Black Forest Labs just dropped FLUX 3, a single model that generates video with native audio, images, and even robot actions from one set of weights.
Meanwhile NVIDIA brought AI agents to every major creative app at SIGGRAPH, Disney opened an AI ad studio for small businesses, and Elon Musk says Grok will produce a full AI feature film by December.
(Read time - 5 minutes)
We’ve put together a full Claude Code masterclass for beginners.
No slides. No theory. Just us opening Claude Code live and showing you exactly how to use it from scratch.
No coding background needed. Over 2,000 people have already signed up.
If you’ve been curious about Claude Code but weren’t sure where to begin, this is the perfect place to start.
Black Forest Labs, the team behind the original Stable Diffusion architecture, launched FLUX 3.
It is a single multimodal model trained simultaneously on images, video, and audio, generating up to 20-second video clips with synchronized sound, dialogue, and ambient noise from one architecture.
🔍 Highlights:
🎬 Up to 20-second video with native audio including dialogue, sound effects, and ambient noise, all generated in sync from a single prompt
📊 93% win rate over Luma Ray 3.2 and 77% over Runway Gen-4.5 in human evaluations, with closer results against Seedance 2.0 and Gemini Omni Flash
🔧 Text-to-video, image-to-video, and video-to-video workflows plus keyframe-controlled transitions and multilingual dialogue generation
🤖 Same weights power FLUX-mimic for robotics, already running on Audi production lines, proving the architecture works beyond creative tasks
📸 HeadshotPro: Studio-style AI headshots for LinkedIn, bios, and creator profiles.
🌐 Framer 3.0: AI Agents now live on the Framer canvas, editing live pages, components, styles, and CMS content in-place.
📱 Ocoya: AI-powered social media management tool that generates captions, schedules posts, and automates content across 30+ platforms.
📷 Topaz Photo AI: Next-gen release with NeuroStream tech that cuts VRAM requirements by 95%.
✨ NVIDIA brought AI agents directly into Adobe Creative Cloud, Blender, Unreal Engine, and Houdini at SIGGRAPH 2026. (continue reading)
🎬 Disney opened a closed beta of Disney Ad Creative Studio, that turns brand logos, product shots, and guidelines into connected-TV video ads with scripts, video, and music. (continue reading)
📷 Elon Musk announced Grok Imagine will produce a feature-length AI adaptation of Homer’s Odyssey by year-end. (continue reading)
🔍 NVIDIA launched a Synthetic Video Detector that spots deepfakes in 22 milliseconds per frame with up to 92% accuracy on uncompressed video. (continue reading)
New prompt for editorial magazine covers. Credit to @Sairah_0
Prompt: Create a premium editorial magazine cover featuring [SUBJECT] as the central figure, filling 60–70% of the composition in a cinematic three-quarter pose.
Add 5–8 asymmetrical collage panels with related locations, action shots, objects, textures, or historical details. Use overlapping paper-cut styling.
Include a bold condensed masthead, editorial captions, dates, coordinates, labels, stamps, barcode, and multilingual typography where relevant.
Style: luxury magazine meets vintage screen-printed poster. Use aged ivory, black, burnt orange, charcoal, sepia, faded ink, halftone texture, worn paper, and cinematic color grading.
Photorealistic 85mm editorial photography, soft lighting, rich shadows, crisp detail, Swiss grid layout, balanced negative space, 8K, print-ready, no warped text or AI artifacts.
Our sister newsletter The AI Break delivers actionable AI tutorials twice a week. Here are some recent favorites:
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.