🚨 WIRED reports that OpenAI is preparing to launch a standalone app for Sora 2, its video-generation model. Key details: The app looks like TikTok, but every video is AI-generated. Vertical feed, swipe navigation, “For You” recommendations. Users can create up to 10s clips, remix others, like/comment/share. No gallery uploads — everything must be generated inside. Identity verification: confirm your face → others can generate clips with you in them (you get notified every time it’s used). Internal OpenAI testing last week got very positive feedback. Limitation: frequent refusals due to copyright filters. Why it matters: TikTok’s uncertain future (Oracle deal in motion) leaves a gap. OpenAI has a rare chance to claim the short-video market and onboard millions into an AI-native feed. 👁️‍🗨️ Fun fact: I actually launched something similar long before Sora 2. In my project AI Dreamin, I publish daily “machine dreams” — surreal AI-generated clips posted automatically, like fragments from a synthetic subconscious. It’s already running at scale, proving that an endless AI-only content feed is not just theory. JOIN 🌀 https://t.me/aidreamin

What if Cossacks walked the Ralph Lauren runway? 🇺🇦⚔️ Last week I made this short AI fashion campaign video: Cossacks x Polo Ralph Lauren. It looks like a luxury editorial, but it’s fully AI-generated. 👉 Why I did this: - To show how AI can mix cultural heritage + modern fashion. - To test workflows for creating viral short-form videos. - To push the idea of Fashion Сi4 — an AI influencer collective built around Ukrainian Cossacks. 👉 How I made it (step by step): 1. Prompts - Chat GPT. 2. Image Generation — I created hyper-realistic stills in Gemini with vertical 9:16 ratio. 3. Upscaling & Cleanup — used Runway + Magnific to polish skin, clothes, and details. 4. Video Animation — imported stills into Veo3 + Kling for motion: breathing, wind in shirts, horse movement. 5. Editing — stitched the 8-second cuts in CapCut with subtle transitions and warm color grading. 6. Publishing — posted as Reel/TikTok with caption: “What if Cossacks became Ralph Lauren models?” 👉 Why this matters: - One video = multiple content pieces: Reel, Stories, BTS carousel, newsletter case study. - AI videos like this can be a template for brands to tell stories fast. - This project = creative + cultural + viral. If you want me to share the exact prompts + workflow template, comment Cossack below, and I’ll drop the full guide.

Veo 3 just dropped a game-changer. For the first time, you can generate vertical videos (9:16) directly — built for TikTok, Reels, and Shorts. No more hacks or awkward cropping. This is the clearest sign yet that Google is positioning Veo as the backbone of short-form AI video. And here’s where it gets even more interesting — the new pricing model: Veo 3 (high quality): $0.40 per second (down from $0.75) Veo 3 Fast (lower quality, quicker): $0.15 per second (down from $0.40) Do the math: A 30-second high-quality clip now costs $12 instead of $22.50. A 30-second fast clip costs just $4.50. That’s not just cheaper — it makes scaled production of vertical AI video actually viable. Expect your feeds on TikTok, Reels, and Shorts to be flooded very soon. Question is: 👉 Are we witnessing the start of high-quality AI vertical content going mainstream, or just the acceleration of “AI slop” everywhere? More AI tools, hacks, and monetization strategies here: Den Manu 🇺🇦