WEBVTT 00:00:00.000 --> 00:00:02.220 When you first try creating images with AI, you 00:00:02.220 --> 00:00:05.059 probably start simple, right? Like draw a dog 00:00:05.059 --> 00:00:08.119 or maybe a futuristic city. You get a picture, 00:00:08.300 --> 00:00:10.800 it's functional. But there's this fundamental 00:00:10.800 --> 00:00:13.740 difference, this leap from just being a user 00:00:13.740 --> 00:00:15.820 to becoming a real creative director. It happens 00:00:15.820 --> 00:00:18.679 when you stop sort of asking the AI and start... 00:00:18.550 --> 00:00:21.429 well, commanding it. Absolutely. Yeah, welcome 00:00:21.429 --> 00:00:23.769 to the deep dive, everyone. Today we are opening 00:00:23.769 --> 00:00:26.210 up a whole stack of source material that basically 00:00:26.210 --> 00:00:28.769 proves the secret to getting those professional, 00:00:29.210 --> 00:00:31.730 really jaw -dropping AI images. It isn't some 00:00:31.730 --> 00:00:35.670 hidden software feature. It's the specific instruction 00:00:35.670 --> 00:00:38.450 set you feed it. So we're going to unlock your 00:00:38.450 --> 00:00:39.829 control over things you probably didn't even 00:00:39.829 --> 00:00:42.310 realize you could dictate. We're talking precise 00:00:42.310 --> 00:00:45.619 studio lighting, specific camera lenses. even 00:00:45.619 --> 00:00:47.840 dialing in the emotional texture. And this deep 00:00:47.840 --> 00:00:50.079 dive is all about tying that technical language 00:00:50.079 --> 00:00:52.939 directly back to your business goals. Exactly. 00:00:53.100 --> 00:00:55.179 We've got case studies here covering everything 00:00:55.179 --> 00:00:59.119 from tricky e -commerce standardization all the 00:00:59.119 --> 00:01:01.920 way to hyper -personalized marketing stuff. So 00:01:01.920 --> 00:01:04.680 our mission today is pretty clear. We want to 00:01:04.680 --> 00:01:06.760 show you how to speak that professional language. 00:01:07.000 --> 00:01:09.799 That way, your AI results become consistent, 00:01:09.939 --> 00:01:12.319 they become scalable, and honestly, perfect for 00:01:12.319 --> 00:01:15.959 the job. Every single time let's kind of unpack 00:01:15.959 --> 00:01:18.799 this shift in thinking okay, so the moment you 00:01:18.799 --> 00:01:21.560 bring money into the picture transactions trust 00:01:21.560 --> 00:01:24.400 Suddenly becomes the most important visual element 00:01:24.400 --> 00:01:26.579 doesn't it thinking about an e -commerce store 00:01:27.069 --> 00:01:29.689 We need to move past just a good -looking image. 00:01:29.930 --> 00:01:32.950 It needs to feel trustworthy, high value, like 00:01:32.950 --> 00:01:35.290 a real product. You got it. Take, say, a quick 00:01:35.290 --> 00:01:37.609 phone snapshot of a leather wallet. To get that 00:01:37.609 --> 00:01:39.510 ready for a high -end website, you can't just 00:01:39.510 --> 00:01:41.370 tell the AI, make the product look good. It won't 00:01:41.370 --> 00:01:42.890 cut it. You have to actually speak the language 00:01:42.890 --> 00:01:44.670 of a studio photographer. Right. So if we need 00:01:44.670 --> 00:01:46.730 that sense of professional dimension, that 3D 00:01:46.730 --> 00:01:49.469 feel, what specific lighting should we be asking 00:01:49.469 --> 00:01:52.930 for? Demanding, even. OK. So you command a three 00:01:52.930 --> 00:01:55.370 -point lighting setup. This is standard stuff. 00:01:55.659 --> 00:01:58.939 in a photo studio. You specify the soft key light, 00:01:59.099 --> 00:02:01.400 that's your main light source, maybe at a 45 00:02:01.400 --> 00:02:04.859 degree angle. Then you need a fill light. Its 00:02:04.859 --> 00:02:07.480 job is to soften the really dark shadows the 00:02:07.480 --> 00:02:09.879 key light creates. But here's the bit most people 00:02:09.879 --> 00:02:13.599 miss, the rim light. The rim light, that's usually 00:02:13.599 --> 00:02:15.860 positioned behind the product, isn't it? What 00:02:15.860 --> 00:02:17.699 effect does that give the customer looking at 00:02:17.699 --> 00:02:20.879 the image? It creates this thin, bright outline 00:02:20.879 --> 00:02:22.960 right around the object. It literally separates 00:02:22.960 --> 00:02:24.280 the product from the background. It makes it 00:02:24.280 --> 00:02:26.719 look genuinely three -dimensional, high -end, 00:02:26.919 --> 00:02:29.039 and that separation, it communicates quality 00:02:29.039 --> 00:02:31.919 instantly. And then we combine that light control 00:02:31.919 --> 00:02:34.659 with specific commands about the lens. If I want 00:02:34.659 --> 00:02:37.139 that classic, really professional look, where 00:02:37.139 --> 00:02:39.659 the wallet is sharp, but the background is just 00:02:39.659 --> 00:02:43.020 beautifully smooth, buttery, how do I ask for 00:02:43.020 --> 00:02:45.360 that? I'm guessing blurry background is too vague. 00:02:45.560 --> 00:02:48.139 Oh yeah, way too generic. You tell the AI, use 00:02:48.139 --> 00:02:50.800 an 85mm lens that compresses the scene really 00:02:50.800 --> 00:02:54.580 nicely, and pair it with a wide f2 .8 aperture. 00:02:55.080 --> 00:02:58.680 Ah, f2 .8. So that's the technical lever controlling 00:02:58.680 --> 00:03:00.979 the depth of field. How much is in focus? That's 00:03:00.979 --> 00:03:03.319 the one. That combo gives you a shallow depth 00:03:03.319 --> 00:03:05.680 of field, so your product stays perfectly sharp. 00:03:05.870 --> 00:03:09.810 tack sharp, but you get that soft, lovely blur 00:03:09.810 --> 00:03:12.349 behind it. Photographers call it bokeh. It just 00:03:12.349 --> 00:03:15.389 screams professional presentation, signals, quality, 00:03:15.990 --> 00:03:18.449 makes customers feel more confident clicking 00:03:18.449 --> 00:03:20.930 buy. That's really fascinating. It shows the 00:03:20.930 --> 00:03:24.129 AI is actually modeling real -world optics. And 00:03:24.129 --> 00:03:26.949 that level of detail, that attention, it scales 00:03:26.949 --> 00:03:28.789 up beautifully too, doesn't it? Yeah. Especially 00:03:28.789 --> 00:03:31.110 thinking about HR or marketing departments. Oh, 00:03:31.110 --> 00:03:33.069 totally. Think about the nightmare scenario. 00:03:33.229 --> 00:03:35.530 Trying to standardize team headshots when everyone 00:03:35.530 --> 00:03:37.849 sends in photos from home. You know, terrible 00:03:37.849 --> 00:03:41.009 lighting, messy backgrounds. Well, this source 00:03:41.009 --> 00:03:43.449 material we looked at suggests a prompt that 00:03:43.449 --> 00:03:45.879 solves that. Like, instantly. OK, how do you 00:03:45.879 --> 00:03:48.500 unify dozens of really mismatched photos just 00:03:48.500 --> 00:03:51.240 using a text prompt? You demand total unity. 00:03:51.479 --> 00:03:54.659 You instruct. Identical background. Specify the 00:03:54.659 --> 00:03:58.020 material. Even, say, light gray concrete, subtly 00:03:58.020 --> 00:04:01.719 blurred. Then, strict framing and sizing. Mandate 00:04:01.719 --> 00:04:05.460 a 1 .1 square ratio. And this is critical. Tell 00:04:05.460 --> 00:04:08.020 the AI to make sure everyone's eyes are level 00:04:08.020 --> 00:04:10.860 on the same horizontal line. That alone would 00:04:10.860 --> 00:04:13.259 make a huge difference. The final gallery wouldn't 00:04:13.259 --> 00:04:15.879 look like some amateur collage. But what about 00:04:15.879 --> 00:04:17.939 the lighting? How do we get sophisticated lighting 00:04:17.939 --> 00:04:20.500 without sending everyone to a studio? You specify 00:04:20.500 --> 00:04:22.000 the Rembrandt lighting technique. You're not 00:04:22.000 --> 00:04:24.180 just asking for good lighting. Rembrandt lighting, 00:04:24.180 --> 00:04:27.319 it creates this very specific small triangle 00:04:27.319 --> 00:04:29.240 of light on the cheek that's opposite the main 00:04:29.240 --> 00:04:31.560 light source. Ah, OK. So the lighting itself 00:04:31.560 --> 00:04:34.709 has an artistic quality. It adds shadow, depth, 00:04:35.209 --> 00:04:37.389 a bit of gravitas, maybe. Precisely. It looks 00:04:37.389 --> 00:04:39.910 intentional. It looks professional, unified. 00:04:40.269 --> 00:04:42.410 It transforms basic headshots into something 00:04:42.410 --> 00:04:44.250 that feels like it came from a proper high -end 00:04:44.250 --> 00:04:47.829 corporate photo shoot. So, studying back, if 00:04:47.829 --> 00:04:49.870 the main goal is standardization across a whole 00:04:49.870 --> 00:04:52.370 brand, what's the single biggest advantage of 00:04:52.370 --> 00:04:54.769 using these really specific technical terms? 00:04:55.370 --> 00:04:58.009 Specificity ensures consistent output, which 00:04:58.009 --> 00:05:00.930 instantly communicates professionalism and company 00:05:00.930 --> 00:05:04.120 unity. Right, so moving beyond just the basics 00:05:04.120 --> 00:05:07.120 of branding, these detailed prompts let us actually 00:05:07.120 --> 00:05:10.199 control the narrative. We can create massive 00:05:10.199 --> 00:05:13.540 perceived value just through hyper -personalization 00:05:13.540 --> 00:05:15.600 and playing with cinematic scale. Okay, let's 00:05:15.600 --> 00:05:17.959 talk about hyper -personalized sales outreach. 00:05:18.300 --> 00:05:21.959 Say you're emailing a potential client, ABC Corp, 00:05:22.160 --> 00:05:23.860 their logo's orange, they're based in Singapore, 00:05:23.920 --> 00:05:26.240 you want your email to really stand out. Generic 00:05:26.240 --> 00:05:29.000 stock photos are just... useless here, right? 00:05:29.120 --> 00:05:31.319 Totally worthless. So what you do is you embed 00:05:31.319 --> 00:05:34.569 their brand details right into an image. maybe 00:05:34.569 --> 00:05:36.449 an image of yourself. Let's say you're pictured 00:05:36.449 --> 00:05:38.810 working at a desk in the image. You tell the 00:05:38.810 --> 00:05:40.730 AI, change the notebook cover in the shot to 00:05:40.730 --> 00:05:42.649 the client's exact brand hex code. Let's say 00:05:42.649 --> 00:05:46.050 it's hashtag FFA500 for Orange, and then maybe 00:05:46.050 --> 00:05:48.470 ask it to subtly print their logo on that notebook 00:05:48.470 --> 00:05:51.110 cover too. The subtlety is key there. Making 00:05:51.110 --> 00:05:53.370 it feel natural like it belongs in the scene, 00:05:53.610 --> 00:05:55.970 not just obviously pasted on. That shows an incredible 00:05:55.970 --> 00:05:57.730 amount of effort, even if it was AI generated. 00:05:58.189 --> 00:06:00.430 It really does. And then for the background, 00:06:00.850 --> 00:06:03.670 don't just put office window. Specify the Singapore 00:06:03.670 --> 00:06:06.810 skyline even better. With the iconic Marina Bay 00:06:06.810 --> 00:06:08.689 Sands building clearly visible in the distance, 00:06:09.209 --> 00:06:10.970 you didn't just mention an office, you mentioned 00:06:10.970 --> 00:06:14.089 their city, their landmark. That level of context, 00:06:15.129 --> 00:06:17.170 it's kind of startling. Yeah, it makes the client 00:06:17.170 --> 00:06:20.819 feel seen. important, recognized. It probably 00:06:20.819 --> 00:06:23.000 stops them scrolling through their inbox, right? 00:06:23.279 --> 00:06:24.720 It shows you respect their business enough to 00:06:24.720 --> 00:06:26.220 do your homework. And we can apply that same 00:06:26.220 --> 00:06:29.000 idea, that power, to scale and perceived value, 00:06:29.279 --> 00:06:32.199 too. Like, if you're promoting a local event, 00:06:32.319 --> 00:06:35.319 maybe a workshop, a standard digital flyer, kind 00:06:35.319 --> 00:06:37.899 of boring, let's be honest. The goal here is 00:06:37.899 --> 00:06:40.839 to command this impressive, maybe even global, 00:06:41.120 --> 00:06:43.279 scene for that simple flyer. Okay, so instead 00:06:43.279 --> 00:06:45.459 of just showing the flyer, we tell the AI to 00:06:45.459 --> 00:06:47.920 place it, maybe on a giant digital billboard. 00:06:48.000 --> 00:06:50.439 Somewhere iconic, like Shibuya Crossing in Tokyo, 00:06:50.759 --> 00:06:53.779 at night. Suddenly our local event lifts internationally 00:06:53.779 --> 00:06:57.399 famous. Yes, exactly. And then you layer in the 00:06:57.399 --> 00:06:59.240 technical cinematography commands. You demand 00:06:59.240 --> 00:07:01.379 a low angle shot. Looking up at the billboard, 00:07:01.399 --> 00:07:03.540 that makes it look monumental, really important. 00:07:04.040 --> 00:07:06.040 And maybe add a deliberate motion blur effect 00:07:06.040 --> 00:07:08.319 to the crowds and the cars whizzing by below. 00:07:08.579 --> 00:07:11.079 Ah, the motion blur adds that cinematic tension. 00:07:11.310 --> 00:07:15.050 It emphasizes this is a busy, living, global 00:07:15.050 --> 00:07:19.189 hub, and your ad is the clear, stable focus point 00:07:19.189 --> 00:07:21.790 within all that energy. Whoa! I mean, just imagine 00:07:21.790 --> 00:07:24.050 scaling that feeling of global importance, right? 00:07:24.050 --> 00:07:26.569 That high -octane city vibe, applying that instantly 00:07:26.569 --> 00:07:28.670 to all your social media posts. You're not just 00:07:28.670 --> 00:07:30.310 generating a picture anymore, you're generating 00:07:30.310 --> 00:07:33.889 emotion. So, thinking about maximum impact, maximum 00:07:33.889 --> 00:07:36.990 surprise for the viewer, which hits harder. Is 00:07:36.990 --> 00:07:39.949 it the hyper -personalization, like seeing your 00:07:39.949 --> 00:07:42.889 own company's hex code? Or is it that dramatic 00:07:42.889 --> 00:07:45.170 cinematic scale, the motion blur at Shibuya? 00:07:45.269 --> 00:07:47.569 It's the contextual relevance, like placing the 00:07:47.569 --> 00:07:49.709 client's logo naturally or making a simple ad 00:07:49.709 --> 00:07:51.970 look internationally famous. Okay, let's shift 00:07:51.970 --> 00:07:54.990 focus now to automation. Because often the challenge 00:07:54.990 --> 00:07:57.870 isn't just making one beautiful image. It's building 00:07:57.870 --> 00:08:00.589 a repeatable system, a workflow, to generate 00:08:00.589 --> 00:08:03.629 maybe hundreds of them, all consistently, for 00:08:03.629 --> 00:08:06.290 large projects. Yeah, think about a wedding photographer. 00:08:06.470 --> 00:08:08.959 They might need to edit hundreds, maybe thousands 00:08:08.959 --> 00:08:11.720 of photos, all with the same romantic kind of 00:08:11.720 --> 00:08:14.660 vintage style. Doing that manually, it's just, 00:08:14.660 --> 00:08:17.339 it's prohibitive. It takes way too long. So the 00:08:17.339 --> 00:08:20.079 prompt here isn't really for generating one photo 00:08:20.079 --> 00:08:23.040 from scratch. It's more about building an application, 00:08:23.300 --> 00:08:26.540 like AI Core, around a single preset prompt, 00:08:26.660 --> 00:08:29.379 a preset that every single uploaded image will 00:08:29.379 --> 00:08:31.660 then run through for styling. Precisely. And 00:08:31.660 --> 00:08:33.720 that preset prompt has to define the aesthetic 00:08:33.720 --> 00:08:36.279 very clearly. Things like vintage film look, 00:08:36.759 --> 00:08:39.080 soft golden hour lighting. Here's the really 00:08:39.080 --> 00:08:41.039 tricky technical problem that always crops up 00:08:41.039 --> 00:08:42.860 in batch editing, especially with styles like 00:08:42.860 --> 00:08:45.740 that. Color correction. So the prompt absolutely 00:08:45.740 --> 00:08:48.539 must mandate making sure the white of the wedding 00:08:48.539 --> 00:08:50.980 dress stays pure white and does not turn yellow. 00:08:51.500 --> 00:08:53.639 Why is that specific instruction about the white 00:08:53.639 --> 00:08:56.059 so crucial? What happens otherwise? Well, when 00:08:56.059 --> 00:08:58.919 you dial in that golden hour look or a vintage 00:08:58.919 --> 00:09:01.419 film aesthetic, you're essentially flooding the 00:09:01.419 --> 00:09:04.440 image with warm tones, yellows, oranges, reds. 00:09:04.600 --> 00:09:07.899 The AI, it naturally tends towards color saturation 00:09:07.899 --> 00:09:09.980 or sometimes what we call prompt drift. It gets 00:09:09.980 --> 00:09:12.740 pulled by the dominant style request. Ah, so 00:09:12.740 --> 00:09:15.100 the warmth bleeds into the white fabric, making 00:09:15.100 --> 00:09:17.919 it look yellowish or muddy. Exactly. You know, 00:09:17.919 --> 00:09:20.360 I still wrestle with prompt drift myself sometimes, 00:09:20.419 --> 00:09:22.840 especially with really delicate colors like pure 00:09:22.840 --> 00:09:26.299 white or specific pastels. So building that specific 00:09:26.299 --> 00:09:29.240 guardrail into the preset prompt telling it explicitly 00:09:29.240 --> 00:09:31.940 to keep that wedding dress pure white is incredibly 00:09:31.940 --> 00:09:34.360 smart. It prevents just massive amounts of manual 00:09:34.360 --> 00:09:37.259 rework later and keeps the client happy. That 00:09:37.259 --> 00:09:39.879 kind of hyper -precision must also be absolutely 00:09:39.879 --> 00:09:42.700 essential for things like data -driven A -B marketing 00:09:42.700 --> 00:09:45.379 tests. Say a travel company is testing ads for 00:09:45.379 --> 00:09:48.279 different destinations in Vietnam. Maybe Da Nang 00:09:48.279 --> 00:09:51.480 versus Ha Giang versus Phu Quoc. Data integrity 00:09:51.480 --> 00:09:54.000 is paramount there. Oh, totally. The integrity 00:09:54.000 --> 00:09:56.100 is all about testing only one variable at a time, 00:09:56.279 --> 00:09:58.879 the destination image. So your instruction to 00:09:58.879 --> 00:10:02.320 the AI has to be crystal clear. Keep 100 % of 00:10:02.320 --> 00:10:05.899 the text, OverWay, Logo, and CTA button identical 00:10:05.899 --> 00:10:08.259 across all variations. Right, because if the 00:10:08.259 --> 00:10:10.559 logo shifts slightly, or the font weight changes 00:10:10.559 --> 00:10:13.419 even a tiny bit between the Da Nang version and 00:10:13.419 --> 00:10:16.559 the Fu Quoc version, you're accidentally testing 00:10:16.559 --> 00:10:18.860 multiple variables, your data gets compromised. 00:10:19.019 --> 00:10:21.559 You don't know why one ad performed better. That's 00:10:21.559 --> 00:10:23.500 the key. You lock down everything except the 00:10:23.500 --> 00:10:25.860 image, then you detail the image variations, 00:10:25.940 --> 00:10:29.919 maybe focusing on the feeling. Piling pass, dramatic 00:10:29.919 --> 00:10:32.200 layers of mountains and mist for one option, 00:10:32.460 --> 00:10:35.620 or maybe fukwok, pristine white sand, clear turquoise 00:10:35.620 --> 00:10:38.279 water for the tranquil option. The visual changes, 00:10:38.440 --> 00:10:40.320 maybe drastically, but all the other crucial 00:10:40.320 --> 00:10:43.659 ad elements locked down tight. And this consistency 00:10:43.659 --> 00:10:45.659 challenge, it also applies directly to things 00:10:45.659 --> 00:10:48.759 like brand mascots. Right. If you have, say, 00:10:49.159 --> 00:10:51.500 a robot mascot that needs to show lots of different 00:10:51.500 --> 00:10:53.559 emotions across different campaigns or website 00:10:53.559 --> 00:10:55.980 pages, the robot itself can't suddenly change 00:10:55.980 --> 00:10:58.759 shape or color. Exactly. You have to demand absolute 00:10:58.759 --> 00:11:01.059 consistency of its shape, its proportions, its 00:11:01.059 --> 00:11:03.320 metallic color, everything across all the outputs. 00:11:04.080 --> 00:11:07.019 Then you layer on the specific emotional action 00:11:07.019 --> 00:11:09.629 you need. If the robot needs to look confused, 00:11:09.809 --> 00:11:11.990 you might ask for its head tilted, maybe with 00:11:11.990 --> 00:11:13.590 a little question mark symbol spinning above 00:11:13.590 --> 00:11:16.149 it. So the emotion is dynamic, the pose changes, 00:11:16.789 --> 00:11:19.289 but the core brand asset, the robot's design, 00:11:19.629 --> 00:11:22.210 remains perfectly static. And just thinking practically, 00:11:23.029 --> 00:11:25.169 for maximum efficiency down the line for the 00:11:25.169 --> 00:11:28.509 design team, you add that final, really practical, 00:11:28.710 --> 00:11:31.549 technical requirement. Ask for PNG images with 00:11:31.549 --> 00:11:34.580 a transparent background. Ah, yes. That saves 00:11:34.580 --> 00:11:36.679 just a huge amount of time later when someone 00:11:36.679 --> 00:11:39.139 needs to place that mascot image onto a complex 00:11:39.139 --> 00:11:41.659 web page background or into a brochure layout. 00:11:42.000 --> 00:11:44.820 no manual cutting out needed. So, when you're 00:11:44.820 --> 00:11:47.159 dealing with potentially hundreds, even thousands 00:11:47.159 --> 00:11:49.539 of images in these automated workflows, what 00:11:49.539 --> 00:11:52.000 is the single biggest operational risk that using 00:11:52.000 --> 00:11:54.840 detail prompting really helps mitigate? It minimizes 00:11:54.840 --> 00:11:57.620 inconsistencies, ensuring every output maintains 00:11:57.620 --> 00:12:00.120 brand style and saves hours of manual rework. 00:12:00.480 --> 00:12:02.539 Hashtag, tag, tag, big idea, recap and outro. 00:12:02.779 --> 00:12:05.799 Okay, so the big idea weaving through all of 00:12:05.799 --> 00:12:08.299 this, it's really this fundamental shift in authority, 00:12:08.340 --> 00:12:11.039 isn't it? You absolutely have to move away from 00:12:11.039 --> 00:12:13.620 just sort of asking the AI for a picture, you 00:12:13.620 --> 00:12:15.559 need to start telling it exactly what you want 00:12:15.559 --> 00:12:18.659 it to create. And you do that using the sophisticated 00:12:18.659 --> 00:12:22.240 domain -specific language, the technical vocabulary 00:12:22.240 --> 00:12:25.480 of photography, of lighting, of design. That's 00:12:25.480 --> 00:12:28.559 how you achieve results that are precise, consistent, 00:12:28.700 --> 00:12:31.980 and crucially Instantly scalable. Yeah based 00:12:31.980 --> 00:12:33.620 on everything we've looked at in the source material 00:12:33.620 --> 00:12:36.240 It seems like every truly successful prompt the 00:12:36.240 --> 00:12:38.139 ones that deliver professional results They're 00:12:38.139 --> 00:12:40.500 basically a detailed instruction set and they 00:12:40.500 --> 00:12:43.240 usually describe three key things the precise 00:12:43.240 --> 00:12:45.679 setting the specific feeling or emotion you want 00:12:45.679 --> 00:12:48.289 to evoke and underpinning it all the business 00:12:48.289 --> 00:12:49.990 goal you're trying to achieve, that level of 00:12:49.990 --> 00:12:52.129 detail, that depth, it elevates you from just 00:12:52.129 --> 00:12:55.289 being a passive user to being an active, creative 00:12:55.289 --> 00:12:57.769 image director. You're taking back control from 00:12:57.769 --> 00:13:00.210 the machine in a way. Absolutely. Which means 00:13:00.210 --> 00:13:02.970 the key to getting these game -changing images, 00:13:04.210 --> 00:13:07.169 it's genuinely in your hands. It's not locked 00:13:07.169 --> 00:13:09.309 away in some secret new tech that only a few 00:13:09.309 --> 00:13:12.669 people have access to. And if using this kind 00:13:12.669 --> 00:13:15.090 of highly technical language makes our image 00:13:15.090 --> 00:13:18.309 output so much better, it really raises an important 00:13:18.309 --> 00:13:20.389 question, doesn't it? What happens when we start 00:13:20.389 --> 00:13:22.870 applying that same level of technical design 00:13:22.870 --> 00:13:25.769 -focused specificity to the prompts we write 00:13:25.769 --> 00:13:30.419 for, say, text generation or even video AI? That's 00:13:30.419 --> 00:13:31.820 a really interesting thought to leave people 00:13:31.820 --> 00:13:33.240 with. That's the challenge for you listening 00:13:33.240 --> 00:13:35.620 today. Maybe start practicing just by applying 00:13:35.620 --> 00:13:38.019 one new technical term from our discussion, perhaps 00:13:38.019 --> 00:13:40.980 F2 .8, to control your depth of field or specify 00:13:40.980 --> 00:13:43.019 Rembrandt lighting for more sophisticated shadows 00:13:43.019 --> 00:13:45.480 into your very next image prompt. Just see the 00:13:45.480 --> 00:13:46.940 immediate professional difference it can make. 00:13:47.440 --> 00:13:48.820 Thank you so much for joining us for this deep 00:13:48.820 --> 00:13:49.000 drive.