Hackernews posts about Imagen
Imagen is an artificial intelligence (AI) model that uses text-to-image diffusion to generate photorealistic images from input text descriptions.
- What I learned competing against a convnet (Karpathy 2014) (karpathy.github.io)
- Want your images back? That'll be $5 (www.lutr.dev)
- Krea 2: SOTA open-weights 12B image model (www.krea.ai)
- Moebius: 0.2B image inpainting model with 10B-level performance (hustvl.github.io)
- Image Compression (www.makingsoftware.com)
- Un-0: Generating Images with Coupled Oscillators (unconv.ai)
- Dissecting Apple's Sparse Image Format (ASIF) (schamper.dev)
- Minimus container images are now free (images.minimus.io)
- Rust Coreutils cp Ended Up Breaking Ubuntu Image Builds with Incompatibility (www.phoronix.com)
- Better Images of AI (betterimagesofai.org)
- HTML is a native image format, hear me out (hmml.eddocu.com)
- Image Toolbox (T8RIN) (github.com)
- Generating Images with a 2025 Android (www.duration.ai)
- Decoding 137MHz signals from passing satellites to get images with a $10 SDR (www.jonhilty.com)
- Enhance RAW image processing with Core Image [video] (developer.apple.com)
- UK PM gives tech firms ultimatum to block explicit images on children's phones (www.theguardian.com)
- Signal says UK plan to scan devices for nude images 'endangers us all' (www.theregister.com)
- Show HN: Turn images into audio that can be decoded with a spectrogram (nsspot.herokuapp.com)
- Neural Image Compression with Gemini 3 (bertolami.com)