WEBVTT 00:00:00.000 --> 00:00:03.759 Welcome to the Deep Dive. Today we are diving 00:00:03.759 --> 00:00:07.179 into this stack of sources about Google VO3 and, 00:00:07.240 --> 00:00:12.460 you know, the whole wild world of viral AI generated 00:00:12.460 --> 00:00:15.019 video. Yeah, it feels like things are like really 00:00:15.019 --> 00:00:17.539 changing fast, doesn't it? Totally. And our mission 00:00:17.539 --> 00:00:19.780 in this Deep Dive is to really unpack this guide 00:00:19.780 --> 00:00:22.859 we have here. It's not just about the tech itself, 00:00:23.019 --> 00:00:24.879 like how to click the buttons to make a video. 00:00:25.019 --> 00:00:28.000 Right. It's about the strategy behind. creating 00:00:28.000 --> 00:00:30.519 stuff that's actually engaging actually shareable 00:00:30.519 --> 00:00:33.439 using ai and you know maybe how to make some 00:00:33.439 --> 00:00:35.479 money doing it too exactly right because the 00:00:35.479 --> 00:00:38.299 source really hits this point hard like vo3 feels 00:00:38.299 --> 00:00:40.640 like this game changer obviously yeah you can 00:00:40.640 --> 00:00:42.899 create video content cinematic looking stuff 00:00:42.899 --> 00:00:46.179 at a fraction of the traditional cost in time. 00:00:46.219 --> 00:00:49.359 A tiny fraction. But having access to the tool, 00:00:49.520 --> 00:00:52.479 that by itself isn't enough anymore. That's the 00:00:52.479 --> 00:00:55.179 key nugget, I think. The real leverage, according 00:00:55.179 --> 00:00:57.719 to leading creators quoted in the source, is 00:00:57.719 --> 00:01:01.780 understanding why people engage. Like what actually 00:01:01.780 --> 00:01:03.460 makes someone stop scrolling? We're gonna try 00:01:03.460 --> 00:01:05.459 and break that down for you. Okay, so let's unpack 00:01:05.459 --> 00:01:08.700 this a bit. The source lays out what it calls 00:01:08.700 --> 00:01:12.019 a surprisingly simple formula for creating videos 00:01:12.019 --> 00:01:14.319 that tend to go viral. Oh, I'm always up for 00:01:14.319 --> 00:01:17.079 a formula. What is it? It boils down to this. 00:01:17.799 --> 00:01:22.019 Setup plus subversion equal surprise. Okay, I 00:01:22.019 --> 00:01:26.680 can get that. So the setup is like what you expect. 00:01:27.019 --> 00:01:29.980 Yeah, exactly. The setup is establishing something 00:01:29.980 --> 00:01:32.180 really familiar to the audience. Setting a norm. 00:01:32.340 --> 00:01:34.079 Yeah, setting up a norm they already recognize. 00:01:34.599 --> 00:01:36.599 Think like a standard commercial style you see 00:01:36.599 --> 00:01:39.079 all the time or maybe a typical product review 00:01:39.079 --> 00:01:42.459 video, a movie trailer trope, you know. Something 00:01:42.459 --> 00:01:44.500 known. Something that grounds the viewer right 00:01:44.500 --> 00:01:46.540 away and something known. Gotcha. So you build 00:01:46.540 --> 00:01:49.200 that expectation. Yeah. And then comes the subversion. 00:01:49.640 --> 00:01:51.879 Precisely. That's where you introduce something 00:01:51.879 --> 00:01:54.379 totally unexpected, maybe really absurd or just 00:01:54.379 --> 00:01:56.920 genuinely funny that completely breaks the pattern 00:01:56.920 --> 00:01:59.159 you just set up. The twist. It's the twist. It 00:01:59.159 --> 00:02:01.319 pulls the rug out from under them. Okay. Here's 00:02:01.319 --> 00:02:03.000 where it gets really interesting because that 00:02:03.000 --> 00:02:06.840 unexpected twist. That's aversion. The source 00:02:06.840 --> 00:02:09.960 says it creates surprise, right? And that actually 00:02:09.960 --> 00:02:12.680 triggers a dopamine release, which feels good. 00:02:12.719 --> 00:02:14.500 It totally does feel good. And that feeling, 00:02:14.560 --> 00:02:17.039 that little emotional hit from the surprise is 00:02:17.039 --> 00:02:20.259 what makes people want to share it. It taps into 00:02:20.259 --> 00:02:23.539 some, you know, really basic human psychology. 00:02:23.580 --> 00:02:25.860 We're wired to find patterns. And when you disrupt 00:02:25.860 --> 00:02:29.240 one in a clever way. It just grabs our attention. 00:02:29.439 --> 00:02:32.340 Yeah, totally. It can cause a really strong reaction, 00:02:32.580 --> 00:02:35.849 laughter, shock, curiosity. And that reaction, 00:02:35.949 --> 00:02:38.370 that's what drives the sharing behavior online. 00:02:38.710 --> 00:02:40.650 It makes perfect sense. Yeah. Your brain is like, 00:02:40.729 --> 00:02:42.650 whoa, what just happened? You have to see this. 00:02:42.729 --> 00:02:45.430 Right. Exactly. And the source gives a perfect 00:02:45.430 --> 00:02:48.729 example of this formula in action with this viral 00:02:48.729 --> 00:02:51.189 pharmaceutical commercial parody called Pupramin. 00:02:51.330 --> 00:02:53.169 Oh, yeah. I saw this reference. The source says 00:02:53.169 --> 00:02:56.509 it reportedly cost only about $500 in AI credits. 00:02:56.689 --> 00:02:59.490 Just $500. And was made in less than a day. Yeah. 00:02:59.770 --> 00:03:02.030 Compared to like a typical half a million dollar 00:03:02.030 --> 00:03:04.400 pharma shoot. Yeah. That's... That's kind of 00:03:04.400 --> 00:03:06.240 mind -blowing. It's a massive, massive difference 00:03:06.240 --> 00:03:09.460 in scale. And this video, Pupperman, absolutely 00:03:09.460 --> 00:03:12.159 nails the formula. How so? The setup is spot 00:03:12.159 --> 00:03:14.659 on. It opens exactly like a standard pharma ad. 00:03:14.860 --> 00:03:17.280 Okay. You have this character, Sarah, middle 00:03:17.280 --> 00:03:19.939 -aged, talking about her depression, how other 00:03:19.939 --> 00:03:21.439 treatments haven't worked. You're right, the 00:03:21.439 --> 00:03:23.340 usual tropes. She's sitting in a comfortable 00:03:23.340 --> 00:03:26.800 living room. The tone is somber, measured. She 00:03:26.800 --> 00:03:29.080 mentions trying new pill. Okay, so you're totally 00:03:29.080 --> 00:03:31.539 building that expected pharma ad pattern. Totally 00:03:31.539 --> 00:03:35.219 typical. But then comes the subversion. The twist 00:03:35.219 --> 00:03:37.860 isn't like some complicated brain chemistry thing. 00:03:38.000 --> 00:03:41.520 No, not at all. The pill makes you secrete pheromones 00:03:41.520 --> 00:03:43.919 that attract adorable puppies to your doorstep. 00:03:44.099 --> 00:03:47.500 Wow. Which is just peak absurdity, but delivered 00:03:47.500 --> 00:03:50.069 so straight. Totally straight. And the source 00:03:50.069 --> 00:03:52.689 points out these absurd, hilarious lines they 00:03:52.689 --> 00:03:54.949 deliver with perfect deadpan pharma sincerity. 00:03:55.550 --> 00:03:58.009 Like, it's really difficult to be depressed when 00:03:58.009 --> 00:04:00.650 cute dogs show up at your doorstep. Or my favorite, 00:04:00.789 --> 00:04:03.370 the non -sequitur one. I used to feel so empty, 00:04:03.409 --> 00:04:05.789 but now I feel joy and mild concern about how 00:04:05.789 --> 00:04:08.050 a pee stain got on the ceiling. The pee stain. 00:04:08.569 --> 00:04:11.530 It's so specific and weirdly relatable. I know. 00:04:11.650 --> 00:04:14.009 And the one about the ex -husband, my puppy listens 00:04:14.009 --> 00:04:16.089 twice as good as my ex -husband and only climbs 00:04:16.089 --> 00:04:18.509 into the lap of half as many of my friends. It's 00:04:18.509 --> 00:04:21.449 brilliant comedic writing, you know, really leveraging 00:04:21.449 --> 00:04:24.209 that formula. Yeah. The video maintained the 00:04:24.209 --> 00:04:28.490 format, the serious tone, even the tiny disclaimers 00:04:28.490 --> 00:04:31.689 at the end while being completely bizarre. That 00:04:31.689 --> 00:04:33.769 dissonance. That dissonance, like the source 00:04:33.769 --> 00:04:36.449 says, is absolutely key to why it was so shareable. 00:04:36.550 --> 00:04:39.160 You just... You just did not expect that. So 00:04:39.160 --> 00:04:42.180 you've got this amazing, unexpected idea following 00:04:42.180 --> 00:04:44.819 the formula. How do you actually make something 00:04:44.819 --> 00:04:48.300 like that using AI tools like VO3? Right, the 00:04:48.300 --> 00:04:51.100 practical steps. The source lays out, you know, 00:04:51.139 --> 00:04:53.439 a whole step -by -step workflow. Yeah, it starts 00:04:53.439 --> 00:04:56.079 with ideation, finding that kernel, that initial 00:04:56.079 --> 00:04:58.060 hook. That kernel of an idea hook. Something 00:04:58.060 --> 00:05:00.459 inherently unexpected. And for Pupperman, the 00:05:00.459 --> 00:05:02.279 first idea wasn't the puppy thing, right? It 00:05:02.279 --> 00:05:04.660 was more standard. No, the source says the initial 00:05:04.660 --> 00:05:07.189 thought was just... A standard pharmaceutical 00:05:07.189 --> 00:05:09.670 commercial. Okay. And they quickly realized that 00:05:09.670 --> 00:05:13.470 would be cinematic blase garbage, just a race 00:05:13.470 --> 00:05:15.990 to the bottom in terms of creativity and impact, 00:05:16.230 --> 00:05:18.930 especially now with AI making that easy. Wait, 00:05:18.990 --> 00:05:22.730 but why is a standard pharma commercial garbage? 00:05:24.230 --> 00:05:26.350 Don't they still work for their target audience? 00:05:26.870 --> 00:05:29.990 What specifically about that approach makes it 00:05:29.990 --> 00:05:33.470 a race to the bottom? in this ai context that's 00:05:33.470 --> 00:05:35.370 a good point it's not that they don't work for 00:05:35.370 --> 00:05:37.829 pharma companies in a traditional marketing sense 00:05:37.829 --> 00:05:41.649 okay but in the context of viral ai video that 00:05:41.649 --> 00:05:43.589 you're trying to get people to share organically 00:05:43.589 --> 00:05:47.230 just making something look cinematic isn't unique 00:05:47.230 --> 00:05:49.490 anymore right the tools make that easy ai tools 00:05:49.490 --> 00:05:52.490 can do that easily so if your idea isn't surprising 00:05:52.490 --> 00:05:55.170 or entertaining if it's just the expected norm 00:05:55.170 --> 00:05:57.730 it won't cut through it won't stand out and you'll 00:05:57.730 --> 00:05:59.870 be competing on price because anyone can make 00:05:59.870 --> 00:06:02.420 a standard looking ad. It's a race to the bottom 00:06:02.420 --> 00:06:05.100 on creativity and cost when the tech commoditizes 00:06:05.100 --> 00:06:08.399 the standard. Ah, okay. That reframes it. So 00:06:08.399 --> 00:06:10.279 they pivoted because the standard approach wouldn't 00:06:10.279 --> 00:06:12.939 stand out. Exactly. First pivot was, what if 00:06:12.939 --> 00:06:14.639 it's a placebo and they just ship you a dog? 00:06:14.800 --> 00:06:17.819 Kegels. Okay. Then the refinement was the pheromones 00:06:17.819 --> 00:06:19.199 and the attracted puppies. That was the winner 00:06:19.199 --> 00:06:21.879 idea. Can AI help with that brainstorming part? 00:06:22.040 --> 00:06:24.769 Finding those unexpected twists? Oh, yeah, totally. 00:06:24.990 --> 00:06:28.829 You can prompt large language models, LLMs, like 00:06:28.829 --> 00:06:31.509 ChatGPT. How would you prompt that? Say something 00:06:31.509 --> 00:06:34.889 like, give me 10 ideas for a pharmaceutical commercial 00:06:34.889 --> 00:06:39.230 with a satirical twist, like SNL or Tim and Eric. 00:06:39.490 --> 00:06:41.790 So you can even specify the kind of comedy style. 00:06:42.050 --> 00:06:44.990 You can. And the source suggests trying multiple 00:06:44.990 --> 00:06:48.269 models, ChatGPT for general ideas, maybe Claude 00:06:48.269 --> 00:06:51.149 for a bit more nuanced take. And Brock for edgier 00:06:51.149 --> 00:06:54.259 stuff. And Grok, especially. especially in its 00:06:54.259 --> 00:06:57.199 less inhibited modes for edgier, crazier stuff, 00:06:57.439 --> 00:06:59.519 the expert and the source actually said, you 00:06:59.519 --> 00:07:02.399 could probably have ChatGPT come up with a viral 00:07:02.399 --> 00:07:04.860 twist. I would say it would pitch you some crazy 00:07:04.860 --> 00:07:07.939 ideas. Okay, so you've got your wild idea honed 00:07:07.939 --> 00:07:10.660 with AI help. Then you move to writing the script, 00:07:10.759 --> 00:07:13.860 also with AI. Primarily through iterative refinement 00:07:13.860 --> 00:07:16.339 with ChatGPT. Iterative, okay. You start with 00:07:16.339 --> 00:07:18.620 a basic prompt describing your concept, the length, 00:07:18.639 --> 00:07:20.560 structure, the twist, the tone you want. And 00:07:20.560 --> 00:07:22.019 then you just use whatever it gives you. First 00:07:22.019 --> 00:07:24.769 shot. Uh, not really. You review what it gives 00:07:24.769 --> 00:07:27.769 you. Some lines are going to be just generic. 00:07:27.949 --> 00:07:30.209 Yeah, probably most. But some will have a spark, 00:07:30.410 --> 00:07:33.629 something unique. The key is iterative refinement. 00:07:33.910 --> 00:07:36.529 You feed the good lines back to it and ask for 00:07:36.529 --> 00:07:39.509 more in that specific style. Like, give me more 00:07:39.509 --> 00:07:41.490 like this one. Exactly. Like, okay, give me 10 00:07:41.490 --> 00:07:43.990 more one -liners like that pee stain joke. Make 00:07:43.990 --> 00:07:46.769 them more unhinged and incorporate non -sequiturs. 00:07:46.769 --> 00:07:48.629 Right. So it's not just one prompt and you're 00:07:48.629 --> 00:07:52.720 done. It's a conversation with the AI. Definitely 00:07:52.720 --> 00:07:55.079 not. The source has a great quote on this. They 00:07:55.079 --> 00:07:57.740 say it's good. Like 95 percent of the jokes aren't 00:07:57.740 --> 00:08:00.220 that funny, but 5 percent actually are. And if 00:08:00.220 --> 00:08:03.220 you just prompted enough. Huh. So you got to 00:08:03.220 --> 00:08:05.139 sift through a lot. You just have to be persistent. 00:08:05.420 --> 00:08:08.639 Keep generating and refining. Probably 70 percent 00:08:08.639 --> 00:08:11.040 of the lines in this are chat GPT, but it's not 00:08:11.040 --> 00:08:13.639 like a one shot prompt. OK, so you've got your 00:08:13.639 --> 00:08:16.379 script refined with the AI. Hitting that setup 00:08:16.379 --> 00:08:18.620 subversion formula. Now you need to actually 00:08:18.620 --> 00:08:22.279 make the video clips using Google VO3. Right, 00:08:22.339 --> 00:08:24.920 the generation phase. How do you translate that 00:08:24.920 --> 00:08:27.759 script into prompts for the video tool? You can 00:08:27.759 --> 00:08:30.819 use ChatGPT to help structure those prompts, 00:08:30.939 --> 00:08:32.940 maybe by giving it a detailed example format 00:08:32.940 --> 00:08:35.399 and asking it to follow that. But the source 00:08:35.399 --> 00:08:38.360 notes that VO3 is actually pretty simple. Oh, 00:08:38.399 --> 00:08:41.059 simpler prompts. That's nice. Less prompt engineering 00:08:41.059 --> 00:08:43.740 needed. Yeah. You basically just need to describe 00:08:43.740 --> 00:08:46.259 what's happening in the shot. It actually does 00:08:46.259 --> 00:08:49.120 not need complicated props. Interesting. A clear 00:08:49.120 --> 00:08:51.919 overview of the scene and the dialogue for that 00:08:51.919 --> 00:08:55.500 moment is often enough. Got it. Describe the 00:08:55.500 --> 00:08:58.360 scene, add the dialogue, and the source had a 00:08:58.360 --> 00:09:02.759 really crucial tip here, right? Avoiding weird 00:09:02.759 --> 00:09:05.659 subtitles. Yes, absolutely crucial. To avoid 00:09:05.659 --> 00:09:08.200 getting those weird gibberish subtitles burned 00:09:08.200 --> 00:09:11.360 directly into your video by VO3. Yeah. What do 00:09:11.360 --> 00:09:13.820 you do? You must not use quotation marks around 00:09:13.820 --> 00:09:16.220 dialogue in your VO3 prompts. No quotes at all. 00:09:16.299 --> 00:09:18.000 Just put the dialogue text there. No quotes. 00:09:18.259 --> 00:09:20.480 Okay, that is a big practical tip. No quotes 00:09:20.480 --> 00:09:22.879 around dialogue in VO3 prompts. So you're ready 00:09:22.879 --> 00:09:25.299 to generate the clips. Right. You access VO3, 00:09:25.419 --> 00:09:26.960 select the highest quality model you want to 00:09:26.960 --> 00:09:29.279 use. Settings. For settings, the source recommends 00:09:29.279 --> 00:09:32.340 generating one output per prompt. Why just one? 00:09:32.679 --> 00:09:36.899 To save money. Or... Both cost and review. You 00:09:36.899 --> 00:09:38.940 want to manage the credits because they add up 00:09:38.940 --> 00:09:41.340 around two, three dollars or a hundred credits 00:09:41.340 --> 00:09:44.120 per generated shot they mention. Right. But also 00:09:44.120 --> 00:09:46.679 you want to review each one. You input prompts 00:09:46.679 --> 00:09:49.080 individually. That lets you iterate and tweak 00:09:49.080 --> 00:09:51.500 if a shot isn't working. Makes sense. Check as 00:09:51.500 --> 00:09:53.340 you go. You don't want to generate a whole batch 00:09:53.340 --> 00:09:55.500 and find out the character looks totally wrong 00:09:55.500 --> 00:09:57.419 in all of them or the camera angle is weird. 00:09:57.620 --> 00:10:01.580 And the source says to expect variation. Like 00:10:01.580 --> 00:10:03.659 you won't use every clip it spits out. Oh yeah. 00:10:03.980 --> 00:10:06.320 Definitely expect a pretty high ratio of generated 00:10:06.320 --> 00:10:09.100 clips to usable ones. They suggest maybe 3 .1 00:10:09.100 --> 00:10:12.299 or even 5 .1. Wow, five generated for every one 00:10:12.299 --> 00:10:14.340 you use. You've got to generate a bunch to get 00:10:14.340 --> 00:10:16.860 the few perfect shots you need for your edit. 00:10:17.039 --> 00:10:18.720 What were the big challenges they mentioned for 00:10:18.720 --> 00:10:21.960 VO3 generation? What's tricky? Character consistency 00:10:21.960 --> 00:10:25.299 is a huge one. VO3 really struggles to make the 00:10:25.299 --> 00:10:27.659 same character look the same across different 00:10:27.659 --> 00:10:31.240 shots. Ugh, that's tough for narrative. That's 00:10:31.240 --> 00:10:33.659 like the biggest issue still, right? It is. The 00:10:33.659 --> 00:10:35.639 best workaround they suggest is describing the 00:10:35.639 --> 00:10:38.159 character in extreme detail. Like how detailed? 00:10:38.500 --> 00:10:42.840 And I mean like extreme. And doing that consistently 00:10:42.840 --> 00:10:45.539 in every single prompt for that character. Okay, 00:10:45.580 --> 00:10:47.960 every single time. Also, apparently text -to 00:10:47.960 --> 00:10:50.720 -video generation is often better for consistency 00:10:50.720 --> 00:10:52.980 than using an image reference, interestingly. 00:10:53.360 --> 00:10:55.879 Hmm, that's good to know. Text -to -video over 00:10:55.879 --> 00:10:59.159 image -to -video for consistency. What else is 00:10:59.159 --> 00:11:01.320 tricky? Audio generation can be inconsistent. 00:11:01.659 --> 00:11:03.440 Sometimes it works. Sometimes it doesn't produce 00:11:03.440 --> 00:11:05.539 audio at all. Sometimes the quality varies a 00:11:05.539 --> 00:11:08.840 lot. Okay, so audio's hit or miss. Yeah. And 00:11:08.840 --> 00:11:11.659 the subtitle thing, again, they mentioned it 00:11:11.659 --> 00:11:13.720 twice, so it must be a real pain point. Yes. 00:11:13.799 --> 00:11:16.100 They reiterate it in the pro tips section because 00:11:16.100 --> 00:11:18.600 it's apparently that important. Absolutely no 00:11:18.600 --> 00:11:20.799 quotation marks around dialogue and the prompts 00:11:20.799 --> 00:11:23.279 if you want to avoid those weird burned in subtitles. 00:11:23.480 --> 00:11:26.940 Seriously. I got it. No quotes. So you've got 00:11:26.940 --> 00:11:29.399 your pile of usable clips after generating and 00:11:29.399 --> 00:11:30.720 iterating. Now you got to put it all together. 00:11:30.779 --> 00:11:33.460 Editing and finalizing. Yep. Import your clips 00:11:33.460 --> 00:11:36.720 into your editor. Final Cut Pro, CapCut, DaVinci 00:11:36.720 --> 00:11:41.179 Resolve. Honestly, the free options now are incredibly 00:11:41.179 --> 00:11:43.500 powerful. Yeah. Resolve is amazing for free. 00:11:43.850 --> 00:11:46.250 Totally. Assemble the clips according to your 00:11:46.250 --> 00:11:49.029 script. And pacing and timing are like super 00:11:49.029 --> 00:11:51.610 crucial here, right? Especially for comedy. Absolutely 00:11:51.610 --> 00:11:54.289 critical. You got to adjust the cuts, the pauses 00:11:54.289 --> 00:11:57.250 between lines, how the dialogue lands. That's 00:11:57.250 --> 00:11:58.870 where the art comes in. That's what makes the 00:11:58.870 --> 00:12:01.409 jokes hit or build suspense effectively. And 00:12:01.409 --> 00:12:04.870 then add the polish, music, sound effects. Exactly. 00:12:05.590 --> 00:12:07.590 Transitions, background music, make sure it's 00:12:07.590 --> 00:12:11.009 licensed. Sound effects, color correction. Make 00:12:11.009 --> 00:12:13.610 it feel like a finished piece. Right. Then export 00:12:13.610 --> 00:12:15.649 it for wherever you're planning to post it, vertical 00:12:15.649 --> 00:12:18.610 for shorts or reels, widescreen for YouTube. 00:12:19.070 --> 00:12:21.429 Cool. So beyond that core workflow, the source 00:12:21.429 --> 00:12:24.990 gives some pro tips from AI viral strategists. 00:12:25.110 --> 00:12:27.549 These are like more advanced techniques to make 00:12:27.549 --> 00:12:29.509 it feel more cinematic and intentional. Yeah, 00:12:29.549 --> 00:12:31.669 like describing camera movements directly in 00:12:31.669 --> 00:12:34.570 your prompts. Don't just do static shots. Add 00:12:34.570 --> 00:12:37.870 things like camera slowly dollies in or a slow 00:12:37.870 --> 00:12:40.970 push in on the product or rack focus to the character 00:12:40.970 --> 00:12:43.090 in the background. Ooh, that adds a lot more 00:12:43.090 --> 00:12:45.350 visual dynamism. Takes it beyond just basic generation. 00:12:45.590 --> 00:12:48.309 What about the characters themselves? Get hyper 00:12:48.309 --> 00:12:50.309 -detailed in your character descriptions. It's 00:12:50.309 --> 00:12:52.190 not just for consistency, though that helps. 00:12:52.470 --> 00:12:56.070 It's also to get a really specific look or feel. 00:12:56.289 --> 00:12:58.629 They give an example for Sarah from Pupperman. 00:12:59.129 --> 00:13:02.090 Middle -aged woman, mid -30s, shoulder -length 00:13:02.090 --> 00:13:05.129 dark blonde hair, slight natural wrinkles around 00:13:05.129 --> 00:13:07.889 her eyes, conveying a sense of lived experience. 00:13:08.289 --> 00:13:09.850 Wow, that's like writing directing notes into 00:13:09.850 --> 00:13:12.730 the prompt. Very specific. It really is. And 00:13:12.730 --> 00:13:15.370 you can include emotional cues and subtext to 00:13:15.370 --> 00:13:18.649 guide the AI on the character's internal state. 00:13:18.889 --> 00:13:21.070 How about that look? Like she coyly quips her 00:13:21.070 --> 00:13:23.450 response, a slight smirk playing on her lips, 00:13:23.570 --> 00:13:26.009 fully aware that she's making a subtle joke about 00:13:26.009 --> 00:13:28.570 her ex -husband's past infidelity. You're basically 00:13:28.570 --> 00:13:30.730 writing performance direction for the AI. That's 00:13:30.730 --> 00:13:33.269 fascinating. Pretty much. And add rich environmental 00:13:33.269 --> 00:13:36.370 or prop details. Small things can make an AI 00:13:36.370 --> 00:13:39.070 scene feel more real or deliberately quirky. 00:13:39.409 --> 00:13:42.169 Like the pee stain. Like adding that a faint, 00:13:42.230 --> 00:13:44.549 unexplained pee stain is visible on the ceiling 00:13:44.549 --> 00:13:47.029 when she looks up. Chuckles the pee stain again. 00:13:47.169 --> 00:13:49.649 It's such a weird, specific detail, but it works. 00:13:49.690 --> 00:13:52.289 It makes it memorable. It does. And then, of 00:13:52.289 --> 00:13:53.950 course, they reiterate the big one for a third 00:13:53.950 --> 00:13:56.990 time. No quotes around dialogue in your prompts. 00:13:57.350 --> 00:14:00.350 Okay. Message received. I'm never putting quotes 00:14:00.350 --> 00:14:03.730 around dialogue in a VO3 prompt now. Ever. Loud 00:14:03.730 --> 00:14:06.269 and clear. Good. It's important. Apparently very 00:14:06.269 --> 00:14:09.250 important. So you've used the formula, mastered 00:14:09.250 --> 00:14:11.649 the workflow, applied the pro tips, and you've 00:14:11.649 --> 00:14:15.570 got this technically good, creatively surprising 00:14:15.570 --> 00:14:18.750 video. But the source says that's really only 00:14:18.750 --> 00:14:20.769 half the battle. Right. Here's where it gets 00:14:20.769 --> 00:14:23.509 really interesting. Making a great video isn't 00:14:23.509 --> 00:14:25.850 enough if no one ever actually watches it. Yeah. 00:14:25.909 --> 00:14:28.970 The hook is arguably the single most critical 00:14:28.970 --> 00:14:31.690 element to getting seen. This totally reminds 00:14:31.690 --> 00:14:34.299 me of... you know, Mr. Beast's philosophy, which 00:14:34.299 --> 00:14:36.799 the source brings up. Yes. He's famous for saying 00:14:36.799 --> 00:14:39.220 headline, thumbnail, video content. Right. His 00:14:39.220 --> 00:14:41.399 point being, if the headline doesn't grab them 00:14:41.399 --> 00:14:43.419 on a feed, they'll never even see the thumbnail. 00:14:43.539 --> 00:14:45.740 They just scroll past. And if the thumbnail isn't 00:14:45.740 --> 00:14:48.139 compelling, they won't click to watch the video 00:14:48.139 --> 00:14:50.419 content, no matter how good it is. It's like 00:14:50.419 --> 00:14:52.879 these super tiny micro windows you have to work 00:14:52.879 --> 00:14:56.460 with, like milliseconds. Exactly. On platforms 00:14:56.460 --> 00:15:00.649 like X or Twitter, you have like literally 0 00:15:00.649 --> 00:15:02.929 .2 seconds to hook someone with your text headline 00:15:02.929 --> 00:15:05.509 as they scroll. 0 .2 seconds. And then if they 00:15:05.509 --> 00:15:08.070 click or if it autoplays, you only have the first 00:15:08.070 --> 00:15:10.250 two, three seconds of the video itself to keep 00:15:10.250 --> 00:15:13.269 them watching. That's terrifyingly short amount 00:15:13.269 --> 00:15:15.990 of time. It really is. So your hook has to be 00:15:15.990 --> 00:15:18.970 incredibly powerful. The source gives a fantastic 00:15:18.970 --> 00:15:21.570 example using the Pupperman Video's own promotion 00:15:21.570 --> 00:15:24.730 hook. Oh yeah, that hook was great. It was. I 00:15:24.730 --> 00:15:27.149 used to shoot $500 pharmaceutical commercials. 00:15:27.490 --> 00:15:31.129 I made this for $500 in VO3 credits in less than 00:15:31.129 --> 00:15:34.049 a day. What's the argument for spending $500k 00:15:34.049 --> 00:15:36.950 now? Steal my prompt below. It's kind of a masterclass 00:15:36.950 --> 00:15:39.460 in hooking attention in that small window. Let's 00:15:39.460 --> 00:15:41.659 break down why that works, according to the source. 00:15:41.820 --> 00:15:44.240 Why is that so effective? Okay, first off, it 00:15:44.240 --> 00:15:46.259 immediately establishes credibility. I used to 00:15:46.259 --> 00:15:48.799 shoot $500 commercials. Authority. Like this 00:15:48.799 --> 00:15:50.539 person knows the traditional industry inside 00:15:50.539 --> 00:15:52.340 and out. Right, they have authority. Then it 00:15:52.340 --> 00:15:55.659 creates a stark, compelling contrast. $500 queues 00:15:55.659 --> 00:15:58.120 versus $500. Yeah, those numbers. Those symmetrical 00:15:58.120 --> 00:16:00.440 numbers make it really stick in your head. Totally. 00:16:00.519 --> 00:16:02.960 It's a massive, almost unbelievable difference, 00:16:03.080 --> 00:16:05.200 which makes you curious. Definitely curious. 00:16:05.460 --> 00:16:07.080 Then it places a really provocative question. 00:16:07.850 --> 00:16:11.710 What's the argument for spending $500K now? Challenges 00:16:11.710 --> 00:16:14.029 the status quo. That challenges traditional thinking 00:16:14.029 --> 00:16:16.509 and encourages people to chime in, increasing 00:16:16.509 --> 00:16:19.309 engagement. And finally, it offers tangible value 00:16:19.309 --> 00:16:23.330 right up front. Steal my prompt below. Gives 00:16:23.330 --> 00:16:25.090 you a reason to click. It gives you a direct 00:16:25.090 --> 00:16:27.350 reason to click to see the video and get the 00:16:27.350 --> 00:16:29.710 goods. It hits all the right notes for curiosity, 00:16:30.009 --> 00:16:33.009 authority, contrast, and value. Yeah, that's 00:16:33.009 --> 00:16:35.509 a killer hook. The source does caution, though. 00:16:35.950 --> 00:16:39.070 Find your own provocative voice. Don't just do 00:16:39.070 --> 00:16:41.769 generic clickbait. Right. Make it specific and 00:16:41.769 --> 00:16:44.470 unique. It has to feel authentic to you and the 00:16:44.470 --> 00:16:47.509 content. Ask a fit. OK, so you've got the formula 00:16:47.509 --> 00:16:50.149 down, the workflow, the pro tips, the killer 00:16:50.149 --> 00:16:53.370 hook. Now, for the fun part, how do you actually 00:16:53.370 --> 00:16:57.830 make money with these VO3 skills and AI video 00:16:57.830 --> 00:17:00.610 creations? The source lists about 10 proven ways 00:17:00.610 --> 00:17:02.929 creators are doing this. Liam on us. What are 00:17:02.929 --> 00:17:06.200 they? First. Build a powerful personal brand. 00:17:06.859 --> 00:17:10.240 Create content regularly. Grow an audience. The 00:17:10.240 --> 00:17:13.059 foundation. That audience is your platform, your 00:17:13.059 --> 00:17:15.200 asset for everything else. Like what we're doing 00:17:15.200 --> 00:17:16.920 right now, I guess. Yeah. Building an audience 00:17:16.920 --> 00:17:18.720 interested in this topic. Pretty much, yeah. 00:17:18.799 --> 00:17:21.960 Second, create bespoke commercials for niche 00:17:21.960 --> 00:17:25.470 industries. Like who? Target places that traditionally 00:17:25.470 --> 00:17:28.890 couldn't afford high -quality video, local breweries, 00:17:28.970 --> 00:17:32.309 small businesses, specialized services. You can 00:17:32.309 --> 00:17:35.410 offer cinematic quality for significantly cheaper. 00:17:35.630 --> 00:17:38.029 Okay, that makes sense. Fill a clear gap in the 00:17:38.029 --> 00:17:40.700 market. Undercut the big guys. for smaller clients. 00:17:40.920 --> 00:17:43.119 Third, partner with marketing and advertising 00:17:43.119 --> 00:17:47.519 agencies. Agencies need AI video expertise now 00:17:47.519 --> 00:17:49.859 to stay competitive. Oh, interesting. You can 00:17:49.859 --> 00:17:51.819 position yourself as their go -to specialist. 00:17:52.119 --> 00:17:54.400 Be the AI video person they call when a client 00:17:54.400 --> 00:17:57.000 wants something cutting edge or fast or cheap. 00:17:57.160 --> 00:18:00.259 Exactly. Fourth, develop narrative series with 00:18:00.259 --> 00:18:02.700 brand integration. Create your own show like 00:18:02.700 --> 00:18:04.799 the Moniverse idea mentioned in the source. Right. 00:18:04.859 --> 00:18:07.599 And integrate brands organically for sponsorships. 00:18:07.960 --> 00:18:10.359 Like product placement within your own, you know, 00:18:10.380 --> 00:18:13.740 created universe. That sounds cool. Fifth, sell 00:18:13.740 --> 00:18:16.480 courses, workshops, info products. Package up 00:18:16.480 --> 00:18:18.319 everything you've learned, your workflows, your 00:18:18.319 --> 00:18:20.700 prompt packs, your strategic insights. Teach 00:18:20.700 --> 00:18:22.960 others. Teach others how to do it. That's a classic 00:18:22.960 --> 00:18:25.460 creator economy move. Leveraging your expertise, 00:18:25.799 --> 00:18:29.059 monetize the knowledge. Sixth, offer corporate 00:18:29.059 --> 00:18:31.730 consulting and training. Help larger companies 00:18:31.730 --> 00:18:34.650 figure out how to integrate AI video into their 00:18:34.650 --> 00:18:37.710 existing structures, legal review, marketing 00:18:37.710 --> 00:18:40.930 teams, compliance. So bigger clients. You can 00:18:40.930 --> 00:18:43.390 scale your pricing significantly for bigger clients. 00:18:43.490 --> 00:18:45.890 So you're not just making videos, you're selling 00:18:45.890 --> 00:18:47.970 your knowledge of the process at an enterprise 00:18:47.970 --> 00:18:51.309 level. Smart. Seventh, launch a niche sketch 00:18:51.309 --> 00:18:54.460 comedy channel. Viral shorts attract brands and 00:18:54.460 --> 00:18:57.119 sponsors. Just like successful TikTok or Instagram, 00:18:57.440 --> 00:19:00.339 real creators get paid. So basically aim to become 00:19:00.339 --> 00:19:04.299 the AI -powered SNL or something. Use AI for 00:19:04.299 --> 00:19:07.299 comedy. If that's your jam. Eighth, create narrative 00:19:07.299 --> 00:19:10.259 films or episodic content. Okay. The source specifically 00:19:10.259 --> 00:19:13.420 emphasizes that episodic weekly content is the 00:19:13.420 --> 00:19:16.779 new creator economy. Weekly content? Why weekly? 00:19:17.259 --> 00:19:20.140 It's more adaptable for AI tools and builds audience 00:19:20.140 --> 00:19:23.740 engagement faster than trying to make, say, a 00:19:23.740 --> 00:19:26.200 feature -length film with the current tech. Right. 00:19:26.279 --> 00:19:28.980 Keeps people coming back. Releasing regularly 00:19:28.980 --> 00:19:32.039 builds that relationship with the audience. Ninth, 00:19:32.099 --> 00:19:35.599 pitch brands on custom narrative content. How's 00:19:35.599 --> 00:19:37.500 that different from integration? Instead of them 00:19:37.500 --> 00:19:39.500 sponsoring your show, you create a whole series 00:19:39.500 --> 00:19:42.680 for the brand where their product or message 00:19:42.680 --> 00:19:45.859 is central to the story. Ah. So it's their show 00:19:45.859 --> 00:19:48.480 made by you. That's a really high value offering. 00:19:48.619 --> 00:19:51.259 Okay. So you're their AI powered production studio, 00:19:51.359 --> 00:19:54.119 essentially. Yep. And 10th, target underserved 00:19:54.119 --> 00:19:57.549 niches. Underserved niches. Like what? Find communities 00:19:57.549 --> 00:20:00.589 with huge passion but a low supply of relevant, 00:20:00.769 --> 00:20:03.589 high -quality video content. They mention faith 00:20:03.589 --> 00:20:05.890 -based content like visually retelling Bible 00:20:05.890 --> 00:20:10.029 stories or content for romanticy readers. Romanticy. 00:20:10.250 --> 00:20:12.890 Yeah, like creating book trailers or short scenes 00:20:12.890 --> 00:20:15.690 for that genre. That's an interesting one. Find 00:20:15.690 --> 00:20:17.569 where the passion is but the video hasn't really 00:20:17.569 --> 00:20:20.269 caught up yet. Very smart. And crucially, the 00:20:20.269 --> 00:20:24.180 source says, value your skills. Don't undercharge 00:20:24.180 --> 00:20:26.960 just because you used AI tools to help create 00:20:26.960 --> 00:20:30.819 it. Aim for project values that reflect the results 00:20:30.819 --> 00:20:33.740 you can deliver to a client. They suggest a $5 00:20:33.740 --> 00:20:36.700 ,000 minimum for substantial client projects 00:20:36.700 --> 00:20:39.900 as a potential benchmark. Good reminder to know 00:20:39.900 --> 00:20:42.880 your worth, even when the tools are cheaper or 00:20:42.880 --> 00:20:46.150 faster. Now, the source brings up something really 00:20:46.150 --> 00:20:47.869 important to think about regarding the future, 00:20:48.029 --> 00:20:50.210 kind of the inevitability of commodification. 00:20:50.609 --> 00:20:52.650 Yeah, this is a big challenge everyone using 00:20:52.650 --> 00:20:55.130 these tools needs to face. What is it? As tools 00:20:55.130 --> 00:20:58.710 like VO3 become super easy, super common, just 00:20:58.710 --> 00:21:00.849 making things look cinematic or professional, 00:21:01.130 --> 00:21:02.829 that's not going to be special anymore. It'll 00:21:02.829 --> 00:21:04.609 just be table stakes. It's just going to be the 00:21:04.609 --> 00:21:06.670 expected standard. So if you're just focusing 00:21:06.670 --> 00:21:09.130 on the technical execution, the cinematic look, 00:21:09.289 --> 00:21:12.049 then what? You end up creating that cinematic 00:21:12.049 --> 00:21:14.230 blase garbage we talked about earlier. Right. 00:21:14.309 --> 00:21:16.549 And it absolutely becomes a race to the bottom 00:21:16.549 --> 00:21:18.930 on price. Your technical skill, your ability 00:21:18.930 --> 00:21:21.869 to just make a nice looking video gets commoditized 00:21:21.869 --> 00:21:24.490 because literally anyone can do it. Okay. So 00:21:24.490 --> 00:21:26.890 technical skill isn't the differentiator long 00:21:26.890 --> 00:21:29.470 term. Nope. There's a really key quote from the 00:21:29.470 --> 00:21:32.029 source that sums this up. Making things cinematic 00:21:32.029 --> 00:21:34.869 is no longer difficult. Telling good stories 00:21:34.869 --> 00:21:39.849 is still a rare skill. That hits hard. Storytelling 00:21:39.849 --> 00:21:42.089 is the key. That's the core insight right there. 00:21:42.170 --> 00:21:45.349 In a world already drowning in content and soon 00:21:45.349 --> 00:21:48.869 to be absolutely flooded with AI video. Yeah, 00:21:48.910 --> 00:21:51.430 it's coming. Only the stuff that's truly entertaining, 00:21:51.650 --> 00:21:53.710 truly surprising, that connects emotionally, 00:21:53.930 --> 00:21:56.529 that's narratively compelling. That's what's 00:21:56.529 --> 00:21:58.789 going to cut through the noise. So it's the idea, 00:21:58.869 --> 00:22:01.970 the story, the psychology. Your unique creative 00:22:01.970 --> 00:22:05.390 perspective, your storytelling ability, understanding 00:22:05.390 --> 00:22:08.630 that psychology of why people share and crafting 00:22:08.630 --> 00:22:10.849 that killer hook. Those are your most valuable 00:22:10.849 --> 00:22:13.250 assets, not just operating the tool. Absolutely. 00:22:13.470 --> 00:22:15.589 OK, so the biggest takeaway from this guide, 00:22:15.670 --> 00:22:17.750 from this deep dive, isn't just the workflow 00:22:17.750 --> 00:22:20.130 or even the list of monetization methods, right? 00:22:20.190 --> 00:22:23.170 Not really, no. It's about finding and leveraging 00:22:23.170 --> 00:22:25.390 your unique advantage in this new landscape. 00:22:25.759 --> 00:22:28.319 Yeah, it's your unique creative angle, your authentic 00:22:28.319 --> 00:22:31.740 voice, and that deep understanding of, you know, 00:22:31.759 --> 00:22:34.559 what makes content shareable at a human level. 00:22:34.720 --> 00:22:36.519 The source gives some guiding principles for 00:22:36.519 --> 00:22:39.180 this. Like, lean into your weirdness and your 00:22:39.180 --> 00:22:41.380 passion. Yeah, your unique interests. Whatever 00:22:41.380 --> 00:22:43.660 genuinely fascinates you, whatever your quirky 00:22:43.660 --> 00:22:46.869 interests or obsessions are. That's fertile ground 00:22:46.869 --> 00:22:50.089 for original content ideas that will resonate 00:22:50.089 --> 00:22:51.990 with others who share them. That's where unique 00:22:51.990 --> 00:22:54.710 voices come from. Don't try to be generic. Find 00:22:54.710 --> 00:22:56.670 and serve those underserved niches we talked 00:22:56.670 --> 00:22:58.769 about. Right. Look for passionate communities 00:22:58.769 --> 00:23:01.769 where there's a real demand for content, but 00:23:01.769 --> 00:23:03.750 maybe not a lot of good video supply yet. And 00:23:03.750 --> 00:23:05.890 this one's really interesting. Copy frameworks. 00:23:07.090 --> 00:23:09.630 Not content. Okay, what does that mean? Don't 00:23:09.630 --> 00:23:12.509 just remake the viral video you saw. Study how 00:23:12.509 --> 00:23:15.230 successful videos work. The setup subversion 00:23:15.230 --> 00:23:17.470 formula, the pacing, the editing structure, the 00:23:17.470 --> 00:23:19.529 hook strategy. Learn the structure. Learn the 00:23:19.529 --> 00:23:22.769 mechanics. Then apply those structures to your 00:23:22.769 --> 00:23:25.710 original ideas, your unique perspective, your 00:23:25.710 --> 00:23:28.779 niche. Yeah, don't copy the dance trend. Copy 00:23:28.779 --> 00:23:31.059 the structure of a viral dance trend and put 00:23:31.059 --> 00:23:33.559 your own spin on it. Exactly. They even suggest 00:23:33.559 --> 00:23:36.460 practicing by recreating things just to learn 00:23:36.460 --> 00:23:38.799 the mechanics and then go make your own. That's 00:23:38.799 --> 00:23:41.319 good advice. Practice the form. And ultimately, 00:23:41.440 --> 00:23:45.140 just do what genuinely fascinates you. Authentic 00:23:45.140 --> 00:23:47.880 passion really shows through, you know. Yeah, 00:23:47.940 --> 00:23:50.069 you can tell. And it tends to lead to the most 00:23:50.069 --> 00:23:53.009 original, most compelling work. It all comes 00:23:53.009 --> 00:23:55.450 back to that core idea the creators and the source 00:23:55.450 --> 00:23:57.250 mentioned right at the beginning. What was that 00:23:57.250 --> 00:23:59.769 again? It's not just about your ability to technically 00:23:59.769 --> 00:24:02.690 make something. It's about your ability to understand 00:24:02.690 --> 00:24:04.869 why people would want to watch the whole thing 00:24:04.869 --> 00:24:07.410 and share it with their friends. The why. Yeah. 00:24:07.509 --> 00:24:10.309 That revolution is definitely here, as the source 00:24:10.309 --> 00:24:13.890 puts it. Tools like Google VO3 are making cinematic 00:24:13.890 --> 00:24:16.369 video creation more accessible than it's ever 00:24:16.369 --> 00:24:19.019 been. But the real power. The real advantage 00:24:19.019 --> 00:24:22.660 you have in this new world. It's not just the 00:24:22.660 --> 00:24:26.150 tool. It's you. Yeah. It's your unique vision, 00:24:26.269 --> 00:24:28.470 your ability to tell stories that connect with 00:24:28.470 --> 00:24:31.029 people, your understanding of that human element. 00:24:31.150 --> 00:24:33.470 Think about that Pupperman example one last time. 00:24:33.589 --> 00:24:36.569 It wasn't just the AI making weird visuals. No. 00:24:36.630 --> 00:24:39.650 It was the brilliant, absurd concept behind it 00:24:39.650 --> 00:24:42.430 and the perfect execution of that setup plus 00:24:42.430 --> 00:24:45.210 subversion formula that made it memorable and 00:24:45.210 --> 00:24:48.390 shareable. So what unique story do you have swirling 00:24:48.390 --> 00:24:51.190 around in your head? What underserved niche sparks 00:24:51.190 --> 00:24:53.829 your interest? That's really something to talk 00:24:53.829 --> 00:24:55.369 about. mull over as we wrap up this deep dive. 00:24:55.490 --> 00:24:57.569 Hopefully this gave you some real aha moments 00:24:57.569 --> 00:25:00.309 and sparked some ideas for your own creative 00:25:00.309 --> 00:25:03.430 ventures in this wild new AI video world.