Image Generation

Today's Stories

1. 1

   Black Forest Labs launches FLUX 3 for image, video, and audio generation

   Black Forest Labs released FLUX 3, a multimodal AI model that generates images and combined audio/video clips up to 20 seconds from a single prompt. The model is jointly trained across image, video, and audio rather than combining separate models. It also extends to robotic vision and actions. FLUX 3 represents the company's first public video generation model and frames creative generation, simulation, computer use, and robotics as connected applications of a single capability. Black Forest Labs positions this as "visual intelligence"—models that can perceive, predict, and act across physical and digital environments—rather than treating each modality as a separate tool.

   FLUX 3 will be offered through four product lines: FLUX 3 Video, FLUX 3 Image, and FLUX 3 Act. The release is limited to start, meaning broader availability or pricing details may follow.

2. 2

   Blomkamp's AI 'zombie film' Nightborne panned as slop despite human input

   Director Neill Blomkamp unveiled Nightborne, a 13-minute sci-fi short created entirely with ByteDance's Seedance 2.0 text-to-video generator through his new company Barley Studios. The film, loosely based on Peter Watts' 2014 novel Echopraxia, features 32 human actors whose likenesses were used and dialogue from real concept artists, but was generated shot-by-shot using the AI tool. Despite polished editing by Austyn Daines and Blomkamp's stated ambition to make a full feature in this format, the short bears the hallmarks of AI-generated content ('slop') — gibberish background signs, unnatural character delivery, and a loss of creative direction that critics attributed to the filmmaker being 'washed.' The project highlights that even when skilled filmmakers apply traditional technique, text-to-video models still lack the understanding of human performance nuance needed to produce compelling finished work.

   Blomkamp framed Nightborne as a 'test start' to demonstrate what generative AI is capable of and said he wants to tackle a full feature in this format at some point in the future. The announcement was met with immediate backlash, positioning Blomkamp in a situation similar to director Darren Aronofsky's failed AI Revolutionary War series.

3. 3

   Disney's Ozzy Fox goes viral amid 'AI slop' claims—but it's AI-assisted, not AI-generated

   Disney's new preschool cartoon series Ozzy Fox, launched on YouTube with minimal promotion, has drawn widespread criticism over suspected AI generation due to visual glitches like incomplete props and bad lip-synching. The backlash has paradoxically boosted viewership; each of the two episodes released so far has over 500,000 views. The series is a co-production between Disney and French animation company Animaj, created through Disney's 2024 Accelerator program. Ozzy Fox uses AI as an assistant tool—specifically for sketch-to-pose prediction and motion in-betweening—but the 3D models, storyboards, and key poses are created by human artists who then manually review and correct the work in Maya animation software. The distinction matters: if every imperfect animation gets labeled 'AI slop' when it is not actually AI-generated, the label may lose its power to signal genuine concern when Disney or others deploy actual generative AI cartoons like text-to-video tools.

   Disney has explored generative AI for animation, including a major deal with OpenAI that collapsed when the company deprioritized its Sora AI video platform. How the public and critics respond to Ozzy Fox's actual workflow—and whether they distinguish AI-assisted tools from AI-generated content—will shape perceptions of animation AI going forward.

4. 4

   Alibaba's Qwen-Image-3.0 generates readable text and complex layouts in one pass

   Alibaba's Qwen team released Qwen-Image-3.0, an image generator that accepts prompts up to 4,500 tokens, renders legible text as small as ten pixels, and supports twelve languages natively. It can create complex layouts such as infographics, LaTeX papers, and newspaper pages in a single pass. The ability to generate intricate multi-element layouts with readable text in one step addresses a longstanding weakness in image generation—most systems struggle with text legibility and coherent complex designs. For businesses and creators working with infographics, academic papers, or print layouts, this could reduce the need for manual design work.

   The practical utility remains uncertain because the output is a pixel image rather than an editable format, limiting downstream editing and repurposing compared to native design files.

5. 5

   Cloak: Free Tool Adds Invisible Noise to Block AI Training

   Cloak, a free web tool created by Timothy via Claude Code, layers invisible noise onto images to make AI models classify them as watermarks or harmful content and refuse to train on them. The tool uses a technique called L-infinity perturbations and projected gradient ascent to maximize cosine similarity scores between a CLIP embedding and a target embedding, steering models away from the protected image. Artists face ongoing threats from automated image scrapers that feed training datasets for AI models. Cloak addresses a gap that other protection tools do not fill: it works against mass-scraping pipelines without assuming the image will be used for training (unlike Glaze) or requiring large-scale dataset representation (unlike Nightshade). The protection remains effective even if the image is cropped or resized, offering a practical layer of defense without visible artifacts at default strength.

   Cloak is free with limits: maximum file size 30 megabytes, maximum image length 10,000 pixels, 300 processing passes maximum, and 60 images per hour. Best target prompts include 'watermark', 'blood', 'blurry', 'bad quality', and 'explicit content'—labels scrapers routinely filter out. The creator recommends using Cloak alongside Glaze and Nightshade for maximum protection.