WEBVTT 00:00:00.000 --> 00:00:03.600 $5 ,000 to $10 ,000 and usually a month minimum 00:00:03.600 --> 00:00:06.580 of just waiting around. That is the traditional 00:00:06.580 --> 00:00:08.779 cost of entry. If you want to launch a professional 00:00:08.779 --> 00:00:11.019 looking shoe brand, that's the before picture. 00:00:11.039 --> 00:00:14.820 This is what less than $50. Yeah. And maybe one 00:00:14.820 --> 00:00:18.280 hour of focused work. Welcome back to the deep 00:00:18.280 --> 00:00:20.660 dive. It's good to have you here today. We're 00:00:20.660 --> 00:00:24.609 looking at a pretty radical shift. in the mechanics 00:00:24.609 --> 00:00:27.570 of how businesses get built. We are analyzing 00:00:27.570 --> 00:00:30.829 a comprehensive guide on building a theoretical 00:00:30.829 --> 00:00:34.630 athletic brand aptly named Rival from absolute 00:00:34.630 --> 00:00:36.530 scratch. And I wanna be clear about what scratch 00:00:36.530 --> 00:00:39.250 means here. Yeah, we aren't talking about scribbling 00:00:39.250 --> 00:00:41.689 a logo on a napkin and calling it a day. We are 00:00:41.689 --> 00:00:44.630 talking about high fidelity product design, lifestyle 00:00:44.630 --> 00:00:48.490 photography. even motion video ads, the whole 00:00:48.490 --> 00:00:50.929 thing. The full stack. Yeah. So the roadmap for 00:00:50.929 --> 00:00:52.570 our discussion is really specific. We're going 00:00:52.570 --> 00:00:54.270 to look at the philosophy behind this, which 00:00:54.270 --> 00:00:56.750 the source calls Failfast. Then we're getting 00:00:56.750 --> 00:00:59.630 into the weeds of the toolkit, specifically Nano 00:00:59.630 --> 00:01:02.670 Banana Pro for the visuals and Cling 2 .6 for 00:01:02.670 --> 00:01:05.290 the video work. But I really want to explore 00:01:05.290 --> 00:01:08.069 why these specific prompts work, not just, you 00:01:08.069 --> 00:01:10.510 know, read them out. It's a fascinating case 00:01:10.510 --> 00:01:13.180 study because it stops being about Can I afford 00:01:13.180 --> 00:01:16.180 to try this? And it starts being about, how fast 00:01:16.180 --> 00:01:18.920 can I iterate before I get bored? So let's unpack 00:01:18.920 --> 00:01:21.780 the problem first. Because usually, when we talk 00:01:21.780 --> 00:01:24.120 about startups, we talk about runway. The source 00:01:24.120 --> 00:01:27.079 material paints this picture of the biggest wall 00:01:27.079 --> 00:01:29.579 facing entrepreneurs. It's the gatekeeper. It 00:01:29.579 --> 00:01:33.500 is. Historically, if you had an idea for a sneaker, 00:01:33.780 --> 00:01:36.920 say, a high top for weight lifters, you couldn't 00:01:36.920 --> 00:01:38.799 just visualize it. You had to hire a designer. 00:01:39.379 --> 00:01:41.920 That's $1 ,000 right there. Instantly. Then if 00:01:41.920 --> 00:01:43.579 you wanted people to actually care about it, 00:01:43.799 --> 00:01:46.459 you needed a photographer, models, a studio. 00:01:46.719 --> 00:01:49.000 That's another three grand. You're bleeding cash 00:01:49.000 --> 00:01:51.239 before you've sold a single pair. And the kicker 00:01:51.239 --> 00:01:53.799 is you spend all that money and you still don't 00:01:53.799 --> 00:01:56.159 know if anyone wants the shoe. That's the terrifying 00:01:56.159 --> 00:01:57.939 part. You could spend months and thousands of 00:01:57.939 --> 00:02:00.500 dollars just to find out your idea was, well, 00:02:00.620 --> 00:02:04.140 bad. OK, so this is where the source introduces 00:02:04.140 --> 00:02:07.120 the fail fast and cheap strategy. But I want 00:02:07.120 --> 00:02:09.740 to push on this a bit because fail fast is like. 00:02:10.409 --> 00:02:12.889 Silicon Valley cliche number one, everyone says 00:02:12.889 --> 00:02:15.129 it, what's different here? The difference is 00:02:15.129 --> 00:02:18.930 the denominator. Failing fast usually means fail 00:02:18.930 --> 00:02:22.110 within six months. Here, failing fast means failing 00:02:22.110 --> 00:02:24.750 before your coffee gets cold. The premise is, 00:02:25.229 --> 00:02:27.889 if your design is ugly, or if the market hates 00:02:27.889 --> 00:02:30.650 it, you want to find that out in an hour on your 00:02:30.650 --> 00:02:33.870 laptop, costing you basically zero dollars. You 00:02:33.870 --> 00:02:36.030 don't want to find that out after you've, you 00:02:36.030 --> 00:02:38.289 know... drained your savings account on a prototype 00:02:38.289 --> 00:02:41.050 run. So it changes the risk profile from financial 00:02:41.050 --> 00:02:44.789 to just your time. Exactly. It's purely a time 00:02:44.789 --> 00:02:46.669 investment. So this is where it gets really interesting, 00:02:46.789 --> 00:02:49.129 though. The guide admits that most people are 00:02:49.129 --> 00:02:52.110 using these AI tools completely wrong. Oh, yeah. 00:02:52.550 --> 00:02:55.370 This part. This really resonated with me. Why 00:02:55.370 --> 00:02:57.069 is that? You're usually the one who's all in 00:02:57.069 --> 00:02:58.849 on these tools. Well, I have to make a bit of 00:02:58.849 --> 00:03:00.569 a vulnerable admission here. I struggle with 00:03:00.569 --> 00:03:02.629 this, too. Oh. Usually when I open up an image 00:03:02.629 --> 00:03:04.610 generator and I type in something like, make 00:03:04.610 --> 00:03:08.939 me a cool athletic shoe. The result, it's just 00:03:08.939 --> 00:03:11.680 bad. It looks like a cheap toy. It looks plastic. 00:03:11.819 --> 00:03:14.120 It has that weird AI sheen where everything is 00:03:14.120 --> 00:03:16.719 too smooth. The uncanny value issue? Precisely. 00:03:17.000 --> 00:03:19.879 And I always wonder why. Is the model bad? No, 00:03:19.879 --> 00:03:22.060 I'm bad. The source actually breaks down the 00:03:22.060 --> 00:03:24.879 fix, and it's surprisingly manual. It's a three 00:03:24.879 --> 00:03:28.219 -step strategy. Walk us through that. OK, so 00:03:28.219 --> 00:03:30.819 step one is all human. You go find inspiration. 00:03:31.229 --> 00:03:33.969 You look at Nike, you look at Adidas. You find 00:03:33.969 --> 00:03:36.270 that gold standard image that has the vibe you 00:03:36.270 --> 00:03:39.650 want. But step two is the bridge. You don't just 00:03:39.650 --> 00:03:42.050 look at it and try to describe it yourself. You 00:03:42.050 --> 00:03:46.189 upload that photo to a text -based AI, like ChatGPT, 00:03:46.409 --> 00:03:48.909 and you ask it to describe the lighting, the 00:03:48.909 --> 00:03:50.710 materials, the camera angle. Wait, hold on. Isn't 00:03:50.710 --> 00:03:53.009 that just passing the buck? You're using one 00:03:53.009 --> 00:03:55.289 AI to write the homework for the other AI? It 00:03:55.289 --> 00:03:57.409 feels like cheating, doesn't it? But think of 00:03:57.409 --> 00:04:00.939 it as a translation layer. The AI knows the technical 00:04:00.939 --> 00:04:03.520 vocabulary that is statistically associated with 00:04:03.520 --> 00:04:05.879 high quality images in its training data. It 00:04:05.879 --> 00:04:09.139 knows words like herringbone traction or soft 00:04:09.139 --> 00:04:13.759 studio lighting or volumetric fog, words that 00:04:13.759 --> 00:04:16.019 you and I might not think of. So it's extracting 00:04:16.019 --> 00:04:19.120 the metadata of the aesthetic. Yes. It's giving 00:04:19.120 --> 00:04:22.300 you the cheat codes for the latent space. Then 00:04:22.300 --> 00:04:24.939 step three is feeding that professional description 00:04:25.420 --> 00:04:28.959 into the image generator, NanoBanana Pro. So, 00:04:29.220 --> 00:04:31.139 proving that for a second, it's not just about 00:04:31.139 --> 00:04:33.360 typing a command, it's about learning the language 00:04:33.360 --> 00:04:36.500 of design. You use AI to learn the vocabulary 00:04:36.500 --> 00:04:39.470 of experts so you can speak to the machine. Okay, 00:04:39.610 --> 00:04:41.689 so once you have that vocabulary you have to 00:04:41.689 --> 00:04:44.069 build the identity the source creates this brand 00:04:44.069 --> 00:04:47.490 rival Which is a solid name? Short suggests competition 00:04:47.490 --> 00:04:50.370 and it's easy for the AI to render because it's 00:04:50.370 --> 00:04:52.209 not a complex string of characters They start 00:04:52.209 --> 00:04:54.829 with the logo and I noticed the prompts engineering 00:04:54.829 --> 00:04:56.709 here was just incredibly specific They didn't 00:04:56.709 --> 00:04:58.750 just say make a sports logo. No, not at all. 00:04:58.750 --> 00:05:02.100 They asked for a minimalist athletic logo a split 00:05:02.100 --> 00:05:04.899 V design like a mountain peak, and specific sans 00:05:04.899 --> 00:05:07.360 serif typography. Right. But there's a technical 00:05:07.360 --> 00:05:09.379 nuance here that I think is really important 00:05:09.379 --> 00:05:11.180 for anyone listening who wants to actually try 00:05:11.180 --> 00:05:14.879 this. The aspect ratio setting. Yes. The guide 00:05:14.879 --> 00:05:19.680 advises setting Nano Banana Pro to 16 .9. Which 00:05:19.680 --> 00:05:22.040 is wide, like a TV screen. Why does that matter 00:05:22.040 --> 00:05:24.420 for a logo? Logos are usually squares or circles. 00:05:24.879 --> 00:05:27.560 It's about how these diffusion models work. If 00:05:27.560 --> 00:05:30.519 you force it into a square, the AI often tries 00:05:30.519 --> 00:05:33.279 to fill every corner with noise or detail. It 00:05:33.279 --> 00:05:36.420 feels cramped. Ah, okay. The source argues that 00:05:36.420 --> 00:05:39.420 the wide format gives the AI room to breathe. 00:05:39.920 --> 00:05:42.459 It centers the logo and leaves negative space 00:05:42.459 --> 00:05:45.379 on the sides. It prevents the text from getting 00:05:45.379 --> 00:05:47.720 cramped or garbled, which is still a massive 00:05:47.720 --> 00:05:50.779 headache with AI text. That's a great tip. Now 00:05:50.779 --> 00:05:52.660 once you have that logo, the guide highlights 00:05:52.660 --> 00:05:54.930 what it calls a crucial step. You can't just 00:05:54.930 --> 00:05:56.930 keep generating random images. You have to use 00:05:56.930 --> 00:05:59.149 the reference image feature. This is the anchor. 00:05:59.209 --> 00:06:01.029 This is probably the most critical part of the 00:06:01.029 --> 00:06:03.410 whole workflow. Explain how that works technically, 00:06:03.610 --> 00:06:05.889 because usually AI is like a slot machine. You 00:06:05.889 --> 00:06:07.829 pull the lever, you get a whole new result. Right. 00:06:08.209 --> 00:06:10.449 If you don't use a reference, every time you 00:06:10.449 --> 00:06:13.069 ask for a shoe with the rival logo, the AI will 00:06:13.069 --> 00:06:15.470 just hallucinate a new logo. It doesn't know 00:06:15.470 --> 00:06:18.129 your brand. It just knows the concept of a logo. 00:06:18.360 --> 00:06:21.360 Okay. By using the reference image feature, you're 00:06:21.360 --> 00:06:24.220 essentially locking a set of pixel values. You're 00:06:24.220 --> 00:06:26.079 telling the algorithm, do whatever you want with 00:06:26.079 --> 00:06:29.060 the shoe, but this specific pattern of pixels, 00:06:29.600 --> 00:06:32.620 the logo, must remain statistically similar to 00:06:32.620 --> 00:06:35.360 this input. So it constrains the randomness? 00:06:35.600 --> 00:06:37.879 It constrains the randomness in that one specific 00:06:37.879 --> 00:06:40.079 area. Without it, you don't have a brand, you 00:06:40.079 --> 00:06:43.040 just have a folder full of random pictures. And 00:06:43.040 --> 00:06:45.620 why is that reference step so critical for a 00:06:45.620 --> 00:06:48.899 business specifically? Consistency creates trust. 00:06:49.500 --> 00:06:51.939 Customers need to see the same symbol everywhere, 00:06:52.420 --> 00:06:55.019 or it just feels like a scam. So you have the 00:06:55.019 --> 00:06:57.699 logo locked in. Now we move to the physical product. 00:06:58.060 --> 00:07:00.600 The guide walks through creating a six shoe lineup, 00:07:01.079 --> 00:07:03.079 three for men, three for women. And this is where 00:07:03.079 --> 00:07:04.860 that whole vocabulary lesson we talked about 00:07:04.860 --> 00:07:08.319 really pays off. The texture. is king here. I 00:07:08.319 --> 00:07:10.180 was reading the prompt for the men's training 00:07:10.180 --> 00:07:12.860 shoe and it was so detailed it wasn't just white 00:07:12.860 --> 00:07:15.459 shoe. No, it creates a sensory experience. The 00:07:15.459 --> 00:07:18.500 prompt called for breathable white mesh, synthetic 00:07:18.500 --> 00:07:21.639 leather overlays, and a translucent rubber sole 00:07:21.639 --> 00:07:24.220 with a herringbone traction pattern. See, herringbone 00:07:24.220 --> 00:07:25.920 traction pattern is not a phrase I would ever 00:07:25.920 --> 00:07:28.480 think to type. I'd probably say zigzag bottom. 00:07:29.019 --> 00:07:32.439 And if you type zigzag you'd get a cartoon. Herringbone 00:07:32.699 --> 00:07:36.240 triggers a very specific subset of training data 00:07:36.240 --> 00:07:38.980 professional product photography. That's the 00:07:38.980 --> 00:07:41.339 difference between a concept sketch and a manufacturing 00:07:41.339 --> 00:07:44.139 ready visual. So it's simulating the physics 00:07:44.139 --> 00:07:46.500 of light based on the material definition. It 00:07:46.500 --> 00:07:50.300 is. The AI understands how light bounces off 00:07:50.300 --> 00:07:53.660 mesh differently than it does off leather. If 00:07:53.660 --> 00:07:56.399 you don't specify the material, the AI defaults 00:07:56.399 --> 00:07:58.980 to the average of all shoe images, which usually 00:07:58.980 --> 00:08:01.740 results in that smooth plastic look. Specificity 00:08:01.740 --> 00:08:04.180 forces realism. It really does. And then they 00:08:04.180 --> 00:08:06.379 just iterated on this. They made the midnight 00:08:06.379 --> 00:08:09.279 in black and gray, the energy in white and orange, 00:08:09.339 --> 00:08:11.759 and the zen in light blue. Yeah, and because 00:08:11.759 --> 00:08:13.579 they use that reference anchor we talked about, 00:08:14.180 --> 00:08:16.670 the logo is perfect on all of them. It took them 00:08:16.670 --> 00:08:18.589 seconds to create a full seasonal collection. 00:08:18.910 --> 00:08:21.850 So we move from a general idea to a manufacturing 00:08:21.850 --> 00:08:24.850 -ready visual? Right. Specificity in the prompt 00:08:24.850 --> 00:08:27.110 yields reality in the image. OK. So you have 00:08:27.110 --> 00:08:29.649 the shoes. But a white background photo is just 00:08:29.649 --> 00:08:34.049 a catalog entry. It's dry. It's not an ad. Context 00:08:34.049 --> 00:08:36.730 sells. The source makes a really big point of 00:08:36.730 --> 00:08:38.809 this. You need lifestyle photography. You need 00:08:38.809 --> 00:08:41.909 to sell the dream of athleticism, the sweat, 00:08:42.090 --> 00:08:44.750 the effort, not just the rubber and glue. The 00:08:44.750 --> 00:08:47.570 setup they used for the lifestyle shots was really 00:08:47.570 --> 00:08:50.490 evocative. They went for a cinematic sports photography 00:08:50.490 --> 00:08:53.269 look. The fit male athlete doing a deadlift. 00:08:54.029 --> 00:08:56.190 But look at the lighting choices. They suggested 00:08:56.190 --> 00:08:59.480 moody studio lighting or golden hour. And there 00:08:59.480 --> 00:09:01.820 was a specific camera tip there, too, using a 00:09:01.820 --> 00:09:04.419 35 millimeter lens in the prompt. Does the AI 00:09:04.419 --> 00:09:06.700 actually know what a lens does? It does. It's 00:09:06.700 --> 00:09:09.120 incredible. A 35 millimeter lens is standard 00:09:09.120 --> 00:09:11.820 for photojournalism. It gives a slight distortion, 00:09:12.039 --> 00:09:14.519 a wider field of view, a sense of being there. 00:09:14.620 --> 00:09:17.600 OK. If you don't specify the lens, AI tends to 00:09:17.600 --> 00:09:19.899 use a telephoto look where everything is flattened 00:09:19.899 --> 00:09:22.460 and perfect. Looks like a stock photo. 35 millimeter 00:09:22.460 --> 00:09:24.179 looks like a story. And that leads to the problem 00:09:24.179 --> 00:09:26.039 we touched on earlier, the plastic mannequin 00:09:26.039 --> 00:09:28.519 look. Humans are hard for AI. They often look. 00:09:28.889 --> 00:09:31.110 Vacant? How did they solve that in a lifestyle 00:09:31.110 --> 00:09:34.309 shot? Sweat. Literally. They add natural skin 00:09:34.309 --> 00:09:36.830 texture and sweat to the prompt. Why does that 00:09:36.830 --> 00:09:40.549 trick the brain? Because AI models are biased 00:09:40.549 --> 00:09:42.750 toward symmetry and perfection. They want to 00:09:42.750 --> 00:09:45.289 make the skin smooth. But real life is messy. 00:09:45.750 --> 00:09:48.669 Real skin has pores, blemishes, moisture. Right. 00:09:49.070 --> 00:09:52.889 By forcing the AI to render sweat, You are forcing 00:09:52.889 --> 00:09:56.629 it to introduce noise and texture. That imperfection 00:09:56.629 --> 00:09:58.690 is what signals to your brain, this is a real 00:09:58.690 --> 00:10:01.669 human. That's wild. We are adding dirt to make 00:10:01.669 --> 00:10:04.850 it look clean. It's the paradox of realism. So, 00:10:04.870 --> 00:10:08.110 probing that, how do we stop the model from looking 00:10:08.110 --> 00:10:10.289 like a plastic mannequin? We request natural 00:10:10.289 --> 00:10:13.190 skin texture and sweat to ground it in biological 00:10:13.190 --> 00:10:15.450 reality. Now, here is where it gets really interesting 00:10:15.450 --> 00:10:19.049 for me. We are moving from still images to motion. 00:10:19.269 --> 00:10:21.240 This is the moment of wonder. It really does 00:10:21.240 --> 00:10:24.220 feel like magic. Using cling 2 .6. Yeah, taking 00:10:24.220 --> 00:10:26.200 that still image of the deadlift, which is already 00:10:26.200 --> 00:10:28.379 impressive and making it move. The guide says 00:10:28.379 --> 00:10:30.580 to upload the lifestyle shot and then describe 00:10:30.580 --> 00:10:33.320 the movement, but it warns about glitches. And 00:10:33.320 --> 00:10:35.539 we've all seen these AI videos where people morph 00:10:35.539 --> 00:10:38.139 into demons or grow a third arm. The dreaded 00:10:38.139 --> 00:10:41.039 AI shimmer. It just breaks the immersion instantly. 00:10:41.200 --> 00:10:43.720 So how do they avoid that? The source seems pretty 00:10:43.720 --> 00:10:46.080 conservative here. The advice is essentially, 00:10:46.820 --> 00:10:49.179 restrain yourself. Don't ask for a backflip. 00:10:49.259 --> 00:10:52.159 Keep it simple. Exactly. AI video generation 00:10:52.159 --> 00:10:54.799 isn't actually moving 3D models. It's predicting 00:10:54.799 --> 00:10:58.100 the next set of pixels. Complex movements like 00:10:58.100 --> 00:11:00.440 a backflip require a lot of prediction, which 00:11:00.440 --> 00:11:03.399 leads to errors. The source suggests asking for 00:11:03.399 --> 00:11:05.860 subtle movements, natural breathing, muscles 00:11:05.860 --> 00:11:09.519 tensing, or maybe slowly tying shoelaces. Or 00:11:09.519 --> 00:11:12.960 just move the camera a slow dolly in to focus 00:11:12.960 --> 00:11:15.120 on the shoes. It's interesting that less is more 00:11:15.120 --> 00:11:17.350 here. Well think about a luxury ad from Nike 00:11:17.350 --> 00:11:20.330 or Adidas. Is it always frantic action? No. Often 00:11:20.330 --> 00:11:22.730 it's slow motion, controlled, high definition 00:11:22.730 --> 00:11:25.470 focus. It's confident. Exactly. Fast movement 00:11:25.470 --> 00:11:28.570 hides mistakes. Slow movement shows off quality. 00:11:28.889 --> 00:11:31.470 Why does subtle movement work better for advertising 00:11:31.470 --> 00:11:33.669 here? It connects emotionally without breaking 00:11:33.669 --> 00:11:35.669 the visual illusion. So you have the product, 00:11:35.830 --> 00:11:38.909 the photos, the video. But you don't have a story 00:11:38.909 --> 00:11:41.309 yet. And this is the part I felt was surprisingly 00:11:41.309 --> 00:11:45.080 deep in the source material. the digital storefront. 00:11:45.299 --> 00:11:47.080 Which brings us to the final piece of the puzzle. 00:11:47.519 --> 00:11:50.360 They use Nano Banana Pro to generate mock -ups 00:11:50.360 --> 00:11:53.519 of a website and an Instagram grid. Why bother 00:11:53.519 --> 00:11:55.139 doing this if you haven't built the site yet? 00:11:55.320 --> 00:11:58.620 Because context changes perception. Yeah. You 00:11:58.620 --> 00:12:00.620 might love your logo when it's on a white PDF 00:12:00.620 --> 00:12:04.879 background. But rival uses this stark black and 00:12:04.879 --> 00:12:07.200 white aesthetic right when they generated the 00:12:07.200 --> 00:12:09.340 Instagram grid mock -up They get instantly see 00:12:09.340 --> 00:12:11.639 does this look cool or does it look depressing? 00:12:11.919 --> 00:12:14.259 Does the black logo just disappear when the website 00:12:14.259 --> 00:12:17.679 is in dark mode? It's a vibe check. It's a holistic 00:12:17.679 --> 00:12:20.840 system check you are visualizing the whole ecosystem 00:12:20.840 --> 00:12:23.379 of the brand If the Instagram grid looks messy, 00:12:23.980 --> 00:12:26.740 your brand identity is wrong, even if the individual 00:12:26.740 --> 00:12:28.899 photos are good. So you're building the store 00:12:28.899 --> 00:12:31.899 before you have the inventory. Yes. You validate 00:12:31.899 --> 00:12:34.440 the aesthetic before you spend a dime on code 00:12:34.440 --> 00:12:36.480 or Shopify themes. We're going to take a very 00:12:36.480 --> 00:12:37.940 short break. But when we come back, I want to 00:12:37.940 --> 00:12:41.620 zoom out. We've built the brand. We have the 00:12:41.620 --> 00:12:44.720 assets. But does that mean we actually have a 00:12:44.720 --> 00:12:48.340 business? Midroll sponsor placeholder. Okay, 00:12:48.379 --> 00:12:50.720 let's unpack this. We've gone through the fail 00:12:50.720 --> 00:12:54.519 fast philosophy We've geeked out on prompt engineering 00:12:54.519 --> 00:12:57.919 using chat GPT to write the prompts using sweat 00:12:57.919 --> 00:13:00.879 for realism using reference images to lock the 00:13:00.879 --> 00:13:03.120 logo He built a theoretical empire in about 20 00:13:03.120 --> 00:13:05.799 minutes But what is the big idea here because 00:13:05.799 --> 00:13:08.559 we aren't just talking about shoes This applies 00:13:08.559 --> 00:13:11.580 to coffee brands tech startups clothing lines. 00:13:11.639 --> 00:13:14.769 I think we're talking about access We are shifting 00:13:14.769 --> 00:13:17.809 from a world where capital, that $10 ,000 we 00:13:17.809 --> 00:13:20.009 talked about at the start, was the gatekeeper. 00:13:20.269 --> 00:13:23.029 To a world where iteration is the gatekeeper. 00:13:23.330 --> 00:13:25.509 Precisely. If you have the patience to learn 00:13:25.509 --> 00:13:27.649 the prompts, to use the reference anchors, to 00:13:27.649 --> 00:13:30.370 refine the textures, you can compete with a brand 00:13:30.370 --> 00:13:32.990 that has a million dollar budget. At least visually. 00:13:33.129 --> 00:13:35.190 Just to recap the toolkit for anyone taking notes. 00:13:35.529 --> 00:13:37.629 It's pretty simple. It's Nano Banana Pro for 00:13:37.629 --> 00:13:40.629 the assets. The logo, the product, the lifestyle 00:13:40.629 --> 00:13:43.629 shots. And it's cling 2 .6 for the motion and 00:13:43.629 --> 00:13:45.909 the video ads. And the bridge between them is 00:13:45.909 --> 00:13:49.590 that three -step process. Inspiration, description, 00:13:49.809 --> 00:13:52.610 via chat, GPT, and then generation. That is the 00:13:52.610 --> 00:13:54.929 secret sauce. Yeah. That is what separates the 00:13:54.929 --> 00:13:57.149 amateurs from the pros. So the source material 00:13:57.149 --> 00:13:59.470 ends with a bit of a challenge. It says the perfect 00:13:59.470 --> 00:14:01.970 time to start is now. It leaves us with this 00:14:01.970 --> 00:14:03.950 comparison table. You know, traditional means 00:14:03.950 --> 00:14:07.149 high risk, high cost. AI means low risk, time 00:14:07.149 --> 00:14:10.440 investment. I want to leave you, the listener, 00:14:10.740 --> 00:14:12.460 with something else to chew on, something that 00:14:12.460 --> 00:14:15.700 wasn't in the source, but feels inevitable. What's 00:14:15.700 --> 00:14:19.059 that? If the barrier to entry drops to zero, 00:14:19.620 --> 00:14:21.639 if anyone can create a professional -looking 00:14:21.639 --> 00:14:25.120 brand in an hour, then looking professional isn't 00:14:25.120 --> 00:14:26.799 a competitive advantage anymore. That's true. 00:14:26.899 --> 00:14:28.639 If everyone looks like Nike, then looking like 00:14:28.639 --> 00:14:30.980 Nike doesn't matter. Exactly. So the value shifts. 00:14:31.019 --> 00:14:33.299 It shifts away from the visual assets and back 00:14:33.299 --> 00:14:36.240 to the story, back to the why. You can build 00:14:36.240 --> 00:14:38.960 rival in an hour, but can you make me care about 00:14:38.960 --> 00:14:41.799 it? That is something AI still can't quite do 00:14:41.799 --> 00:14:45.259 for you. The tools make you a builder, but they 00:14:45.259 --> 00:14:47.440 don't make you a storyteller. That's still on 00:14:47.440 --> 00:14:50.059 you. It's a brave new world for builders. Thanks 00:14:50.059 --> 00:14:52.340 for diving in with us. Always a pleasure. See 00:14:52.340 --> 00:14:52.899 you on the next one.