The Yoneda lemma is uniquely determined by its relationship to every other result in mathematics!

I am (slowly) re-reading the Tolkien legendarium (of which Lord of the Rings is a small part). The whole body of work is so incredible and there's nothing else like it... it dilutes other worlds of fiction. Wait - your story doesn't have a comprehensive history/mythology spanning

Our latest Gemini image generation and editing model is quite good (see the examples in the thread below). It lets you indulge in a bit of creative fun, helps you make new business cards, etc! Try it out at gemini.google.com Google Gemini App

Vision tokenizers are stuck in 2020🤔while language models revolutionized AI🚀 Language: One tokenizer for everything Vision: Fragmented across modalities & tasks Introducing AToken: The first unified visual tokenizer for images, videos & 3D that does BOTH reconstruction AND

Happy to share (with a bit of delay tho) our paper on quantifying visual information loss in VLMs --- "Lost in Embeddings: Information Loss in Vision-Language Models" is accepted to EMNLP 2025 findings: arxiv.org/pdf/2509.11986 💃code is also released: github.com/lyan62/vlm-inf…

FLUX is now in Photoshop. A dream coming true. 🥹 bfl.ai/blog/flux1-kon…

🚀Excited to share our recent work on test-time scaling for feed-forward Gaussian splatting: we learn a recurrent model ReSplat that is able to iteratively improve the reconstruction quality in a feed-forward manner! haofeixu.github.io/resplat/

「残酷な天使のテーゼ」が発売されてから、今日で30年。 相変わらずエヴァ観てない。ごめん。

Super cool paper from a few days ago :) I had the intuition this was true for a while now, and we even proved injectivity holds at initialisation in a recent paper (arxiv.org/pdf/2507.08802), but this paper really improved on our result, showing injectivity holds across training!

We’ve been thinking a lot about how generative AI can make editing feel faster, smarter, and more intuitive without losing the control creators love. Today at #AdobeMAX, the #AdobeFirefly team previewed Layered Image Editing: bringing the power of layers and compositing together

Excited to announce our MIT Press book “Neuroevolution: Harnessing Creativity in AI Agent Design” by Sebastian Risi (Sebastian Risi), Yujin Tang (Yujin Tang), Risto Miikkulainen, and myself. We explore decades of work on evolving intelligent agents and shows how neuroevolution can

Can we apply gradient descent to discrete changes? In our new #SIGGRAPHAsia paper, we show that gradient descent can work on shape grammars, as in CAD and procedural modeling, but only if the grammars are designed correctly!

As you may have heard, “The Thing” has been added to the Library of Congress. We made the movie to push against the edges of what we could pull off, and it carries that strain in its bones. Seeing it now treated with the same care and reverence we had making it means a great deal