Xavi Giró

X and @elonmusk have failed into promoting the values of democracy and human rights. Time to leave this platform. We learned a lot here, thanks to those who made it possible. Find me on LinkedIn and Bluesky.

@KenRoth The biggest risk of AI is the concentration of power in a few dominant providers of proprietary AI assistants. The only solution to AI sovereignty is open source foundation models.

We're looking forward to connecting with the machine learning community at @icmlconf! We have talks happening at our booth all week, plus a chance to connect 1:1 with our researchers. Full schedule: amzn.to/3SBXPXA

When frontier labs suddenly cut costs, or say "we found a way to dramatically cut inference memory!!" This is what they found. The secret sauce. Vision always wins! Of course, my very capable ex-colleagues who now work on Gemini and Claude already found out years ago:

Amazon researchers have accepted papers at @icmlconf spanning machine learning, causal reasoning, LLM inference, agentic systems, vision-language models, graph learning, robotics, and more. Explore the full list. #ICML2026 amazon.science/conferences-an…

The video of my Stanford CS25 guest lecture, From Language Models to Native Multimodal Intelligence, is now online. I discussed how the core ideas behind LLMs has shaped multimodal AI, from architectures to training paradigms and scaling, and where the next challenges may lie. 🧠🌐 🎥: youtube.com/watch?v=NDdc39…

Amazon's latest machine learning research is headed to Seoul. We'll be at ICML with accepted papers, live demos, and researchers presenting across agentic AI, robotics, and more: amzn.to/4vHY5CJ #ICML2026

GANs are back! "Scalable GANs with Transformers" shows that GANs are finicky but with the right setup, they can be scaled. arxiv.org/abs/2509.24935

We explored the impact of variability sources in generative modeling. Turns out, we've been neglecting the error bars associated with training variability all along! We should aim to report results that we are sure of their scientific validity, instead of seed engineering!

We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2. Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication. We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating. 🧵👇

🚀 We're hiring a Research Scientist, Machine Learning Research (MLR), Barcelona ☀️ 👉 Apply here: jobs.apple.com/en-us/details/… #MachineLearning #MLResearch #AI #Hiring #Apple #Barcelona #ResearchScientist #DeepLearning

Who wants free coffee?! Join the Ioannidis lab.

On 1 July: 'Multimodal Foundation Models – from Research to Innovation", joint workshop by ELLIOT, @elias_project, @ELLISBarcelona & the ELLIS Program on Multimodal Learning Systems. 📍 Barcelona 🇪🇸 + online 🔗 More info: elliot-ai.eu/events/multimo…

Just caught up with the recent GLM-5.2 release. The best open-weight model today. Architecture-wise, it's build on the GLM-5 and GLM-5.1 architecture that I covered previously, which means it's reusing the Multi-head Latent Attention (MLA) and DeepSeek Sparse Attention (DSA) mechanisms from DeepSeek V3.2. (I wrote about it here: magazine.sebastianraschka.com/p/technical-de…) What's new is that they added an IndexShare mechanism. (That's a cross-layer reuse trick for DSA where instead of recomputing the sparse-attention top-k indexer in every layer, GLM-5.2 runs the full indexer only once every four layers and lets the following layers reuse those selected token indices. This keeps the same DSA idea but makes 1M-token inference much cheaper.)

@ClementDelangue @Dan_Jeffries1 Everyone, please join Project Tapestry thealliance.ai/projects/tapes…

Local models can't benefit from batch parallelism as easily, but you can still parallelise over the token axis. So here's an open text diffusion model! >1000 tokens/sec for accelerated tokenmaxxing, yay!🫨

Poster is up #241! Join us at Session 4 from 4:30pm (June 6) @CVPR #CVPR2026 for our @GoogleDeepMind paper "Unique Lives, Shared World"... Learning from Single-Life Videos. We introduce a new learning paradigm - only from the experiences of ONE person @SaynaEbrahimi @TengdaHan

I followed up on two misconduct cases at top ML conferences. TLDR; academic dishonesty pays 😓 Bans (especially cross-venue bans) are non-existent and hard to enforce [1/3]

@dimadamen @CSProfKGD @ETH_en @mapo1 @xiwang1212 @leto__jean @JitendraMalikCV @pulkitology Hiking wearing the badge is next-level. Greetings !

I had an exceptionally enjoyable &thought provoking week @ETH_en invited by @mapo1 @xiwang1212 @leto__jean.. presenting our latest works on "Modelling in an Ego-Sensed World" [tagged from pics @JitendraMalikCV @CSProfKGD @pulkitology & others - Andrea, Laura, Javier, Aylosha] 1/N

📢 New @heyjasper release ! 📢 MONET 🌸 : An Apache2.0 deduped and recaptioned dataset of 105M samples unlocking reproducible text-to-image research. Nano T2I 🖌️ : A codebase to train your own T2I model 🤗 @huggingface: huggingface.co/datasets/jaspe… 💻: github.com/gojasper/nano-… Very excited about this new release, pushing the boundaries of open and reproducible T2I research. Congrats to the team! Benjamin Aubin Gonzalo Quintana @onurxtasar @UlaLaParis @_jeev2 @dh7net @clipdropapp @heyjasperai