Abhishek Prabhu’s Post

The second video in this series is a more abstract take, inspired by Elton John’s concept for Bennie and the Jets. This one was tricky, especially when it came to keeping the portals and characters consistent. I’ve since revisited it with KLING 1.6 and Act-One to refine the details, but I’m holding off on a full update until I get access to Veo2 to truly push the "World Builder" potential. What’s wild is that, the same week I wrapped this project, it felt almost instantly obsolete. Filmmaking is evolving at a speed that feels tied to Moore’s Law, and staying ahead—or even just keeping up—requires constant rethinking. The mixed emotions that come with this are hard to ignore, but it’s also a reminder of why experimenting and playing with these tools matters so much. More behind-the-scenes and workflow insights are on the way, but for now, let’s take a step back, have fun, and maybe sing a little song. _____ Recently, I had the opportunity to produce a live show featuring an LED wall, which allowed me to push the boundaries of what’s possible with publicly available AI video software. After thousands of iterations across 18 videos, each 2-3 minutes long, I’m proud to say I’ve come out the other side with results I’m excited to share. As the project begins to roll out, I’ll reveal more, but with tools like Google Veo2 Pika 2.0 and new advancements emerging every week, it’s clear how important it is to experiment and refine the creative potential of these technologies. For most of these videos (including the one below), I used Midjourney to generate base images, animated them with KLING - AI Videos 1.5 and Runway, and brought everything together in Adobe Premiere. (As always, excuse the compression)

The most interesting part about AI video generation isn't the videos - it's how it's making us question what we value in content. When everything can be generated instantly, originality becomes less about technical execution and more about having something worth saying. The bottleneck is moving from 'can we make this?' to 'what's worth making?' Like early CGI movies, technical capability doesn't guarantee good content. Expect a lot of spectacular but empty visuals before people figure out how to use it well.

Here's a 3 minute look at a whole VCCP agency working hard on getting Ai to do the hardest thing of all - be consistent. A quick word to Ben Hopkins (or Morten Legarth) "The potential of generative video is much talked about and we’ve seen some great examples in fashion and music videos. They play well to the limitations of popular tools, but when it comes to creative control and consistency things get sketchy." Sketchy is a good word to use. A very polite way to put it. By the way, you wanna know how I'm working to get my character to have the same face in every shot? (because I've tried screaming at Midjourney "just stop" for a couple of weeks, and it didn't work) I'm generating scene stills with the main character, then animate them in Runway, and then... and this is beautiful – use DEEPFAKE to make the face consistent all through the video.

The new Magnific AI update is here—and it’s game-changing. Here’s why: You can now create Editorial Portraits in Mystic 2.5, producing visuals so realistic you’ll do a double take. But I didn’t stop there. 🎥 I had to see how they’d look in motion. The result? Mind-blowing. (Yes, it works beyond just editorial portraits.) This opens up endless creative possibilities—whether you’re crafting cinematic animations or exploring new artistic mediums. Curious to see these in action?

Testing #ad ideas in #Midjourney V6. A lot harder to achieve than a one off prompt. V6 demands rerolling, tweaking, remixing and blending. FYI. Just so you know. (If you hit follow it lights a little fire for me to share more of what I know... lol. If you have a question, I'll be happy to answer in comments.) Repost to a friend (or client, or boss) in need. (A D this could be useful for your network. Piotr Bombol , you too.) #advertisingai #aiphotography #aiart #creativedirector #artdirector #creative #ai #stablediffusion ##DalleE #comfyui

Imagine a Script - not just written, but visualized. No, it is not replacing the actors... but it is fast tracking concepts, and moving from a rough idea - to a visualized possibility in hours - not months. GenAI can AMP + RAMP the Creative Professional. Let me show you how. The full step-by-step breakdown for this HOW TO, is in my Substack. Start here: https://lnkd.in/edvRm4mX Creating a SCENE of Dialogue, AI Amped Part: (1) Script + Midjourney + Runway's Act-One + ElevenLabs + Human Element (2) The Script + the Acting + the Characters... (3) The Animation + The Voices + The Sound Bed (4) Putting it all together PS - Don't judge it for what it ISN'T... Imagine how it could be USEFUL to you as a Creative Professional. Want to read it for FREE? ::: Comment ::: 'WHAT ABOUT...' - followed by HOW you could use this technique in what you do. ...and get access to my Substack for a whole WEEK for free - so you can read this whole series (and the 278 OTHER educational posts in the archives.) Again.. don't tell me HOW IT IS NOT GREAT (I see the tell-tale AI signatures)... instead, tell me how it can be HUGE for you right now. INSPIRE ME! Brian Sykes theBrianSykes AI Lab #Midjourney #Runway #ActOne #Udio #AIVideo

Merry Christmas, everyone! 🎄 This Christmas, we wanted to give you something special. Introducing Jingle Bells – an AI-generated movie about Delhi under attack by the AQI Monster and Santa Claus stepping in to save the day. This project was an experiment to see how far we could push AI tools in storytelling and editing, all within 24 hours. But it wasn’t without its challenges: Lack of HD generations: Most image-to-video models struggle with clarity. Character consistency: It’s improving, but still miles away from reliable. Limited control: Tools like Runway need better time remapping and keyframe precision – the basics of good animation. Frustrating prompting: Even with experience, it often took 20+ generations to get one usable shot. Reviewing your work is key in editing, but waiting and tweaking repeatedly was exhausting. Still, the progress is undeniable. A year ago, this wasn’t even possible. Challenges aside, I’m optimistic about where this technology is heading!

Hey there fellow video creators! I've been seeing a lot of buzz around this whole AI thing and how it's changing the game for us video folks. At first, I'll admit, I was a little worried. I mean, how can a bunch of algorithms possibly replace the creativity and skill that us humans bring to the table, right? But the more I've played around with these AI video tools, the more I've realized they're not here to replace us - they're here to make our lives easier. 🤖 Think about it - all those super tedious, time-consuming tasks like color grading, visual effects, even editing itself? AI can handle a lot of that heavy lifting, freeing us up to focus on the real artistry. Instead of spending hours upon hours fiddling with keyframes and node-based compositing, we can pour that energy into crafting the perfect narrative, directing powerful performances, and injecting each frame with our unique creative vision. Sure, the machines are getting smarter every day, but they'll never be able to fully replicate the magic that comes from the human touch. At the end of the day, we're the ones with the storytelling chops, the emotional intelligence, and that special je ne sais quoi that breathes life into every production. So let's accept and learn from these AI tools, use them to our advantage, and take our craft to new heights.🎥✨

Hey Video Wizards! 🧙♂️ Ever wondered how your fave films get those jaw-dropping special effects? Well, it's a kind of magic called AI! This clever tech can spot and tag all the action in your videos, making it a cinch to separate scenes and add some serious sparkle to your edits with automated effects. 🎞️ It's no longer a huge editing puzzle, 'cause AI has got your back. Get ready to up your film game without sweating it out. How cool is that? Got questions or wanna share your AI-powered creations? Hit us up in the comments!