.849 adamabdulhamid: all of these, the models have like a variety of capabilities. Some are text text based inputs. You know, you give it a prompt, you want to say, like, make a video that does this or make an image that looks like this? Some are are image to video, right? It’s like, takes an image as input, some are both. It takes like an image and a prompt and produces a video.
172 00:18:14.949 ⇒ 00:18:20.309 adamabdulhamid: And basically, I think that the core you know thing we need from the tool is to be able to
173 00:18:20.649 ⇒ 00:18:21.364 adamabdulhamid: like,
174 00:18:22.149 ⇒ 00:18:45.049 adamabdulhamid: categorize and maintain the like, the like, the provenance of all the information that you’re producing. You like you said. You know you asked, you know 4. 0. To generate an image of this character, and then you want to use that image as the seed for some vo. 2, to generate a little clip of animating it. And, like the producer needs to see. Like, okay, I used this image to produce this video. And here’s what the output was.
175 00:18:45.049 ⇒ 00:19:05.859 adamabdulhamid: And you know, maybe with this prompt and like, I like this style. But maybe I want to try with a different seed image to see if the you know new style looks good or something like that. So I feel like that workflow like right now, I assume it’s just people just like with files, you know, in on their Mac, just like, you know, keeping track of like. Oh, I you know I use this prompt. I use this thing so like a basic like
176 00:19:05.859 ⇒ 00:19:17.859 adamabdulhamid: orchestration tool, to say, like, I want to use this model, generate an image and then feed it to an you know, new model kind of see what the what the content that you’re building looks like, you know, a as you’re sort of like working in this tool.
177 00:19:19.640 ⇒ 00:19:20.800 Uttam Kumaran: Yeah, does that make sense.
178 00:19:21.900 ⇒ 00:19:23.609 Andrew Duplessie: Fine with me. Go ahead, John.
179 00:19:23.610 ⇒ 00:19:25.499 Uttam Kumaran: Yeah, I think the other thing is like.
180 00:19:26.370 ⇒ 00:19:35.459 Uttam Kumaran: you know, this is where it’s like, I sort of want to start in 2 ways, one like thinking about. That’s why I kind of been sending a couple of links. But like.
181 00:19:35.660 ⇒ 00:19:41.090 Uttam Kumaran: what does like, okay, good. And like, great look like.
182 00:19:41.090 ⇒ 00:19:41.480 Andrew Duplessie: Yeah.
183 00:19:41.480 ⇒ 00:19:45.540 Uttam Kumaran: Like working backwards from there in terms of like the final output.
184 00:19:45.930 ⇒ 00:19:47.809 Uttam Kumaran: which would be kind of helpful.
185 00:19:48.806 ⇒ 00:19:50.379 Uttam Kumaran: Second is
186 00:19:52.320 ⇒ 00:19:58.769 Uttam Kumaran: to your point like, I really want to see the steps, that of the production process, and then almost like.
187 00:19:59.290 ⇒ 00:20:07.899 Uttam Kumaran: take off what this tool augments or completely replaces, that will help you also get us be able to, for the deck
188 00:20:08.190 ⇒ 00:20:14.730 Uttam Kumaran: put together like what the potential impact is like. Hey, this tool, when, like.
189 00:20:14.880 ⇒ 00:20:18.439 Uttam Kumaran: for example, you could say, v, 1 of this tool will augment
190 00:20:18.570 ⇒ 00:20:36.099 Uttam Kumaran: 30% of the production from 0 to the some 20 second thing and then completely replace 5%. V 2 will take it to like 40% like 50% augmented meaning like they use it as an assist. Then 30% can be almost like
191 00:20:36.360 ⇒ 00:20:44.280 Uttam Kumaran: like it just runs on its own, or something like for me, like I want I need to see completely broken down like that.
192 00:20:44.763 ⇒ 00:20:52.860 Uttam Kumaran: Because there, there are pieces that I think we lack the context on what is the current production process? And I think it’ll help you
193 00:20:53.220 ⇒ 00:21:08.040 Uttam Kumaran: basically show your depth on like, how you’re looking at production as an engineering problem like it is a manufacturing problem. And like, we can help, you basically start to structure it that way. And then you pick off. We, we basically can prioritize either, like
194 00:21:08.410 ⇒ 00:21:12.820 Uttam Kumaran: for what you what is must have for like the raise.
195 00:21:12.920 ⇒ 00:21:18.160 Uttam Kumaran: And then by basically like ease of doing. And then how long it’s gonna take.
196 00:21:19.300 ⇒ 00:21:27.940 Uttam Kumaran: And then we can sort of just like that’s your pro. That’s basically your your roadmap for this internal tool, and if you want to prioritize
197 00:21:28.445 ⇒ 00:21:43.820 Uttam Kumaran: certain workflow augmentations before others, you can do that. But then it also it’ll also allow you to understand. One core thing about AI is that you don’t. I don’t think biting off the entire piece is worth it at all.
198 00:21:43.970 ⇒ 00:21:54.400 Uttam Kumaran: I think, even if you were to get 25% speed improvements. You have a really great business here. So if you’re able to get 50 like you’re solid
199 00:21:54.630 ⇒ 00:21:57.830 Uttam Kumaran: coming into this and thinking to augment the 100% and like.
200 00:21:58.050 ⇒ 00:22:01.769 Uttam Kumaran: give something, get something like that’s too big of a scope.
201 00:22:02.250 ⇒ 00:22:04.630 Uttam Kumaran: So I I really think about
202 00:22:05.010 ⇒ 00:22:10.800 Uttam Kumaran: trying to lower that so you can really control the outcomes, and then you just stitch the pieces together.
203 00:22:10.900 ⇒ 00:22:11.840 Andrew Duplessie: Now.
204 00:22:13.585 ⇒ 00:22:17.944 adamabdulhamid: Yeah, I I actually I I’ve I agree with that a lot. So
205 00:22:18.380 ⇒ 00:22:24.386 adamabdulhamid: let me see if I can share my screen. I there was like I saw some demo. Oh, I don’t have
206 00:22:25.550 ⇒ 00:22:30.504 adamabdulhamid: I got a new laptop, so I have like no permissions to share anything yet. But let me let me send this
207 00:22:30.880 ⇒ 00:22:31.610 adamabdulhamid: this
208 00:22:32.620 ⇒ 00:22:46.178 adamabdulhamid: I just sent a tweet in that group chat there’s like. So I saw. I don’t know if this is like a real mature company or something it sounds like, seems some kid who’s like made this like AI Photoshop thing. I actually think this is like not the right
209 00:22:46.840 ⇒ 00:22:53.279 adamabdulhamid: like ui right away. Like, I think this is like, too, yeah, like, this is like full blown
210 00:22:53.969 ⇒ 00:23:09.640 adamabdulhamid: like, you know you. It looks like Imovie, but just like you have your like cursor style, you know, like chat on the right or something, and like this seems tough like I don’t expect the models are, gonna be so good that you can just say, like, Hey, like, fix this change. This, you know.
211 00:23:09.640 ⇒ 00:23:09.990 Uttam Kumaran: Yeah.
212 00:23:09.990 ⇒ 00:23:14.330 adamabdulhamid: I think it’s gonna need to be like smaller scope. To begin with, like, okay, I want to use.
213 00:23:14.330 ⇒ 00:23:14.870 Uttam Kumaran: Yeah.
214 00:23:14.870 ⇒ 00:23:15.230 adamabdulhamid: Even tomorrow.
215 00:23:15.230 ⇒ 00:23:15.680 Uttam Kumaran: Totally.
216 00:23:16.070 ⇒ 00:23:16.900 adamabdulhamid: Yeah.
217 00:23:16.900 ⇒ 00:23:21.830 Uttam Kumaran: But you’re but you’re totally right in that. My, my sort of beef with folks like
218 00:23:22.230 ⇒ 00:23:29.129 Uttam Kumaran: this kid is, I’m like, just build faster videos, then don’t like, think, don’t sell this tool
219 00:23:29.360 ⇒ 00:23:30.429 Uttam Kumaran: if you figure it out.
220 00:23:30.430 ⇒ 00:23:30.920 Uttam Kumaran: Yeah.
221 00:23:30.920 ⇒ 00:23:36.669 Uttam Kumaran: if you learn if you found oil, you’re like, why, like, you know what I mean like, why, why.
222 00:23:36.670 ⇒ 00:23:38.290 adamabdulhamid: Yeah, yeah.
223 00:23:38.290 ⇒ 00:23:42.839 Uttam Kumaran: Up against a video editing tool is impossible, like not a place you want to be in.
224 00:23:44.010 ⇒ 00:23:49.970 adamabdulhamid: Yeah. And and my, I mean, my guess is like a kid like this, like they don’t know. They don’t have.
225 00:23:49.970 ⇒ 00:23:51.740 Uttam Kumaran: No idea what they’re getting into. Yeah.
226 00:23:51.740 ⇒ 00:24:01.409 adamabdulhamid: Yeah, they don’t, they don’t. They also don’t know like they don’t know what like the content creation business is like they don’t have the right people to be like, okay, I made this tool like, why don’t you just use it for what it’s intended?
227 00:24:01.410 ⇒ 00:24:01.800 adamabdulhamid: Yeah, but
228 00:24:01.800 ⇒ 00:24:18.870 adamabdulhamid: and just make better videos or faster videos or cheaper videos. And so, yeah, I think I think that’s totally right. It’s like, we need to start with that as the goal. And then and then build like, okay, what’s the what’s the thing that the AI systems are best at right now that can augment that workflow and and start there.
229 00:24:21.890 ⇒ 00:24:23.959 Uttam Kumaran: So I know we had this. I mean, I think.
230 00:24:24.420 ⇒ 00:24:27.790 Uttam Kumaran: like you, I don’t think the ui is like the
231 00:24:28.220 ⇒ 00:24:30.559 Uttam Kumaran: the ui is probably the easiest thing out of all of it.
232 00:24:30.560 ⇒ 00:24:31.549 Andrew Duplessie: Yeah, I agree, yeah.
233 00:24:32.810 ⇒ 00:24:41.850 Uttam Kumaran: like, whatever you I mean, basically whatever you want, I mean. But you know, like, it’s whatever what what works may not be, what you need to sell. So like.
234 00:24:42.180 ⇒ 00:24:45.039 Uttam Kumaran: I think that’s up. It’s up to you. This is where, like
235 00:24:46.650 ⇒ 00:24:51.630 Uttam Kumaran: I think we should almost set like what you need for the raise versus like.
236 00:24:52.030 ⇒ 00:24:52.420 Andrew Duplessie: Yeah.
237 00:24:52.420 ⇒ 00:24:53.970 Uttam Kumaran: Need on a demo.
238 00:24:53.970 ⇒ 00:24:54.450 Andrew Duplessie: Yeah.
239 00:24:54.450 ⇒ 00:24:57.909 Uttam Kumaran: Saying that those are like so far apart, but.
240 00:24:58.220 ⇒ 00:24:58.850 Andrew Duplessie: No, it’s true.
241 00:24:58.850 ⇒ 00:24:59.780 Uttam Kumaran: The same 2 things.
242 00:24:59.780 ⇒ 00:25:00.500 Andrew Duplessie: Things.
243 00:25:00.670 ⇒ 00:25:04.580 Uttam Kumaran: Like. They don’t say the so if they don’t serve the same purpose
244 00:25:04.730 ⇒ 00:25:13.339 Uttam Kumaran: like you want to raise money, and then you also want to have the the inklings of a product that’s actually going to to work and test some hypothesis.
245 00:25:13.630 ⇒ 00:25:28.430 Uttam Kumaran: That’s why I sort of asked about a little bit about the timeline, so to know whether we could prioritize one or the other, or if you’re not sort of set on the timeline, then at least we need to set some milestones for what you need to see.
246 00:25:30.460 ⇒ 00:25:31.990 Uttam Kumaran: To be able to then
247 00:25:32.270 ⇒ 00:25:42.159 Uttam Kumaran: like. Of course, I think the biggest milestone is is raising. So if you were to say cool that maybe at some point, what do you need to see from the
248 00:25:42.340 ⇒ 00:25:43.989 Uttam Kumaran: proof of concept?
249 00:25:44.150 ⇒ 00:25:48.929 Uttam Kumaran: In order to then be like, okay, we’re we’re 5% through.
250 00:25:49.140 ⇒ 00:25:51.749 Uttam Kumaran: I have the vision. Now, I can go do that right like.
251 00:25:51.750 ⇒ 00:25:52.510 Andrew Duplessie: Yeah.
252 00:25:52.510 ⇒ 00:25:54.180 Uttam Kumaran: I want to sort of back into.
253 00:25:54.180 ⇒ 00:25:56.419 Andrew Duplessie: Yeah, that makes sense. So I think it’s this.
254 00:25:56.680 ⇒ 00:26:04.930 Andrew Duplessie: I think it’s it’s we are showing that this tool can create like a relatively concise
255 00:26:05.330 ⇒ 00:26:09.020 Andrew Duplessie: episode, or something that can go on social media
256 00:26:09.330 ⇒ 00:26:22.779 Andrew Duplessie: as a part of that that tool. I mean the the output, the content that comes from this tool goes viral or does really well on the Internet. So like when I’m walking into the the investor meetings in the fall, let’s say like.
257 00:26:23.360 ⇒ 00:26:28.147 Andrew Duplessie: if there was a timeline, I would say, like, I really want to be pitching this like mid September.
258 00:26:29.130 ⇒ 00:26:30.010 Andrew Duplessie: like.
259 00:26:30.490 ⇒ 00:26:36.310 Andrew Duplessie: I have the content in hand which did really well on social, and I have the data to show for it.
260 00:26:36.820 ⇒ 00:26:42.130 Andrew Duplessie: And then I have this tool that I can pull up in front of an investor and click around.
261 00:26:42.460 ⇒ 00:26:44.379 Andrew Duplessie: And the story is is like
262 00:26:45.210 ⇒ 00:26:57.249 Andrew Duplessie: we, we show this model where, like we pair a producer that we’ve kind of coached and brought in, that has their own Eq. And IQ. And we pair it with our tool, which can always be augmented with additional
263 00:26:57.570 ⇒ 00:26:59.849 Andrew Duplessie: parameters, tips, and tricks
264 00:26:59.990 ⇒ 00:27:21.000 Andrew Duplessie: all that stuff, and it’s like a hit making tool. And so like, that’s that’s like the outcome that I want. So essentially you, Tom, I don’t think we’re doing anything differently that you can’t just take a image off my journey and pull it into runway today. But it’s because we’re dressing it up, and we’re showing what the future could look like that. It has, you know, like that value
265 00:27:22.110 ⇒ 00:27:30.009 Andrew Duplessie: like, that’s what I think it. That’s what I think. It simply is. So I guess on the, for on a level of like a 1st step, it is that that ui that I showed which is like
266 00:27:30.660 ⇒ 00:27:32.599 Andrew Duplessie: up top, I can create
267 00:27:32.710 ⇒ 00:27:44.510 Andrew Duplessie: content and kind of like, how runway does their ui, it’s like on the left. You create your content. You can kind of drag it into video. It’s just like copy. We’ll copy that to be honest, but I don’t know if that’s even something that is doable or whatnot.
268 00:27:46.950 ⇒ 00:27:49.830 Uttam Kumaran: Yeah, I guess that’s where I’m almost like, tell me how
269 00:27:50.320 ⇒ 00:27:55.769 Uttam Kumaran: different this is than just using runway like, what is the
270 00:27:56.320 ⇒ 00:28:11.879 Uttam Kumaran: is the sauce in the company, the fact that you’re running and you’re forcing everyone to use it. Is it in the fact that you do need to maintain a custom workflow, because maybe the meet, the medium, or like the type of content, is different.
271 00:28:12.090 ⇒ 00:28:14.359 Uttam Kumaran: like, can you talk to that a little bit.
272 00:28:14.980 ⇒ 00:28:21.160 Andrew Duplessie: Yeah, we should debate this cause I’d rather come out with something good wrong. Feel free to debate.
273 00:28:21.160 ⇒ 00:28:23.920 Uttam Kumaran: Cause you can still achieve your outcome. And just like
274 00:28:24.110 ⇒ 00:28:27.060 Uttam Kumaran: you’re like we just use runway ever like.
275 00:28:27.060 ⇒ 00:28:28.580 Andrew Duplessie: Yeah, I, I.
276 00:28:28.580 ⇒ 00:28:30.000 Uttam Kumaran: That is one way.
277 00:28:30.000 ⇒ 00:28:32.770 Andrew Duplessie: For investors. I have to show it’s on a different platform.
278 00:28:32.770 ⇒ 00:28:35.460 Uttam Kumaran: Okay. Okay, then. Well, then, that’s that’s the debate.
279 00:28:35.460 ⇒ 00:28:40.390 Andrew Duplessie: That. Well, yeah, that’s that’s definitely like the 1st app. But just internally for us, like.
280 00:28:40.560 ⇒ 00:28:44.950 Andrew Duplessie: I think that. I think that it’s
281 00:28:46.650 ⇒ 00:29:10.510 Andrew Duplessie: we’re a studio that’s not known, for, like mass output of shit AI content, it’s more like we’re creating really curated tastemaker content. That’s like bluey rugrats Dexter, like I I was thinking the other day, like all these Ips are for sale, too. Right rocket power all these Ips. You can go and option them for a certain amount of money, and one of the investors called me is like dude like.
282 00:29:10.710 ⇒ 00:29:29.709 Andrew Duplessie: pick your favorite cartoon from the nineties or 2 thousands, and let’s just option it like, let’s just buy the IP, and that would be amazing, right? And that that could be like our approach. And maybe we pivot to that. But this is where I think it starts right. It’s like a an internal platform that creates like true hits. Does that make sense.
283 00:29:29.710 ⇒ 00:29:32.199 Uttam Kumaran: To me about like the like.
284 00:29:34.790 ⇒ 00:29:38.559 Uttam Kumaran: Like the balance sheet of like a studio like this, like what?
285 00:29:38.800 ⇒ 00:29:43.789 Uttam Kumaran: Where does cost? Where does what is cost? How does revenue come in to something like this?
286 00:29:44.380 ⇒ 00:29:50.579 Andrew Duplessie: I don’t think we’ll make any money at first, st like I think we have to figure out
287 00:29:50.790 ⇒ 00:29:53.009 Andrew Duplessie: if the content is like
288 00:29:53.950 ⇒ 00:30:02.860 Andrew Duplessie: even good, even good like at a at a bare minimum. So I don’t. I don’t see us making money, but I think down the road it’s like we create
289 00:30:03.050 ⇒ 00:30:07.500 Andrew Duplessie: the rugrats, and our distribution initially is on Tiktok.
290 00:30:07.800 ⇒ 00:30:12.449 Andrew Duplessie: We build up enough data that Netflix wants to do a deal with us
291 00:30:12.580 ⇒ 00:30:23.390 Andrew Duplessie: or a different streamer. There’s this. There’s hundreds of examples of this. Blue is one of them. They did it with Prime. There’s a reality show that’s about pop popping like a red balloon. It’s like a dating show. Netflix.
292 00:30:23.390 ⇒ 00:30:25.010 Uttam Kumaran: Yeah, yeah, so you, too, yeah.
293 00:30:25.010 ⇒ 00:30:36.909 Andrew Duplessie: That started on Tiktok. So you’re seeing all these shows start on Tiktok 1st and then build data and get pushed to the main streamers, but I think, like the end. The end goal for us is like
294 00:30:37.380 ⇒ 00:30:55.670 Andrew Duplessie: we’ve produced the bluey, the rugrat. We have like 8 or 9 shows that we’ve proven out. And we’re like, okay, like, we, we understand this, we really do. And then a studio comes in and says, Hey, like, this is our AI unit. These guys get it. They produce these hits. You know that that to me is like the dream scenario. In a year or 2.
295 00:30:55.670 ⇒ 00:31:03.730 Uttam Kumaran: So potential. So I guess what I’m asking also is like, you could accomplish that without AI. It’s just the
296 00:31:04.010 ⇒ 00:31:07.220 Uttam Kumaran: time and the capital needed would be yeah.
297 00:31:07.220 ⇒ 00:31:08.650 Andrew Duplessie: Oh, yeah. Oh, yeah.
298 00:31:08.650 ⇒ 00:31:13.879 adamabdulhamid: It’s just it’s just like a content studio with better unit economics like you can just do it faster and more cheaply.
299 00:31:13.880 ⇒ 00:31:14.359 Andrew Duplessie: That’s a good one.
300 00:31:14.360 ⇒ 00:31:29.339 adamabdulhamid: Because you, and because you can do both of those, you can make more. And then and then you can like, I think the the social aspect is kind of important, like we’re using, or one option that we don’t necessarily have to, but is to use social to like gauge.
301 00:31:29.340 ⇒ 00:31:29.940 Uttam Kumaran: Yeah.
302 00:31:29.940 ⇒ 00:31:33.480 adamabdulhamid: You know, consumer interest in the content itself. So you’re like we made it.
303 00:31:33.480 ⇒ 00:31:34.130 Uttam Kumaran: My focus.
304 00:31:34.130 ⇒ 00:31:34.740 adamabdulhamid: Good.
305 00:31:34.890 ⇒ 00:31:55.539 adamabdulhamid: Yeah. Instead of like a focus group or something you just like, use direct consumer feedback on, you know, socials, on Tiktok, on Instagram, or whatever and it’s like you could totally do this without AI. And you just need an army of producers with, like, you know, traditional animation, tools and and whatever sketch tools they use. It’s just slower and and more expensive.
306 00:31:57.150 ⇒ 00:32:18.430 adamabdulhamid: And one other thought I had earlier about like, Okay, you know, you’re like, what’s the secret sauce? Why not just use runway? I I agree, Andrew, you mentioned like, probably you just need this for like investor purposes, to be like, okay, we have our own, you know, sort of tool. And and there’s some secrets off there. The other thing that I think is I don’t necessarily know how important this will be, but I do think it’s a nice property. Is that?
307 00:32:18.960 ⇒ 00:32:48.430 adamabdulhamid: If you build your own, you can like use all the different Apis like it lets you access all the models and potentially even open source models. Or, like, you know, maybe there’s a great open Source image generation model that is like, you know, 1 10th the cost of like, you know. You know, Gpt 4 or whatever. And so you know, like you can let the producers, or, like the tool itself, say, like, Okay, for these kinds of queries, we’ll use these models, whatever you know. Kind of like cursor is in that sense.
308 00:32:48.880 ⇒ 00:33:08.079 Andrew Duplessie: That’s actually a good idea. It’s like. And it’s also something that could sell an investor meeting. If if I can put in a prompt, and then you show me clean runway mid journey. If there’s like, if there’s a feature we can create where I see all 6 outputs at the same time, and investor be like, Oh, my God, I get it.
309 00:33:08.080 ⇒ 00:33:09.920 adamabdulhamid: Exactly that. That’s what I was talking about. When I’m like
310 00:33:10.360 ⇒ 00:33:34.980 adamabdulhamid: the like, the orchestration aspect. I think you want to see like what. And then you you want to see like, okay, I use this prompt. And here was like the cling output. Here’s the mid journey output. Here’s the Vo 2 output, and you can see like, Oh, I love vo, 2 style so like you. You can like, say, like, I want this one, and you can like. See them all, though it’s like somehow, like managed in this platform. So you can kind of like see the arc of the the content generation process.
311 00:33:36.980 ⇒ 00:33:43.299 Uttam Kumaran: Okay, I feel pretty. I feel better now. It seems like you’re focused on, at least on animation.
312 00:33:43.880 ⇒ 00:33:44.590 Uttam Kumaran: It’s like.
313 00:33:44.590 ⇒ 00:33:47.030 Andrew Duplessie: Kids the angle like kids content.
314 00:33:47.030 ⇒ 00:33:56.440 Andrew Duplessie: The only the only reason I say that Tom, is because it seems like the quality coming out of these generators is not good enough for adult eyes. But maybe kids, eyes.
315 00:33:56.940 ⇒ 00:34:00.550 Uttam Kumaran: So like, I know, if I have to. I’m I’m doing. I’m looking at this in the reverse way.
316 00:34:00.550 ⇒ 00:34:01.529 Uttam Kumaran: I see what you mean.
317 00:34:01.530 ⇒ 00:34:08.630 Andrew Duplessie: Yeah. Like, if Adam hands me a piece of content, I’m thinking what I can get views on believable views, organic quality views. And it’s like
318 00:34:08.940 ⇒ 00:34:12.980 Andrew Duplessie: copying disney fireworks. Right? It’s like shit that kids watch.
319 00:34:12.989 ⇒ 00:34:13.559 Uttam Kumaran: Sure.
320 00:34:13.560 ⇒ 00:34:17.399 Andrew Duplessie: That’s just my assumption to start. I don’t wanna be in the business of kids content. But it’s.
321 00:34:17.400 ⇒ 00:34:25.340 Uttam Kumaran: No, but I don’t think that I don’t think like this tool necessarily like gets you like puts you in like, use the animation focus whatever.
322 00:34:26.170 ⇒ 00:34:28.890 Uttam Kumaran: Just think, I’m trying to. Basically
323 00:34:29.050 ⇒ 00:34:32.899 Uttam Kumaran: they, if we were to do like a 4 or 6 week sprint
324 00:34:33.050 ⇒ 00:34:42.950 Uttam Kumaran: like, what can we accomplish that ticks off as many of your boxes, and at least like either, shut some directions off or like reinforces a few.
325 00:34:43.459 ⇒ 00:34:58.640 Uttam Kumaran: I think if we consider the sort of animations that you mentioned. I hear you on a little bit of like having the seed. There’s also some like this paper that came out recently about how they this team
326 00:34:59.410 ⇒ 00:35:06.149 Uttam Kumaran: generated like a full, like a basically a minute of fully synthetic Tom and Jerry, like an episode that really like.
327 00:35:06.810 ⇒ 00:35:07.150 adamabdulhamid: Yeah.
328 00:35:07.150 ⇒ 00:35:07.530 Andrew Duplessie: Yes.
329 00:35:07.530 ⇒ 00:35:11.730 Uttam Kumaran: Yeah, like that was great. That’s like, what’s in yeah.
330 00:35:11.730 ⇒ 00:35:14.790 Andrew Duplessie: Sorry to interrupt. You give me an idea, but like that makes me think like
331 00:35:15.380 ⇒ 00:35:24.289 Andrew Duplessie: we should start with someone who draws those characters like a human, and then we use that base drawing to extrapolate scenes, and shit like that.
332 00:35:24.290 ⇒ 00:35:28.470 Uttam Kumaran: Yeah, I’ll base. I’m just gonna go re. I’ll go read the paper again and try to remember what like, what.
333 00:35:28.470 ⇒ 00:35:28.820 Andrew Duplessie: Yeah.
334 00:35:28.820 ⇒ 00:35:32.790 Uttam Kumaran: Was the other thing in terms in terms of like.
335 00:35:32.980 ⇒ 00:35:35.309 Uttam Kumaran: I mean, I think certainly we should
336 00:35:36.120 ⇒ 00:35:40.080 Uttam Kumaran: go through like that, like maybe even. It’s just like a
337 00:35:40.550 ⇒ 00:35:44.270 Uttam Kumaran: few hours, or like a half day where we break down
338 00:35:44.530 ⇒ 00:35:49.369 Uttam Kumaran: like the production process, because I don’t think it’s worth.
339 00:35:50.160 ⇒ 00:36:08.150 Uttam Kumaran: This is like a it’s very luckily this is like a very, it’s like a well defined manufacturing process. But I just wanna make sure we see that. And we know what pieces we’re going after, and it’ll help you prioritize if you need to increase the scope of the tool. And then, in terms of like execution. Yeah, do the people I’m working with.
340 00:36:08.885 ⇒ 00:36:16.020 Uttam Kumaran: I just called him yesterday, and they they are. They’re working for. They’re doing some AI work for riot.
341 00:36:17.040 ⇒ 00:36:17.820 Andrew Duplessie: My games.
342 00:36:17.820 ⇒ 00:36:27.320 Uttam Kumaran: Yeah, they’re doing some AI animation work for riot like 3D. 3D. Rendering and stuff for their games. Team and then one other
343 00:36:27.610 ⇒ 00:36:31.206 Uttam Kumaran: other big logos. So like good people,
344 00:36:32.070 ⇒ 00:36:45.630 Uttam Kumaran: like, we’re partnered with them there. I just like they were. They’ve done sort of this set of working creative. And we have, like 2 folks internally, that I’ve done this, too. But I want to time box this to like something where it’s like
345 00:36:46.170 ⇒ 00:36:48.610 Uttam Kumaran: we work backwards from an outcome.
346 00:36:48.810 ⇒ 00:36:49.250 Andrew Duplessie: Yeah.
347 00:36:49.250 ⇒ 00:36:54.499 Uttam Kumaran: And the priority, like, I think the ui. If your if your aim is like.
348 00:36:55.320 ⇒ 00:36:59.860 Uttam Kumaran: have something that’s like ui ready by like
349 00:37:00.390 ⇒ 00:37:04.779 Uttam Kumaran: start of August that that’s like what May, June July
350 00:37:05.250 ⇒ 00:37:12.959 Uttam Kumaran: like. That’s pretty good, and I don’t think I think you can whip up whatever ui you want like. You know, it’s pretty pretty easy like in a week.
351 00:37:13.920 ⇒ 00:37:19.190 Uttam Kumaran: Ivana, or even our people internally like, can can slap a ui onto whatever endpoint want.
352 00:37:19.330 ⇒ 00:37:21.289 Uttam Kumaran: So I’m not.
353 00:37:21.750 ⇒ 00:37:28.840 Uttam Kumaran: I don’t really worry about that, especially if it’s like you’re gonna control it. And like, we only have to generate a few have a few screens.
354 00:37:29.180 ⇒ 00:37:30.000 Andrew Duplessie: Yeah.
355 00:37:30.000 ⇒ 00:37:32.239 Uttam Kumaran: I’m more like, how can I give you enough
356 00:37:32.890 ⇒ 00:37:36.570 Uttam Kumaran: evidence that it’s that it’s like gonna work or not? You know.
357 00:37:36.570 ⇒ 00:37:45.959 Andrew Duplessie: I mean, it’s we probably should look back at like what’s working online and wise. It’s like that horror one that I always reference right?
358 00:37:45.960 ⇒ 00:37:47.489 Uttam Kumaran: 4, 1, though.
359 00:37:47.670 ⇒ 00:37:52.380 Andrew Duplessie: The eighties, one on Youtube, where this one, like.
360 00:37:52.380 ⇒ 00:37:58.850 Uttam Kumaran: Anything that’s like involving like politics, or sort of like Pop culture.
361 00:37:58.850 ⇒ 00:37:59.220 Andrew Duplessie: But this.
362 00:37:59.220 ⇒ 00:38:01.730 Uttam Kumaran: The style is horrible, the style is not really good.
363 00:38:01.730 ⇒ 00:38:02.540 Andrew Duplessie: No, no.
364 00:38:02.540 ⇒ 00:38:05.609 Uttam Kumaran: Same like kind of blob stuff. I don’t even know how to describe it, but.
365 00:38:05.610 ⇒ 00:38:19.056 Andrew Duplessie: Vibes. Vibes work right like samurai over buildings with music like we. We almost you make a good point, which is like we almost need to arrive at like, what like are we creating this or that? And
366 00:38:19.410 ⇒ 00:38:21.330 Uttam Kumaran: I think your opinion on like
367 00:38:22.140 ⇒ 00:38:27.230 Uttam Kumaran: testing short form. And then basically like moving.
368 00:38:27.620 ⇒ 00:38:29.680 Uttam Kumaran: I think that’s fair. I also just think.
369 00:38:31.550 ⇒ 00:38:39.799 Uttam Kumaran: I think you just need to decide. Once you see, the entire manufacturing process, you’re like, if 60% of this can be augmented by AI, are we happy?
370 00:38:40.100 ⇒ 00:38:43.399 Uttam Kumaran: Right? Cause. The mind is always gonna take you to say, like end to end.
371 00:38:43.530 ⇒ 00:38:49.619 Uttam Kumaran: prompt to like show, like push away from that, and just think about like
372 00:38:50.200 ⇒ 00:38:54.920 Uttam Kumaran: if I, if we just took care of like 20% of it, is that enough?
373 00:38:56.070 ⇒ 00:39:00.110 Uttam Kumaran: And again, if you’re raising on like the business model, or if you’re raising on the
374 00:39:00.450 ⇒ 00:39:10.269 Uttam Kumaran: then is that enough to get you there? Because 100 deal is the same problem that all these companies are going after, like one of the reasons to not use runway is because they’re gonna they’re trying to serve like
375 00:39:10.380 ⇒ 00:39:12.460 Uttam Kumaran: an extremely wide audience.
376 00:39:12.680 ⇒ 00:39:13.170 Andrew Duplessie: Yeah.
377 00:39:13.170 ⇒ 00:39:16.350 Uttam Kumaran: And really enter. And probably I think, more enterprise.
378 00:39:16.530 ⇒ 00:39:17.470 Uttam Kumaran: So
379 00:39:18.290 ⇒ 00:39:32.909 Uttam Kumaran: there’s gonna be limitations and costs associated. So if you can think of just like, Hey, for example, pretty class, if you’re like in the manufacturing process, there’s a hundred steps. 30 of them take up 80% of the time. So we’re attacking half of those.
380 00:39:33.270 ⇒ 00:39:33.590 Andrew Duplessie: It’s like.
381 00:39:33.590 ⇒ 00:39:35.440 Uttam Kumaran: The prioritization process.
382 00:39:36.320 ⇒ 00:39:38.070 Uttam Kumaran: That would be helpful.
383 00:39:38.857 ⇒ 00:39:45.270 Uttam Kumaran: But he also I I also hear you on the Ui needs to somewhat show the end to end
384 00:39:46.600 ⇒ 00:39:48.320 Uttam Kumaran: process with a human in loop.
385 00:39:48.620 ⇒ 00:39:49.420 Uttam Kumaran: Yeah.
386 00:39:50.010 ⇒ 00:39:55.799 Andrew Duplessie: Yeah, it’s like, I, I can like, make a visual of like what I think
387 00:39:56.000 ⇒ 00:40:04.509 Andrew Duplessie: like, does it start with a human artist that like sketches a scene and figure out like, what’s the weight on AI versus a human
388 00:40:05.600 ⇒ 00:40:09.359 Andrew Duplessie: because it could be like we use the capital
389 00:40:10.170 ⇒ 00:40:18.370 Andrew Duplessie: partially in the traditional sense of creating animation right, which is essentially a lot of drawings. And this is what we did
390 00:40:18.720 ⇒ 00:40:24.249 Andrew Duplessie: to extend that right, like a scene, or the background, or whatever to save time like, we could demonstrate that.
391 00:40:24.460 ⇒ 00:40:28.269 Andrew Duplessie: But I think the more the most powerful thing that we could come away with this is is.
392 00:40:28.620 ⇒ 00:40:30.469 Andrew Duplessie: I put this on Tiktok. It has.
393 00:40:30.470 ⇒ 00:40:33.570 Andrew Duplessie: Yeah, yeah, I weren’t followers. Each one has a million views.
394 00:40:34.220 ⇒ 00:40:37.569 Andrew Duplessie: and it was mostly I like, that’s the shit. I think that will
395 00:40:37.990 ⇒ 00:40:42.499 Andrew Duplessie: blow my mind blow people’s minds, but I don’t know if that’s I don’t know if I don’t know.
396 00:40:42.960 ⇒ 00:40:54.509 Andrew Duplessie: Yeah, I don’t know if it’s like we sitting here right now, we’re like, Oh, we’re gonna tell a story about a baby skeleton that becomes an adult skeleton, and we’re gonna fall, you know, like, what dumb shit are we coming up with?
397 00:40:54.840 ⇒ 00:40:57.036 Andrew Duplessie: I don’t know. Have you guys seen that that
398 00:40:58.190 ⇒ 00:41:02.469 Andrew Duplessie: the short film with the Fox. I think I sent it to both. You.
399 00:41:02.630 ⇒ 00:41:04.989 Uttam Kumaran: I think some something won like an award. Right recently.
400 00:41:05.330 ⇒ 00:41:08.740 Andrew Duplessie: Yeah, one like a film festival, or some show.
401 00:41:08.740 ⇒ 00:41:19.519 adamabdulhamid: I I mean, I do. I do kind of feel like that’s like the right thing to start with is like, pick a specific thing I mean. And, Andrew, you, you definitely will know the like, the production process the best here. But I feel like
402 00:41:19.750 ⇒ 00:41:37.230 adamabdulhamid: I feel like I remember I was talking about this when we last chatted a couple weeks ago, or whatever like, I think we should just pick like some simple idea. It could even be like the same Tom and Jerry idea like just we want. We’re gonna write like a very short like you know, Storyline, for like a 60 second Tom and Jerry one.
403 00:41:37.320 ⇒ 00:42:04.540 adamabdulhamid: and then, just like, see what the process would be like of leveraging AI tools to generate it like. Go sign up for runway sign up for, like, you know, Chatgpt pro or whatever, and just like start just like using the tools like there’s been like several twitter threads on like exactly the workflow these people are using to actually go generate the content as well like that Lord of the Rings one you sent, Andrew. I think the guy like even like detailed, he was like, here’s exactly what I did. I would use this tool to generate the image, and I would take.
404 00:42:04.540 ⇒ 00:42:05.010 Andrew Duplessie: Yeah. Man.
405 00:42:05.010 ⇒ 00:42:26.769 adamabdulhamid: And I put it in this other tool, like, I think, just like going to do that, for, like, you know. 1 1 example, then, I think, will give us way. More understanding of like what we want like, where is all the time spent, or like the majority of the time spent, which steps are like really ripe for you know. Ai like augmentation, or even automation, or whatever.
406 00:42:27.460 ⇒ 00:42:34.219 Andrew Duplessie: I. I agree with that. And I think this like, for example, like there’s so many like, Do this one. This one’s already looking better.
407 00:42:34.845 ⇒ 00:42:39.769 Andrew Duplessie: You could do like we could just do shit like this, and a voice over, and that could be our thing, and it could.
408 00:42:39.770 ⇒ 00:42:40.570 adamabdulhamid: Yeah. Totally.
409 00:42:40.570 ⇒ 00:42:41.310 Andrew Duplessie: Every time.
410 00:42:41.550 ⇒ 00:42:45.350 Andrew Duplessie: You know, this is like a robot in a car, different.
411 00:42:45.700 ⇒ 00:42:46.840 Andrew Duplessie: But like I mean, I feel like.
412 00:42:46.840 ⇒ 00:42:51.940 adamabdulhamid: If you like, from my perspective, like if you were, if if I’m an investor.
413 00:42:52.430 ⇒ 00:43:09.159 adamabdulhamid: And someone came to me and said, Hey, like we made these 10, you know, clips and it and it took us all in, you know, for producer time, and like AI tool costs whatever it costs. I don’t know X dollars, and previously it would have cost 2 x dollars, or 5 x dollars, or 10 x dollars.
414 00:43:09.160 ⇒ 00:43:09.520 Andrew Duplessie: Yeah.
415 00:43:09.520 ⇒ 00:43:14.590 adamabdulhamid: Like that’s very compelling, even if they don’t get like tons of views on Tiktok. Like as long as the quality is like above, something.
416 00:43:14.590 ⇒ 00:43:15.020 Andrew Duplessie: Okay.
417 00:43:15.020 ⇒ 00:43:28.919 adamabdulhamid: You’re like, Oh, it’s kind of cool. Then I think they start to see like, oh, actually like they’re generating this content for way more cheaply. And you know, you kind of like. See in your mind, like the idea of flywheel being like very, very open.
418 00:43:29.380 ⇒ 00:43:35.720 Uttam Kumaran: Yeah, that’s also I I agree, like, are you competing on costs? Are you competing on iteration cycles? Because
419 00:43:36.360 ⇒ 00:43:38.000 Uttam Kumaran: you’re not? I can’t like
420 00:43:38.270 ⇒ 00:43:42.699 Uttam Kumaran: the product is not gonna help you. I mean, I don’t know. Maybe you can use it in ideation. But like.
421 00:43:43.350 ⇒ 00:43:45.410 Uttam Kumaran: that’s the sort of thing is like.
422 00:43:45.750 ⇒ 00:43:48.009 Uttam Kumaran: yeah, this is, this is great.
423 00:43:48.120 ⇒ 00:43:56.846 Uttam Kumaran: This should that this could see anything shows you that it’s possible. And if you read the description the way this guy did. It is like really rough
424 00:43:57.470 ⇒ 00:43:57.910 Andrew Duplessie: Yeah.
425 00:43:57.910 ⇒ 00:43:59.139 Uttam Kumaran: And but like
426 00:44:00.400 ⇒ 00:44:06.370 Uttam Kumaran: I don’t know, it’s a smooth video about a fox like you know what I mean like, that’s they’re still like lacking some taste.
427 00:44:06.370 ⇒ 00:44:06.740 Andrew Duplessie: Okay.
428 00:44:06.740 ⇒ 00:44:09.850 Uttam Kumaran: Whatever distribution, basically a taste and distribution.
429 00:44:10.320 ⇒ 00:44:14.730 Uttam Kumaran: So that like, that’s something where
430 00:44:15.950 ⇒ 00:44:26.910 Uttam Kumaran: there there are like AI ways to help you even identify distribution and stuff like that, or help you ideate. But that’s why it’s helpful to know, like, what part of this is augmented is the pitch like, hey, we can.
431 00:44:27.080 ⇒ 00:44:35.789 Uttam Kumaran: We can test like 5 times more short things, or the cost to test is like way lower both of those the same thing.
432 00:44:36.406 ⇒ 00:44:40.400 Uttam Kumaran: And then we then move to full scale production.
433 00:44:40.911 ⇒ 00:44:47.979 Uttam Kumaran: Saving like X amount of things that like should have been tested that couldn’t got could that didn’t get tested because we didn’t have time like.
434 00:44:48.600 ⇒ 00:44:51.910 Uttam Kumaran: is it like time to value? Yeah.
435 00:44:51.910 ⇒ 00:44:57.060 Andrew Duplessie: Yeah, no, I think that’s right. I think we’re like the bridge between
436 00:44:57.350 ⇒ 00:45:04.579 Andrew Duplessie: fully just pushing scripts into like an AI machine. Right? Like, I think our company could say.
437 00:45:05.070 ⇒ 00:45:17.374 Andrew Duplessie: producer, bring your script and your story over here, and we’ll kind of help guide you. Create it. Blah blah! But I think our company will be basically judged on like what we release and how well it does.
438 00:45:18.450 ⇒ 00:45:19.960 Andrew Duplessie: But I think it’s like.
439 00:45:21.130 ⇒ 00:45:24.980 Andrew Duplessie: I think we can get more specific, right? Uton, because that’s really what you’re looking for from us.
440 00:45:25.520 ⇒ 00:45:36.549 Andrew Duplessie: Like, if you’re being super direct to Adam. And I like you want the tools we want like specific things that you can then push into a roadmap that you’ll send back to us and be like it costs X.
441 00:45:36.550 ⇒ 00:45:38.821 Uttam Kumaran: Yeah, less less of even like
442 00:45:40.090 ⇒ 00:45:48.180 Uttam Kumaran: the the tools cause, like meaning less of even like what video model or whatever like, that’s fine.
443 00:45:48.360 ⇒ 00:45:50.039 Uttam Kumaran: I’m just trying, almost like.
444 00:45:51.530 ⇒ 00:45:52.770 adamabdulhamid: Like, what’s the goal?
445 00:45:52.770 ⇒ 00:45:55.019 Uttam Kumaran: The 1st milestone, like 1st Basic. If.
446 00:45:55.020 ⇒ 00:45:55.360 Andrew Duplessie: There is.
447 00:45:55.360 ⇒ 00:45:57.539 Uttam Kumaran: Okay, okay, 4 to 6 weeks.
448 00:45:57.920 ⇒ 00:46:09.119 Uttam Kumaran: We want to see a ui that does this and then that allow. If that if we get there, then that unlocks us to go one step deeper here, if in 4 to 6 weeks it’s like not.
449 00:46:09.390 ⇒ 00:46:16.180 Uttam Kumaran: it doesn’t do certain things, or it may open up a couple of things. I just don’t. Wanna I just wanna really time box it because I think you’ll.
450 00:46:16.180 ⇒ 00:46:16.780 adamabdulhamid: Yeah.
451 00:46:16.780 ⇒ 00:46:20.270 Uttam Kumaran: To something that’s like takes in an image produce a little video.
452 00:46:20.470 ⇒ 00:46:27.350 Uttam Kumaran: But like, pretty quickly. So I want to think about like, Yeah, ex- exactly.
453 00:46:28.160 ⇒ 00:46:33.160 adamabdulhamid: I think, yeah, I guess, like, from my perspective, I think the
454 00:46:33.370 ⇒ 00:46:43.460 adamabdulhamid: the like we’ve been emphasizing animated content and also short content. I think those really are just because they’re like current model limitations like.
455 00:46:43.900 ⇒ 00:47:05.870 adamabdulhamid: You know those. Those are the likely. The things that likely will lead to the best outcomes are shorter, content, animated, content, you know. Given the current AI systems. So that’s it’s not like a fundamental like, you know, product decision. Like, we only want to focus on those things. It’s more like, let’s use those as the initial prototype version, because they seem like the highest likelihood of success.
456 00:47:07.470 ⇒ 00:47:31.629 adamabdulhamid: yeah. So that that’s I think, why. And then I I see your question. It’s like, is the goal to just like pump out a ton of content that you could test on socials and like figure out how to monetize that somehow, or is it like eventually to say, Okay, we confirm this one is good. And now let’s go make a like a, you know, full 30 min episode or something, and that I don’t quite know. I don’t know. If, like Andrew, you probably have a better idea. I don’t know if it’s like.
457 00:47:31.680 ⇒ 00:47:50.930 adamabdulhamid: okay, let’s say you had a little short, or something, or like a few shorts in the same, you know, for the same characters that like do well on socials. Is that the kind of thing you can go to a studio and say, like, we want to sell you the rights to this to go make a full show, or you partner with them, or something like that, that I’m not sure exactly how how to like you.
458 00:47:50.930 ⇒ 00:47:51.290 Andrew Duplessie: Yeah.
459 00:47:51.290 ⇒ 00:47:52.040 adamabdulhamid: Test, that.
460 00:47:52.210 ⇒ 00:48:09.950 Andrew Duplessie: Yeah, it’s a good question. I think part of it is like, there is this model called the Overall deal in the studio business. Right. And you go to Steven Spielberg. And you’re like we’re giving you an overall deal. It’s gonna be a hundred 1 million, and we we get the 1st look of any project you create. And I think that model is sort of been
461 00:48:10.060 ⇒ 00:48:20.800 Andrew Duplessie: effed up because of the potential here. So part of what we’re building is like pure hype. Like I, I do think we could sell this just on hype to a studio, because, like, these guys are on the right path.
462 00:48:20.870 ⇒ 00:48:44.010 Andrew Duplessie: and they put out a piece of content that worked like we acquire them and move on like. That’s what I think is happening. That’s kind of how I’m setting up in my head. And that’s very common in the studio business to just acquire humans for time with with a script or even an idea. So I think, like, that’s that’s the one side. The other side is like.
463 00:48:44.150 ⇒ 00:48:51.738 Andrew Duplessie: when you look at some of this content that kids are watching these days. I mean, it’s actually idiotic like. I don’t know if you’ve heard of skibbity toilet.
464 00:48:52.670 ⇒ 00:48:54.180 Uttam Kumaran: Yes, of course, of course.
465 00:48:54.180 ⇒ 00:48:58.079 Andrew Duplessie: Yeah. So you know, yeah, like, my friend’s son watches this shit.
466 00:48:58.080 ⇒ 00:49:00.740 Uttam Kumaran: Really don’t understand it, though that’s like really the extent.
467 00:49:00.740 ⇒ 00:49:01.620 Andrew Duplessie: I mean, look at.
468 00:49:01.620 ⇒ 00:49:02.109 Uttam Kumaran: Like that’s.
469 00:49:04.590 ⇒ 00:49:05.950 adamabdulhamid: Is this AI generated.
470 00:49:06.390 ⇒ 00:49:07.590 Andrew Duplessie: It’s not.
471 00:49:07.740 ⇒ 00:49:11.990 Uttam Kumaran: It’s not, but it’s the it’s the highest viewed, animated.
472 00:49:12.240 ⇒ 00:49:14.429 Andrew Duplessie: This is 25 million views. It’s just.
473 00:49:14.430 ⇒ 00:49:14.780 adamabdulhamid: This like.
474 00:49:14.780 ⇒ 00:49:15.480 Uttam Kumaran: Basically.
475 00:49:15.660 ⇒ 00:49:20.519 adamabdulhamid: Is this like a. And they like keep putting out new episodes or something. Or is this like just like a 1 time like.
476 00:49:20.520 ⇒ 00:49:22.699 Uttam Kumaran: No, they put it one out like every week.
477 00:49:23.690 ⇒ 00:49:24.149 adamabdulhamid: And there’s no.
478 00:49:24.150 ⇒ 00:49:26.200 Uttam Kumaran: There’s like no dialogue or anything.
479 00:49:26.200 ⇒ 00:49:29.290 Andrew Duplessie: There’s no dialogue in some episodes like 3 min like.
480 00:49:29.520 ⇒ 00:49:35.000 Andrew Duplessie: and this they, by the way, they just did a deal with James Cameron for a feature. They have toys and target.
481 00:49:36.390 ⇒ 00:49:41.039 Andrew Duplessie: This IP has been distilled across America, and it looks like absolute shit.
482 00:49:42.030 ⇒ 00:49:44.620 Andrew Duplessie: It’s a Pov number.
483 00:49:44.620 ⇒ 00:49:46.990 Uttam Kumaran: Don’t, don’t even dissect it. Dude. I don’t know.
484 00:49:48.359 ⇒ 00:49:51.099 Andrew Duplessie: Toilet, boy.
485 00:49:52.360 ⇒ 00:49:53.360 Andrew Duplessie: There you go.
486 00:49:54.180 ⇒ 00:50:04.730 Andrew Duplessie: It’s all over Target Walmart, gamestop, whatever. So like. If if we, if we create anything close to this like home run. We’ve won over
487 00:50:06.000 ⇒ 00:50:06.560 Andrew Duplessie: because.
488 00:50:06.560 ⇒ 00:50:07.590 adamabdulhamid: Oh, yeah, that’s that’s.
489 00:50:07.590 ⇒ 00:50:10.110 Andrew Duplessie: The side. The assumption on the finance side is
490 00:50:10.480 ⇒ 00:50:15.190 Andrew Duplessie: they made one with an AI tool. They’re gonna be able to make 50 for no cost like.
491 00:50:17.810 ⇒ 00:50:21.540 adamabdulhamid: Yeah, then, then, I think, like for me, the 1st milestone is just like.
492 00:50:21.780 ⇒ 00:50:27.869 adamabdulhamid: Can you get some reasonable looking like non trivially short, and by non trivially short, I mean, like.
493 00:50:28.420 ⇒ 00:50:39.820 adamabdulhamid: even like 30 seconds or 60 seconds. Like, can you get like 30 to 60? Second little content pieces that are like, somewhat coherent, like as coherent as skivity, toilet, and and.
494 00:50:39.820 ⇒ 00:50:40.160 Andrew Duplessie: No.
495 00:50:40.160 ⇒ 00:50:42.749 adamabdulhamid: Like some, you know, like that that. And then you can like
496 00:50:42.970 ⇒ 00:51:01.160 adamabdulhamid: post them on socials and just like, get that flywheel going a little bit, and then I think, the from an investor perspective, you can also then say it’d be useful to have a like a relative understanding of the unit economics as well to say like this, cost whatever X dollars again and like it would have cost, you know, 10 x dollars or something.
497 00:51:01.400 ⇒ 00:51:04.613 Andrew Duplessie: Yes, I love that. I think that’s right. And
498 00:51:05.450 ⇒ 00:51:17.910 Andrew Duplessie: yeah, and I would look at the platform really as insulation. I think that’s what you’re getting at, Tom is like, well, you could just do this on runway, and we could just say, like, we made this show, the platform to me is insulation and the vehicle to raise
499 00:51:18.580 ⇒ 00:51:26.860 Andrew Duplessie: else it’s they need something visual. They they don’t think very deeply or long about this shit. So that’s that’s what I’m thinking.
500 00:51:27.640 ⇒ 00:51:31.149 Andrew Duplessie: So what? What are we next steps? Wise, you sound, what do we?
501 00:51:31.690 ⇒ 00:51:35.560 Andrew Duplessie: What do we? What do you need from us? And what what should we expect from you?
502 00:51:35.560 ⇒ 00:51:40.227 Uttam Kumaran: I think I have enough. I think what I’m gonna do is basically like, yeah, come back to you with like, what
503 00:51:40.830 ⇒ 00:51:41.180 Andrew Duplessie: Like a room.
504 00:51:41.180 ⇒ 00:51:43.020 Uttam Kumaran: 4 to 6 weeks. Spring could look like.
505 00:51:43.020 ⇒ 00:51:44.300 Andrew Duplessie: Yeah, yeah, perfect.
506 00:51:44.300 ⇒ 00:51:47.909 Uttam Kumaran: We’ll we’ll we’ll do a like. I’ll plan out a longer
507 00:51:48.140 ⇒ 00:51:54.569 Uttam Kumaran: sort of roadmap. But the 1st milestone will basically be exactly like, reasonably looking non trivial, short.
508 00:51:55.897 ⇒ 00:52:00.009 Uttam Kumaran: video, I think I’ll probably make some assumptions on
509 00:52:00.748 ⇒ 00:52:06.670 Uttam Kumaran: like, I’ll speak to my guys and then talk a little bit about like, what inputs they expect
510 00:52:06.800 ⇒ 00:52:11.710 Uttam Kumaran: and what outputs you can expect. And then we can then debate like, how much
511 00:52:12.060 ⇒ 00:52:15.828 Uttam Kumaran: understand? You need to need need to know about what the final output needs to look like.
512 00:52:16.050 ⇒ 00:52:16.490 Andrew Duplessie: Yeah.
513 00:52:16.490 ⇒ 00:52:24.410 Uttam Kumaran: To get there right? Because this isn’t just like, create me this thing. And then it generates everything. There will be some structured inputs that’ll be the workflow.
514 00:52:24.790 ⇒ 00:52:30.160 Uttam Kumaran: Right? That’ll be the stuff that’s like augmenting humans that’s not completely taken care of.
515 00:52:30.610 ⇒ 00:52:33.710 Uttam Kumaran: But let me take this back, and then I’ll tell you what’s possible.
516 00:52:34.230 ⇒ 00:52:34.910 Andrew Duplessie: Okay.
517 00:52:34.910 ⇒ 00:52:38.150 Uttam Kumaran: In parallel, I think after this 1st milestone.
518 00:52:38.390 ⇒ 00:52:42.910 Uttam Kumaran: you should also, we should hopefully have a decision on like what the
519 00:52:43.390 ⇒ 00:52:48.839 Uttam Kumaran: Ui is, gonna look like for the raise. And then that way, you can just
520 00:52:49.520 ⇒ 00:52:53.969 Uttam Kumaran: basically work on that work on the deck in parallel of like, whatever’s next.
521 00:52:54.370 ⇒ 00:53:00.440 Uttam Kumaran: And then, yeah, I just think probably one of the weeks.
522 00:53:00.920 ⇒ 00:53:05.420 Uttam Kumaran: like, probably the 1st week, we’ll just be working to break down step by step by step.
523 00:53:05.690 ⇒ 00:53:07.829 Uttam Kumaran: How if you were to go from
524 00:53:08.140 ⇒ 00:53:15.319 Uttam Kumaran: nothing to this like 30 or 60 second animation, what does the process look like now? Like everything involved, from
525 00:53:15.480 ⇒ 00:53:19.179 Uttam Kumaran: creative ideation to drawing to, then stitching together to editing.
526 00:53:19.380 ⇒ 00:53:19.950 Andrew Duplessie: Yeah.
527 00:53:19.950 ⇒ 00:53:24.388 Uttam Kumaran: And then we’ll just we’ll cross off. So I have some frameworks to sort of do, do that sort of
528 00:53:25.330 ⇒ 00:53:26.830 Uttam Kumaran: like workshop.
529 00:53:29.570 ⇒ 00:53:35.500 Uttam Kumaran: And yeah, I don’t think any of any of what you said is impossible. I just think, the other thing you can bank on a lot
530 00:53:35.610 ⇒ 00:53:38.840 Uttam Kumaran: is that this stuff is getting cheaper and faster like every 3 months or so.
531 00:53:39.455 ⇒ 00:53:40.069 Andrew Duplessie: Okay.
532 00:53:40.070 ⇒ 00:53:43.020 Uttam Kumaran: So, even if you do something
533 00:53:43.560 ⇒ 00:53:46.260 Uttam Kumaran: now, you can expect it to be like
534 00:53:46.550 ⇒ 00:53:49.164 Uttam Kumaran: way better and way cheaper.
535 00:53:50.470 ⇒ 00:53:52.850 Uttam Kumaran: like not too far in the future.
536 00:53:52.850 ⇒ 00:53:53.380 Andrew Duplessie: Yeah.
537 00:53:54.410 ⇒ 00:53:57.799 Uttam Kumaran: And again, I think your competitive advantage is just that
538 00:53:58.190 ⇒ 00:54:00.859 Uttam Kumaran: this is a closed off like ecosystem, like.
539 00:54:00.860 ⇒ 00:54:01.630 Andrew Duplessie: Yeah, yeah.
540 00:54:01.630 ⇒ 00:54:08.330 Uttam Kumaran: Just because someone’s on Youtube and gets a million views like doesn’t mean anything. They’re not in la, like they don’t know, you know. So
541 00:54:08.820 ⇒ 00:54:12.800 Uttam Kumaran: that’s what I think the different like, that’s where I would have almost like
542 00:54:13.820 ⇒ 00:54:18.330 Uttam Kumaran: I care a lot less about like the the tech like engineering piece.
543 00:54:18.440 ⇒ 00:54:32.069 Uttam Kumaran: More. I’m I’m like, can we get you something that works really well? And that does the job. Even if you’re you’re you’re seeing that all the these other people are doing these AI sort of videos. But either they’re just like video editors that are doing that. And there’s no like
544 00:54:32.180 ⇒ 00:54:50.210 Uttam Kumaran: broader plan for monetization or like actualizing those clips and anything, or they’re just making they’re just making like, cpm, so this is a 3rd thing, right? Like that’s that’s the pieces hopefully by now. Everybody you’ll pitch into will have seen one of these AI things
545 00:54:50.540 ⇒ 00:55:01.349 Uttam Kumaran: they will have known it’s like possible, but then they themselves would not have known like, Okay, how can you actually do this? It’s kind of like stuff you’re seeing on Linkedin, where it’s like, I built this AI like Sdr agent, that like
546 00:55:01.620 ⇒ 00:55:03.789 Uttam Kumaran: email, someone and does something. It’s like.
547 00:55:04.640 ⇒ 00:55:06.809 Uttam Kumaran: Okay, I know it’s possible. But, like
548 00:55:07.790 ⇒ 00:55:10.599 Uttam Kumaran: Jack can actually like, do it is the real gap right now.
549 00:55:10.730 ⇒ 00:55:11.320 Andrew Duplessie: Yeah.
550 00:55:11.320 ⇒ 00:55:19.510 Uttam Kumaran: Can you do it in a predictable manner like, it’s not just a random animation about like a random skeleton. It’s like an actual con like you can steer it.
551 00:55:21.093 ⇒ 00:55:25.410 Andrew Duplessie: So okay, it’s the thing to Dominic. Give me an idea which is like
552 00:55:25.630 ⇒ 00:55:42.209 Andrew Duplessie: you have like these, really good, like, you have these people who are really really good at telling stories, but they’re completely disconnected from the AI side right now, and we’re sort of courting them in through this platform and saying so like, if Adam and I go out and we hire 3 or 4 producers slash writers for this.
553 00:55:42.400 ⇒ 00:55:47.580 Andrew Duplessie: they’re gonna have no fucking idea how to use this tool. So you can imagine that’s like the process happening.
554 00:55:47.580 ⇒ 00:55:50.299 Uttam Kumaran: Of course, quartering, coaxing them into this.
555 00:55:50.300 ⇒ 00:55:51.040 Uttam Kumaran: That’s it.
556 00:55:51.040 ⇒ 00:55:51.860 Andrew Duplessie: Yeah.
557 00:55:51.860 ⇒ 00:56:02.400 Uttam Kumaran: Part of what you should describe is like. And I’m seeing this in like just every other type of company, is either the from the top down. The company has to like, really be pro
558 00:56:02.520 ⇒ 00:56:03.700 Uttam Kumaran: adopting AI.
559 00:56:04.300 ⇒ 00:56:13.470 Uttam Kumaran: Otherwise it’s like it kinda gets like a disease. People see it like that. Or they fear it. Basically. So part of your
560 00:56:13.750 ⇒ 00:56:17.239 Uttam Kumaran: opportunity is that you can basically dictate from the top down.
561 00:56:18.920 ⇒ 00:56:24.850 Uttam Kumaran: And again, even if your animators just use mid journey, they may see 10% improvement in lift.
562 00:56:24.850 ⇒ 00:56:28.870 Uttam Kumaran: Yeah, let us alone, like an internal tool that’s catered to them.
563 00:56:29.030 ⇒ 00:56:31.639 Uttam Kumaran: They may see 30 or 40% or 50%, right? So.
564 00:56:31.640 ⇒ 00:56:32.520 Andrew Duplessie: Yeah, yeah.
565 00:56:32.520 ⇒ 00:56:43.759 Uttam Kumaran: That’s the thing is like, that’s the competitive advantage is not just the fact that it exists. The technology, actually, the ability to do this stuff has been was here like a year ago. It’s just like the adoption is not going to happen.
566 00:56:43.760 ⇒ 00:56:44.710 Andrew Duplessie: Yeah, yeah.
567 00:56:44.710 ⇒ 00:56:45.920 Uttam Kumaran: There’s pushback.
568 00:56:46.210 ⇒ 00:56:52.989 Uttam Kumaran: even from very smart people, is like A is heavy like. It’s gonna be a few years before a firm
569 00:56:53.570 ⇒ 00:56:56.980 Uttam Kumaran: is completely built like around something like this.
570 00:56:56.980 ⇒ 00:56:57.360 Andrew Duplessie: Yeah.
571 00:56:57.360 ⇒ 00:56:58.919 Uttam Kumaran: Even in my business
572 00:56:59.250 ⇒ 00:57:16.810 Uttam Kumaran: we use AI every day. I still have people on the team that aren’t using it for like basic things like, and that’s very, very challenging. Whether they fear it. Whether it’s like kind of embarrassing is they don’t want to use it in public, or they don’t want to like. Try it whether they they tried it once, and the outcome wasn’t great like. There’s all these challenges.
573 00:57:17.190 ⇒ 00:57:22.999 Uttam Kumaran: So part of it is even just for you to have the vehicle to adopt it, because you’re like forcing it.
574 00:57:23.140 ⇒ 00:57:23.600 Andrew Duplessie: Yeah.
575 00:57:24.790 ⇒ 00:57:28.019 Andrew Duplessie: Yeah, like, a lot of people won’t have that leverage.
576 00:57:28.260 ⇒ 00:57:29.389 Andrew Duplessie: Yeah, it makes sense.
577 00:57:29.550 ⇒ 00:57:35.369 Uttam Kumaran: Yeah, okay, cool. So let me get back to you this week. I’ll just.
578 00:57:35.370 ⇒ 00:57:35.760 Andrew Duplessie: Dude.
579 00:57:35.760 ⇒ 00:57:43.280 Uttam Kumaran: I’ll send text in the in the group. If you have any other thoughts, just throw it in here and leave me a comment, or send me a voice, note, or whatever.
580 00:57:43.520 ⇒ 00:57:44.040 Andrew Duplessie: Perfect.
581 00:57:44.288 ⇒ 00:57:48.519 Uttam Kumaran: Yeah, I’m gonna aim for something that’s like time box of 4 to 6 weeks or so.
582 00:57:48.940 ⇒ 00:57:51.719 Uttam Kumaran: and then we can debate on scope from there.
583 00:57:51.720 ⇒ 00:57:55.734 Andrew Duplessie: Okay, cool, and I’ll send you. I’ll send some more stuff.
584 00:57:57.240 ⇒ 00:58:03.211 Andrew Duplessie: in the group. But thank you. I’m excited. Yeah, I think this will be fun, at least a learning experience.
585 00:58:03.510 ⇒ 00:58:09.742 Uttam Kumaran: Yeah, I’m I’m pumped. This is really cool, like, I mean, you can see. I mean, we’ve been talking for a while. A lot of people are doing in their
586 00:58:09.950 ⇒ 00:58:10.840 Andrew Duplessie: No.
587 00:58:10.840 ⇒ 00:58:13.580 Uttam Kumaran: The stuff is half decent.
588 00:58:14.665 ⇒ 00:58:17.390 Uttam Kumaran: So yeah, it’s like.
589 00:58:17.390 ⇒ 00:58:22.165 Andrew Duplessie: That was Michael Bay, Michael Bay, did it not? James Cameron?
590 00:58:22.950 ⇒ 00:58:28.219 Andrew Duplessie: I mean, if if we, if if we can create the next like skippity or version of it like
591 00:58:28.420 ⇒ 00:58:30.310 Andrew Duplessie: it’s a huge fucking success. Dude.
592 00:58:30.310 ⇒ 00:58:30.650 Uttam Kumaran: Yeah.
593 00:58:31.110 ⇒ 00:58:31.640 adamabdulhamid: But yeah.
594 00:58:31.640 ⇒ 00:58:33.049 Uttam Kumaran: Very little input costs.
595 00:58:33.050 ⇒ 00:58:44.119 Andrew Duplessie: Yeah, it’s the 1st AI studio to create anything that was a hit, because nothing’s been a hit. There’s been 0. That’s been a hit realistically, there’s nothing AI generated. That was a video that you can say.
596 00:58:44.390 ⇒ 00:58:46.850 Uttam Kumaran: So only stuff that’s like gone viral. But like.
597 00:58:46.850 ⇒ 00:58:47.620 Andrew Duplessie: Once, yeah.
598 00:58:47.620 ⇒ 00:58:49.090 Uttam Kumaran: Not not for like.
599 00:58:49.720 ⇒ 00:58:51.850 Andrew Duplessie: Stolen. IP, yeah, like, like that.
600 00:58:52.110 ⇒ 00:58:58.360 Uttam Kumaran: But it’s almost like it, you know. It went viral because it was AI. For you know, a good example of this, you guys.
601 00:58:58.360 ⇒ 00:58:58.779 adamabdulhamid: See you later.
602 00:58:58.780 ⇒ 00:59:01.279 Uttam Kumaran: You guys see that truck, that slate truck that came out.
603 00:59:01.450 ⇒ 00:59:04.019 Andrew Duplessie: No. Oh, yeah, you can like.
604 00:59:04.020 ⇒ 00:59:13.649 Uttam Kumaran: New. Yeah, it’s this new Ev truck that just got announced. But like part of the messaging was, they never once said it was like electric vehicle.
605 00:59:13.990 ⇒ 00:59:19.280 Uttam Kumaran: and that was like what was cool about is, they said, it’s us made. It’s like a small like. It’s 1 of those Toyota helix.
606 00:59:19.670 ⇒ 00:59:20.770 adamabdulhamid: Sizes.
607 00:59:20.950 ⇒ 00:59:31.810 Uttam Kumaran: And they never said it was electric like that was not the selling point. So almost your selling point should not be the fact that it was AI generated like it should stand on its own.
608 00:59:31.810 ⇒ 00:59:32.539 Andrew Duplessie: We’re not doing this.
609 00:59:32.540 ⇒ 00:59:45.559 Uttam Kumaran: For investors, for investors. Sure. But dude! Why, even this one? I tell people, too, I’m like, why are you even pitching that your product is AI like, just build a better product. The AI is like it there.
610 00:59:45.770 ⇒ 00:59:47.479 Uttam Kumaran: Why, why say that.
611 00:59:47.480 ⇒ 00:59:49.830 Andrew Duplessie: Tell you a weird niche that I know will work.
612 00:59:50.010 ⇒ 00:59:58.490 Andrew Duplessie: And just taking a page out of Skibbity, skibbity, skivity, toilet is all a pov of a person right? You never actually see them in 3rd person.
613 00:59:58.940 ⇒ 01:00:04.379 Andrew Duplessie: If if you did it. A series that was pov disney fireworks, it would be a fucking hit.
614 01:00:04.380 ⇒ 01:00:05.250 Uttam Kumaran: Yeah, yeah.
615 01:00:05.250 ⇒ 01:00:15.540 Andrew Duplessie: What my, what my in laws do is they just take their kids down, they plop on the couch, and they put on Disney fireworks, and they just stare. I’m like what the fuck, and then, of course, they’re the whoever runs that channel is just making bang.
616 01:00:16.260 ⇒ 01:00:19.460 Uttam Kumaran: Yeah, or you do like Pov, spider-man.
617 01:00:19.460 ⇒ 01:00:27.440 Andrew Duplessie: Yeah. Pov, you could do a whole pov channel like. So I don’t know there’s there’s we’ll figure it out. But okay.
618 01:00:27.440 ⇒ 01:00:29.169 Uttam Kumaran: Okay, cool. I’ll get back to you guys.
619 01:00:29.410 ⇒ 01:00:32.269 adamabdulhamid: Yeah, I gotta run, Andrew, Andrew. I’ll follow up with you directly.
620 01:00:32.270 ⇒ 01:00:33.379 Andrew Duplessie: Yeah, yeah, how are you guys.
621 01:00:33.380 ⇒ 01:00:34.179 adamabdulhamid: Okay, thank you.
622 01:00:34.180 ⇒ 01:00:34.870 adamabdulhamid: Nice meeting you, Tom.
623 01:00:35.040 ⇒ 01:00:35.640 Uttam Kumaran: Nice to meet you.