Sensitive content

This media may contain sensitive content.

Загрузка видео...

Не удалось загрузить видео

На главную

Wan 2.2. Testing Composition, Camera Movement, Cohesion and Consistency. #AI #Wan22 #GayAI #Homoerotic #Erotica #BarnLove #1930s #Interracial #Queer

Homopolitan_AI's profile picture

Homopolitan_AI

1,482 subscribers

18,627 просмотров • 11 месяцев назад •via X (Twitter)

NSFW#AI#Wan22#GayAI#Homoerotic#Erotica#BarnLove#1930s#Interracial#Queer
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Introducing Wan 2.2 A video model designed for precise control over every shot, lighting, color, camera and composition Available now in the Freepik AI Suite
0:22

Introducing Wan 2.2 A video model designed for precise control over every shot, lighting, color, camera and composition Available now in the Freepik AI Suite

Magnific

15,183 просмотров • 11 месяцев назад

Seedance 2.0 and Seedance 2.0 Fast deliver cinematic frames, high-speed camera movement, precise follow shots, and unified style with strong consistency. Build smoother, sharper AI videos on BudgetPixel AI.
0:28

Seedance 2.0 and Seedance 2.0 Fast deliver cinematic frames, high-speed camera movement, precise follow shots, and unified style with strong consistency. Build smoother, sharper AI videos on BudgetPixel AI.

BudgetPixel AI

380,088 просмотров • 1 месяц назад

Testing Hailuo AI (MiniMax) T2V-01-Director Model (Part II) In this demo, I only used the [Shake] camera preset and combined it with tokens like "handheld camera shots," "handheld camera wobbles," and "handheld camera swings" to capture natural handheld movement. → text2vid: Hailuo AI (MiniMax) → music: Udio → editing: CapCut
0:44

Testing Hailuo AI (MiniMax) T2V-01-Director Model (Part II) In this demo, I only used the [Shake] camera preset and combined it with tokens like "handheld camera shots," "handheld camera wobbles," and "handheld camera swings" to capture natural handheld movement. → text2vid: Hailuo AI (MiniMax) → music: Udio → editing: CapCut

Dan Pradana

12,023 просмотров • 1 год назад

Testing Wan 2.2 on It’s able to do cool stuff with camera that I can’t seem to do with the other models. Its the dutch angles, the creamy rack focus, and it goes in for really tight wide angle closeups with extremely high detail. And it’s open source!
0:40

Testing Wan 2.2 on It’s able to do cool stuff with camera that I can’t seem to do with the other models. Its the dutch angles, the creamy rack focus, and it goes in for really tight wide angle closeups with extremely high detail. And it’s open source!

CoffeeVectors

27,981 просмотров • 11 месяцев назад

🛋️ Get Comfy With Comfy - Wan 2.2 Animate In just 3 minutes, we’ll cover a powerful model for character animation and movement inside ComfyUI. This short tutorial will give you the essentials so you can start experimenting with character-driven AI animation right away.
3:10

🛋️ Get Comfy With Comfy - Wan 2.2 Animate In just 3 minutes, we’ll cover a powerful model for character animation and movement inside ComfyUI. This short tutorial will give you the essentials so you can start experimenting with character-driven AI animation right away.

ComfyUI

28,343 просмотров • 9 месяцев назад

I’m testing out Viggle and while it has weak spots like motion jitter, and lower fidelity, it has one of the closest approximations to a reusable Ai assets I’ve used. I’ll do a longer multi shot test and see what I can get away with on camera movement.
0:33

I’m testing out Viggle and while it has weak spots like motion jitter, and lower fidelity, it has one of the closest approximations to a reusable Ai assets I’ve used. I’ll do a longer multi shot test and see what I can get away with on camera movement.

Jon Finger

29,795 просмотров • 1 год назад

🎬 Want to create cinematic AI videos in seconds? Here’s exactly how I do it using Seedance 2.0 on Yapper: Most people get this wrong… your results depend on how well you *direct* the AI. Here’s the workflow 👇 ✅ Step 1: Open Yapper Go to Create → Video ✅ Step 2: Choose your model Select Seedance 2.0 This model delivers smooth motion, dynamic camera angles, realistic lighting, and strong scene consistency. ✅ Step 3: Write a strong prompt Your prompt is your director. Be specific: Camera movement (pan, zoom, tracking shot) Lighting (golden hour, neon, soft shadows) Environment (city, desert, luxury interior) ✅ Step 4: Add a reference image (optional) Use this to guide style, composition, or subject for more control and consistency. ✅ Step 5: Hit Generate In seconds, your idea turns into a cinematic sequence. 🚀 Pro tip: If it’s not perfect, tweak your prompt — small changes = big difference. From idea → cinematic video in seconds.
1:00

🎬 Want to create cinematic AI videos in seconds? Here’s exactly how I do it using Seedance 2.0 on Yapper: Most people get this wrong… your results depend on how well you *direct* the AI. Here’s the workflow 👇 ✅ Step 1: Open Yapper Go to Create → Video ✅ Step 2: Choose your model Select Seedance 2.0 This model delivers smooth motion, dynamic camera angles, realistic lighting, and strong scene consistency. ✅ Step 3: Write a strong prompt Your prompt is your director. Be specific: Camera movement (pan, zoom, tracking shot) Lighting (golden hour, neon, soft shadows) Environment (city, desert, luxury interior) ✅ Step 4: Add a reference image (optional) Use this to guide style, composition, or subject for more control and consistency. ✅ Step 5: Hit Generate In seconds, your idea turns into a cinematic sequence. 🚀 Pro tip: If it’s not perfect, tweak your prompt — small changes = big difference. From idea → cinematic video in seconds.

Kuria | AI

43,685 просмотров • 2 месяцев назад

🎬 Cinematic Consistency in the Age of AI Video: Why Prompt Structure Matters ⠀ I created a custom GPT built specifically to solve one of the biggest problems in AI filmmaking: character and visual continuity across shots and sequences. ⠀ Too often, tools like Midjourney Video or Veo3 generate beautiful moments — but fall apart when you try to tell a story. Characters change. Outfits glitch. Locations shift. ⠀ This GPT helps with that. It’s trained to write highly structured cinematic prompts, with attention to: — Composition & camera motion — Character appearance and consistency — Lens & lighting consistency — Emotional tone, color grading, and even ambient audio cues ⠀ 💡 The result? A prompt system that actually thinks like a director and helps you like your DP or 1st AD. ⠀ Here’s a side-by-side comparison using the same simple sequence idea in Veo3 and Midjourney Video, both using my GPT. (don't judge the video result but the GPT assistance as I made this as quickly as possible) 👇
1:36

🎬 Cinematic Consistency in the Age of AI Video: Why Prompt Structure Matters ⠀ I created a custom GPT built specifically to solve one of the biggest problems in AI filmmaking: character and visual continuity across shots and sequences. ⠀ Too often, tools like Midjourney Video or Veo3 generate beautiful moments — but fall apart when you try to tell a story. Characters change. Outfits glitch. Locations shift. ⠀ This GPT helps with that. It’s trained to write highly structured cinematic prompts, with attention to: — Composition & camera motion — Character appearance and consistency — Lens & lighting consistency — Emotional tone, color grading, and even ambient audio cues ⠀ 💡 The result? A prompt system that actually thinks like a director and helps you like your DP or 1st AD. ⠀ Here’s a side-by-side comparison using the same simple sequence idea in Veo3 and Midjourney Video, both using my GPT. (don't judge the video result but the GPT assistance as I made this as quickly as possible) 👇

Jesus Plaza

51,432 просмотров • 1 год назад

A Chinese guy spent $20 on an AI video workflow and claims he can replace anyone in any video. He also recorded a full tutorial on creating your own AI avatar. At first, he was just trying to avoid editing the same clip by hand over and over again He spent $20 on the API credits Then he uploaded one normal video and one picture of the character he wanted n8n sent both files to Cloudinary Cloudinary turned them into clean links Wan 2.2 took the video, took the image, and rebuilt the person inside the clip The face changed The body stayed The movement stayed The camera stayed Then he realized this is not a face swap tool anymore It is a machine for making 100 versions of the same video AI is getting scary because now the product is not the video The product is whoever you put inside it
9:16

A Chinese guy spent $20 on an AI video workflow and claims he can replace anyone in any video. He also recorded a full tutorial on creating your own AI avatar. At first, he was just trying to avoid editing the same clip by hand over and over again He spent $20 on the API credits Then he uploaded one normal video and one picture of the character he wanted n8n sent both files to Cloudinary Cloudinary turned them into clean links Wan 2.2 took the video, took the image, and rebuilt the person inside the clip The face changed The body stayed The movement stayed The camera stayed Then he realized this is not a face swap tool anymore It is a machine for making 100 versions of the same video AI is getting scary because now the product is not the video The product is whoever you put inside it

0xTria

28,269 просмотров • 26 дней назад

This is BLITZ 🏍️ starring myself. I've tested a workflow that combines two of the best AI image editing models right now, Nano Banana and Seedream 4. Using these together feels like a superpower. Seedream 4 for instance generates much more dynamic camera angles BUT Nano Banana does a better job with overall consistency, so utilizing both is super beneficial. I wanted to create a sequence where I'm riding a Cyberbike with an electric booster that overclocks the motor to deliver short, high-risk bursts of extreme speed. The location is a white sand desert. I decided to also test Hailuo AI (MiniMax) start and end frame feature which in my opinion is one of the best since it understands so much about context and camera movement. Start and ends frames let you have much better control over camera movement but where Hailuo truly excels is that it understands speed and camera shake which are essential for any fast paced action sequences. The music was generated with Suno 4.5+ and some additional sound effects like the bike and wind in the desert were sourced from Artlist. Edited with Capcut. There's some more tips and tricks in the thread 👍👍
0:39

This is BLITZ 🏍️ starring myself. I've tested a workflow that combines two of the best AI image editing models right now, Nano Banana and Seedream 4. Using these together feels like a superpower. Seedream 4 for instance generates much more dynamic camera angles BUT Nano Banana does a better job with overall consistency, so utilizing both is super beneficial. I wanted to create a sequence where I'm riding a Cyberbike with an electric booster that overclocks the motor to deliver short, high-risk bursts of extreme speed. The location is a white sand desert. I decided to also test Hailuo AI (MiniMax) start and end frame feature which in my opinion is one of the best since it understands so much about context and camera movement. Start and ends frames let you have much better control over camera movement but where Hailuo truly excels is that it understands speed and camera shake which are essential for any fast paced action sequences. The music was generated with Suno 4.5+ and some additional sound effects like the bike and wind in the desert were sourced from Artlist. Edited with Capcut. There's some more tips and tricks in the thread 👍👍

Travis Davids

60,599 просмотров • 10 месяцев назад

19 year old from Warsaw made $11,200 in a month without ever showing his face films 1 movement on his phone → drops an AI character into it → uploads it as a faceless channel his entire "actor" is one photo from ChatGPT + an image gen model then Wan 2.2 Animate transfers expressions and full body motion onto that character $0 on a studio. $0 on actors. $0 on an editor. Week 1 > $1,340 Week 2 > $2,980 Week 3 > $6,880 1 month > $11,200 the secret is Animate Mix -> it replaces you in the frame but keeps your lighting and background the instruction is already publicly available
8:10

19 year old from Warsaw made $11,200 in a month without ever showing his face films 1 movement on his phone → drops an AI character into it → uploads it as a faceless channel his entire "actor" is one photo from ChatGPT + an image gen model then Wan 2.2 Animate transfers expressions and full body motion onto that character $0 on a studio. $0 on actors. $0 on an editor. Week 1 > $1,340 Week 2 > $2,980 Week 3 > $6,880 1 month > $11,200 the secret is Animate Mix -> it replaces you in the frame but keeps your lighting and background the instruction is already publicly available

Ridark

18,842 просмотров • 1 месяц назад

No, this is NOT another episode of The Last of Us. It is FULLY generated with AI! This is Luma #Ray2 🚀 ⚡️Presenting "RUN"⚡️ Having early access to Ray2, this trailer is created entirely with an advanced state-of-the-AI-art text-to-video model by #LumaDreamMachine The quality of shots, visual consistency, camera movement, and precise understanding of the prompts make it a perfect fit for cinematic action! Welcome to the new era of AI filmmaking. 2025 is gonna be awesome! Thanks to the amazing Luma Team for this fantastic new tool! Kiri David "Navi" Romero tuhin Dan Kondratyuk amit Theo Panagiotopoulos Samarth Sinha Jiaming Song @alexyu00 Andrew Prifer Anjney Midha & all!
1:42

No, this is NOT another episode of The Last of Us. It is FULLY generated with AI! This is Luma #Ray2 🚀 ⚡️Presenting "RUN"⚡️ Having early access to Ray2, this trailer is created entirely with an advanced state-of-the-AI-art text-to-video model by #LumaDreamMachine The quality of shots, visual consistency, camera movement, and precise understanding of the prompts make it a perfect fit for cinematic action! Welcome to the new era of AI filmmaking. 2025 is gonna be awesome! Thanks to the amazing Luma Team for this fantastic new tool! Kiri David "Navi" Romero tuhin Dan Kondratyuk amit Theo Panagiotopoulos Samarth Sinha Jiaming Song @alexyu00 Andrew Prifer Anjney Midha & all!

Digital da Vincis

18,862 просмотров • 1 год назад

I tested Wizstar by creating a cinematic smartphone promo to see how well it could handle commercial-style product videos. And honestly… the result looked surprisingly close to a real ad. What stood out most was how well it handled: → cinematic camera movement → gaming visuals → smooth scene transitions → product consistency across every shot No complex editing. No huge production setup. Just concept → AI-generated commercial. This feels especially powerful for e-commerce brands pushing short-form ads at scale. AI video tools are moving way beyond simple “content generation.” They’re starting to handle much larger parts of the creative workflow. Try it here:
0:56

I tested Wizstar by creating a cinematic smartphone promo to see how well it could handle commercial-style product videos. And honestly… the result looked surprisingly close to a real ad. What stood out most was how well it handled: → cinematic camera movement → gaming visuals → smooth scene transitions → product consistency across every shot No complex editing. No huge production setup. Just concept → AI-generated commercial. This feels especially powerful for e-commerce brands pushing short-form ads at scale. AI video tools are moving way beyond simple “content generation.” They’re starting to handle much larger parts of the creative workflow. Try it here:

RAVI KUMAR SAHU

60,601 просмотров • 1 месяц назад

AI video is quietly shifting from generation to direction. After testing Ray 3.2 on a restaurant launch concept, that became pretty obvious. I started with two static images: one exterior shot, one interior dining scene. A few minutes later, I had a cinematic sequence that actually felt connected. The biggest improvement isn't just realism. It's control. Most AI video workflows still look like this: prompt → generate → regenerate → regenerate again. You're mostly hoping the model understands what you want. Ray 3.2 feels different. Instead of generating isolated clips, you can guide how one scene evolves into the next. You can shape transitions, control camera movement, preserve visual consistency, and build sequences that feel intentionally directed. For a simple restaurant commercial, that makes a huge difference. The exterior naturally flows into the dining experience. The pacing feels deliberate. The motion feels cinematic. The story feels connected. And that's what stood out to me most. For the first time, AI video felt less like prompting and more like directing. Every frame, yours to direct. Try Luma today using my link:
0:31

AI video is quietly shifting from generation to direction. After testing Ray 3.2 on a restaurant launch concept, that became pretty obvious. I started with two static images: one exterior shot, one interior dining scene. A few minutes later, I had a cinematic sequence that actually felt connected. The biggest improvement isn't just realism. It's control. Most AI video workflows still look like this: prompt → generate → regenerate → regenerate again. You're mostly hoping the model understands what you want. Ray 3.2 feels different. Instead of generating isolated clips, you can guide how one scene evolves into the next. You can shape transitions, control camera movement, preserve visual consistency, and build sequences that feel intentionally directed. For a simple restaurant commercial, that makes a huge difference. The exterior naturally flows into the dining experience. The pacing feels deliberate. The motion feels cinematic. The story feels connected. And that's what stood out to me most. For the first time, AI video felt less like prompting and more like directing. Every frame, yours to direct. Try Luma today using my link:

SONIA

41,538 просмотров • 1 месяц назад

This is crazy! With LTX 2 in ComfyUI i was able to keep my original camera movement + it did a full lip-sync on the singing haha. All running locally. I made the original video back in 2021 when Metahumans first came out and we were working in Unreal Engine 4. This music video was some kind of result from my first experiments with Livelink and Control Rig at that time. Now the new AI improved version is not perfect and i didn’t train any lora for character consistency yet. I also didn’t have the original video file anymore so I downloaded the Youtube version as input haha. However the camera result + facial expressions are really promising! I wonder how it would handle more complex and fast shots though.
1:38

This is crazy! With LTX 2 in ComfyUI i was able to keep my original camera movement + it did a full lip-sync on the singing haha. All running locally. I made the original video back in 2021 when Metahumans first came out and we were working in Unreal Engine 4. This music video was some kind of result from my first experiments with Livelink and Control Rig at that time. Now the new AI improved version is not perfect and i didn’t train any lora for character consistency yet. I also didn’t have the original video file anymore so I downloaded the Youtube version as input haha. However the camera result + facial expressions are really promising! I wonder how it would handle more complex and fast shots though.

Jerome | InsaneUnreal

41,338 просмотров • 5 месяцев назад

is genspark ($50M ARR in 6mo) underrated next to chatgpt, claude, gemini, and manus? i spent 34 minutes testing its “super ai agent” and how it actually works. 1/ you write one prompt and it runs through gpt-5, claude sonnet, gemini 2.5, and mistral large then gives you the best response. 2/ you upload one image and it runs it through nano banana, flux, and bytedance then returns the most realistic version. 3/ you generate a video and genspark routes it through wan 2.2, pika labs, gen-2, and runway ml combining the best clips and motion. 4/ you paste an idea for slides and it uses gpt-5, claude 3 opus, gemini advanced, and genspark’s slide model to create a full deck with visuals and market data. 5/ AI slides and building a fundraising deck with a few prompts (wasn't perfect, but it was a good start for my personal workflow) what's cool about genspark is it's inexpensive and does it all. kinda like the costco of LLMs. TLDR; you hand it one prompt + input, and genspark routes sub-tasks across models and tools behind the scenes. it’s orchestration over one ideal model. perspective shift. instead of jumping from model to model, llm to llm, genspark shows what the next generation of ai will look like orchestration over obsession. and for the rest of us using it, the only thing that matters is getting the best outcome. what other AI app should I review next? i'll keep doing it if it's helpful for folks.
34:35

is genspark ($50M ARR in 6mo) underrated next to chatgpt, claude, gemini, and manus? i spent 34 minutes testing its “super ai agent” and how it actually works. 1/ you write one prompt and it runs through gpt-5, claude sonnet, gemini 2.5, and mistral large then gives you the best response. 2/ you upload one image and it runs it through nano banana, flux, and bytedance then returns the most realistic version. 3/ you generate a video and genspark routes it through wan 2.2, pika labs, gen-2, and runway ml combining the best clips and motion. 4/ you paste an idea for slides and it uses gpt-5, claude 3 opus, gemini advanced, and genspark’s slide model to create a full deck with visuals and market data. 5/ AI slides and building a fundraising deck with a few prompts (wasn't perfect, but it was a good start for my personal workflow) what's cool about genspark is it's inexpensive and does it all. kinda like the costco of LLMs. TLDR; you hand it one prompt + input, and genspark routes sub-tasks across models and tools behind the scenes. it’s orchestration over one ideal model. perspective shift. instead of jumping from model to model, llm to llm, genspark shows what the next generation of ai will look like orchestration over obsession. and for the rest of us using it, the only thing that matters is getting the best outcome. what other AI app should I review next? i'll keep doing it if it's helpful for folks.

GREG ISENBERG

61,822 просмотров • 9 месяцев назад

This custom GPT writes Sora 2 prompts that actually work 🤯 Most people prompting Sora 2 are getting garbage output. The problem isn't Sora 2. It's your prompts. Here's what I figured out after generating 100+ UGC videos: Sora 2 needs hyper-specific instructions—shot composition, timing, camera movement, lighting, scene details. Most people prompt like this: "Make a UGC video of someone using my skincare product" And then wonder why the output looks like AI slop. What actually gets results: "9:16 vertical handheld shot, soft natural window lighting, female creator mid-20s in bathroom mirror holding [brand] serum at eye level, genuine testimonial energy, 0-3s hook showing application and subtle reaction, shallow depth of field, clean warm color grade" That's the gap between unusable content and scroll-stopping UGC. So I built a custom GPT that writes these prompts for you: → Tell it your product (supplement, skincare, SaaS, whatever) → It generates a detailed timeline-based Sora 2 prompt → Shot composition, camera angles, timing breakdowns—all included → Optimized for 9:16 TikTok/Reels format → Copy and paste directly into Sora 2 No guessing. No prompt engineering rabbit holes. No "why does this look terrible" frustration. Professional UGC prompts in 10 seconds. Want free access to the Sora 2 Prompt Generator GPT? > Comment "SORA" > Like this post And I'll send it over (must be following so I can DM)
0:23

This custom GPT writes Sora 2 prompts that actually work 🤯 Most people prompting Sora 2 are getting garbage output. The problem isn't Sora 2. It's your prompts. Here's what I figured out after generating 100+ UGC videos: Sora 2 needs hyper-specific instructions—shot composition, timing, camera movement, lighting, scene details. Most people prompt like this: "Make a UGC video of someone using my skincare product" And then wonder why the output looks like AI slop. What actually gets results: "9:16 vertical handheld shot, soft natural window lighting, female creator mid-20s in bathroom mirror holding [brand] serum at eye level, genuine testimonial energy, 0-3s hook showing application and subtle reaction, shallow depth of field, clean warm color grade" That's the gap between unusable content and scroll-stopping UGC. So I built a custom GPT that writes these prompts for you: → Tell it your product (supplement, skincare, SaaS, whatever) → It generates a detailed timeline-based Sora 2 prompt → Shot composition, camera angles, timing breakdowns—all included → Optimized for 9:16 TikTok/Reels format → Copy and paste directly into Sora 2 No guessing. No prompt engineering rabbit holes. No "why does this look terrible" frustration. Professional UGC prompts in 10 seconds. Want free access to the Sora 2 Prompt Generator GPT? > Comment "SORA" > Like this post And I'll send it over (must be following so I can DM)

Mike Futia

13,262 просмотров • 6 месяцев назад

We're giving away the prompt we used to make this AI UGC video. Getting Sora 2 to output hyper-realistic, consistent footage isn't just about the tool. The prompt has to be structured in a very specific way... character description, cinematography, camera motion, lighting, dialogue, audio, authenticity keywords. Get it wrong and the footage looks off. Get it right and everything downstream becomes easier. We’re sharing the exact Claude template we use as step one of our AI UGC workflow. It takes a basic brief (who the subject is, where they are, what they're saying, what device it should look like, their accent) and structures all of that into the exact format Sora 2 needs automatically. It sits at the start of a full workflow that runs through Sora 2 for the hook, ElevenLabs for voice cloning, ChatGPT for subject consistency, Nano Banana for B-roll generation and Kling for animation. The whole thing is broken down in the latest D2C Diaries episode. But this prompt is the foundation. Without it, the rest of the workflow is harder than it needs to be. If you're building in this space or testing AI UGC for your brand, this is worth having. Retweet this post and comment PROMPT and I’ll send it over.
0:41

We're giving away the prompt we used to make this AI UGC video. Getting Sora 2 to output hyper-realistic, consistent footage isn't just about the tool. The prompt has to be structured in a very specific way... character description, cinematography, camera motion, lighting, dialogue, audio, authenticity keywords. Get it wrong and the footage looks off. Get it right and everything downstream becomes easier. We’re sharing the exact Claude template we use as step one of our AI UGC workflow. It takes a basic brief (who the subject is, where they are, what they're saying, what device it should look like, their accent) and structures all of that into the exact format Sora 2 needs automatically. It sits at the start of a full workflow that runs through Sora 2 for the hook, ElevenLabs for voice cloning, ChatGPT for subject consistency, Nano Banana for B-roll generation and Kling for animation. The whole thing is broken down in the latest D2C Diaries episode. But this prompt is the foundation. Without it, the rest of the workflow is harder than it needs to be. If you're building in this space or testing AI UGC for your brand, this is worth having. Retweet this post and comment PROMPT and I’ll send it over.

Joe Marston | Ecom Growth

17,422 просмотров • 4 месяцев назад

This is the best scene in Hell Grind, an entirely AI-made movie, the flashback. Watch it and read this analysis on where we are with AI movies today: time, cost, quality. Overall: Phenomenal technical demo by Higgsfield. Mediocre movie. Good graphics, hints of emotion, but superhero movie level quality in certain scenes at best. Too many cuts. That said, 660x fewer man hours, 50x faster and 36x cheaper than the median US film. Time: The 95 min film took 15 people 14 days. The median US theatrical production takes ~200 people ~2yrs. That’s a 660x improvement in man-hours and 50x in calendar time. Economics: It took $500k, 80% of which was compute. The final footage was cut from ~100hrs of footage generated from text to video / image to video models like Bytedance’s Seedance: a 64:1 “curation” ratio. The median US movie takes ~$18M, with even indie films costing $1-5M. Thats 36x cheaper than median. Quality: Average watch *at best*. Way too many cuts between shots, several characters change accents and have “AI” synthetic voices and characters feel like it’s AI too. Movement, editing and blocking feel artificial too. On the plus side, we’ve more or less solved character consistency, camera angles and realism. The reason the movie wasn’t amazing was more about poor directorial choices than innately unusable video models. Hard to put a number on it but maybe we’re at ~90% on quality that is technically achievable. If Scorsese made an AI movie, I reckon it would be quite good. I know the visceral reaction to anything AI is real and well-studied. But I think it’s folly to fight the inevitability of AI film. It’s too cheap and quick to ignore and almost there on quality. Creators with distribution *will* make AI films and shows and just put them on YouTube. This is the worst quality, slowest and most expensive it will ever be. In the end, good content beats “real” content.
2:07

This is the best scene in Hell Grind, an entirely AI-made movie, the flashback. Watch it and read this analysis on where we are with AI movies today: time, cost, quality. Overall: Phenomenal technical demo by Higgsfield. Mediocre movie. Good graphics, hints of emotion, but superhero movie level quality in certain scenes at best. Too many cuts. That said, 660x fewer man hours, 50x faster and 36x cheaper than the median US film. Time: The 95 min film took 15 people 14 days. The median US theatrical production takes ~200 people ~2yrs. That’s a 660x improvement in man-hours and 50x in calendar time. Economics: It took $500k, 80% of which was compute. The final footage was cut from ~100hrs of footage generated from text to video / image to video models like Bytedance’s Seedance: a 64:1 “curation” ratio. The median US movie takes ~$18M, with even indie films costing $1-5M. Thats 36x cheaper than median. Quality: Average watch *at best*. Way too many cuts between shots, several characters change accents and have “AI” synthetic voices and characters feel like it’s AI too. Movement, editing and blocking feel artificial too. On the plus side, we’ve more or less solved character consistency, camera angles and realism. The reason the movie wasn’t amazing was more about poor directorial choices than innately unusable video models. Hard to put a number on it but maybe we’re at ~90% on quality that is technically achievable. If Scorsese made an AI movie, I reckon it would be quite good. I know the visceral reaction to anything AI is real and well-studied. But I think it’s folly to fight the inevitability of AI film. It’s too cheap and quick to ignore and almost there on quality. Creators with distribution *will* make AI films and shows and just put them on YouTube. This is the worst quality, slowest and most expensive it will ever be. In the end, good content beats “real” content.

Deedy

34,584 просмотров • 1 месяц назад