Google is making progress on their diffusion models... It's... now as good as a Gemini 2.0 Flash Lite. The writing is on the wall, a majority of language AI use in the future will be diffusion models.show more

The latent space of earlier generative models like GANS... can linearly encode concepts of the data. What if the data was model weights? We present weights2weights, a subspace in diffusion weights that behaves as an interpretable latent space over customized diffusion models.show more

Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models... "Block diffusion sequentially generates blocks of tokens by performing diffusion within each block and conditioning on previous blocks. By combining strength from autoregressive and diffusion models, block diffusion overcomes the limitations of both approaches by supporting variable-length, higher-quality generation and improving inference efficiency with KV caching and parallel sampling."show more

LLaDA (the first Large Language Diffusion Model) is *just*... out 💥 and I've built a demo, try out now 👨‍💻 It's mesmerizing to watch the diffusion process 🌀, and it being a diffusion model gives you superpowers like "the 4th word has to be pineapple" 🦸 Demo and weights 👇show more

Nano Banana 2 Lite 🤝 Gemini Omni Flash Now... you can use the Interactions API to build with both models. We created this e-commerce applet in Google AI Studio that turns static product shots into cinematic videos. Here’s how it works: 1⃣ Upload a product photo and set the vibe 2⃣ Nano Banana 2 Lite generates new, brand-accurate assets 3⃣ Gemini Omni Flash renders those images into a high-quality videoshow more

(1/2) Check out "𝐏𝐨𝐥𝐲𝐃𝐢𝐟𝐟: Generating 3D Polygonal Meshes with... Diffusion Models"! Our model operates directly on the polygons of 3D meshes and generates novel shapes as output through an iterative diffusion process.show more

1/ Gemini 2.5 is here, and it’s our most... intelligent AI model ever. Our first 2.5 model, Gemini 2.5 Pro Experimental is a state-of-the-art thinking model, leading in a wide range of benchmarks – with impressive improvements in enhanced reasoning and coding and now #1 on Arena by a significant margin. With a model this intelligent, we wanted to get it to people as quickly as possible. Find it on Google AI Studio and in the Google Gemini for Gemini Advanced users now – and in Vertex in the coming weeks. This is the start of a new era of thinking models – and we can’t wait to see where things go from here.show more

Selected as a best paper finalist at #CVPR2026: PixelDiT... from NVIDIA Research In most image generation models, a pretrained autoencoder compresses the image before any diffusion happens, causing quality loss that accumulates across the entire pipeline. PixelDiT, or Pixel Diffusion Transformers, removes this step entirely. It's a single-stage model that learns the diffusion process directly in pixel space, end-to-end.show more

The Hidden Language of Diffusion Models paper page: tackle... the challenge of understanding concept representations in text-to-image models by decomposing an input text prompt into a small set of interpretable elements. This is achieved by learning a pseudo-token that is a sparse weighted combination of tokens from the model's vocabulary, with the objective of reconstructing the images generated for the given concept. Applied over the state-of-the-art Stable Diffusion model, this decomposition reveals non-trivial and surprising structures in the representations of concepts. For example, we find that some concepts such as "a president" or "a composer" are dominated by specific instances (e.g., "Obama", "Biden") and their interpolations. Other concepts, such as "happiness" combine associated terms that can be concrete ("family", "laughter") or abstract ("friendship", "emotion"). In addition to peering into the inner workings of Stable Diffusion, our method also enables applications such as single-image decomposition to tokens, bias detection and mitigation, and semantic image manipulationshow more

A curious property of diffusion models: when a sample... is mostly noise, the model’s best guess is close to the mean. So early on samples can be pulled toward the center of the dataset, even when no data exists there, before rebounding toward the data manifold.show more

China is making AI a compulsory subject in primary... school. If India wants to lead the future, AI literacy must become as fundamental as reading, writing and arithmetic. That’s what our new Education Minister should be working on.show more

Drawing on our unique, global view of AI usage,... the Microsoft AI for Good Lab’s latest AI Diffusion report highlights that AI is moving from trial to daily utility. Growth is surging in markets like Japan and South Korea as local-language capabilities improve, while disparities between the Global North and South continue to deepen.show more

Woow Google has just rolled out the AI model... Flash Thinking 2.0 This is the first reasoning model capable of accessing YouTube and it changes everything: - Search for a video on your topic - Ask Gemini to think about the video - You'll have a tailor-made result in 10 sec. And it's even connected to Google Search and Maps!show more

Most recent diffusion language model research (that I’ve seen)... seems to be using masking as the noising process. It looks like, however, most closed-source models (Google Gemini Diffusion and possibly Inception Labs’ Mercury) use a different noising process, where instead of masking tokens, they replace them with different tokens (either with a random token or a semantically similar token). I wondered how they were getting such high throughput with the latter noising process, since I believed that optimizing inference with KVCache approximation would be more difficult (for various reasons). I visualized this noising process with tiny-diffusion and compared it to normal unmasking, and was very surprised to see how fast the generation “settles” into a reasonable output, and then only slightly refines afterwards, requiring much fewer steps in total. Unmasking (where tokens are never remasked, the typical implementation) is inherently limited in generation speed by the fact that an increase in tokens decoded per step leads to more errors due to the mismatch between individual and marginal token probability distributions we sample from. The token replacement noising process seems to have a much different set of characteristics. Because we sample each token per step, every token makes “progress” towards the final output each iteration (in addition to *potentially* giving other tokens more information in future steps). Generally, masking has outperformed other noising processes, which is probably why most research focused on it (using smaller models). But the paper referred to in the retweet shows that random replacement as a noising process may scale better as model size increases. Big labs might have noticed these results much earlier (due to having drastically more training resources and being able to test larger models), which may explain the discrepancy in the choice of noising process. I’m gonna test this with larger models, since tiny-diffusion only has 10M parameters.show more

BREAKING 🚨: Google is testing Voice Cloning on AI... Studio, potentially ahead of the anticipated Gemini 3 Flash Native Audio upgrade. Users will be able to record and upload their own voices to let Gemini clone them. Soon? 👀show more

"The vast majority of the team right now is... focused on Empulse" 1047 games CEO reveals that a small number of people are still working on Splitgate, but their full focus for the future will be on their movement shooter Empulseshow more

Google is preparing to release an updated version of... Omni Flash (models/gemini-omni-1.1-flash) The underlying model id is "omni-1p2", indicating that Google is using the Omni Flash 1.2 checkpoint for 1.1 The attached video is an output from it, generated in 43 seconds. I wish they wouldn't release it at all atp...show more

As someone who is pretty up to date in... AI I’d say there are 2 times a year where I am genuinely blown away by something. A breakthrough that keeps me up at night. This is one of them. I mean the video of will smith genuinely freaks me out by how good it is. 2.5 years of AI progress and you think it’s hitting a wall. There’s no point in trying to convince people where this is going. The writing is on the wall. I can bring a horse to water but I can’t force him to drink.show more

Naming a startup is hard. Usually it's chaos: throw... words on a whiteboard, argue, and as fatigue hits, settle on good enough. Untitled offers another way. I turned the method I used to name dozens of startups into an app. It's free to try and uses the latest AI models.show more

Text-to-image diffusion transformer models learn to align text and... image representations as a byproduct of their conditional denoising task. By taking the dot product between the text and image representations of a DiT model (like Flux 2), you can create rich saliency maps.show more