Feliz con el nacimiento de este hijo fruto de años de embarazo de periodismo de investigación. Me lo publica Debate y, si queréis conocer qué hay bajo la punta del iceberg de la IA que todos usamos, es un buen manual... de resistencia. Estos medios le han dado ya la bienvenida:🧵

📢 New Blog Alert! Read a recap of the DP Beyond Algorithms Workshop and learn more about how to be part of the community working on sociotechnical aspects of deploying DP! opendp.org/blog/reflectio…

Until now, pricing structure on rideshare apps has been opaque for both drivers and riders. 🚗 To help fix this, the The Workers' Algorithm Observatory and researchers from Princeton University created the FairFare app to crowdsource payment info from drivers. Now, a new law in Colorado mandates transparency.

📢📢New paper: AI as Normal Technology. I’ve worked on this w/ Sayash Kapoor for the last 2 years. The full version will be our next book! Utopian & dystopian AGI visions are quite similar. A far more important axis of disagreement—will AI be superintelligence or normal technology?

Excited to be presenting our spotlight paper on provable decompositions of predictive uncertainty today at #iclr2025! Come visit poster #501 from 3-5:30pm in Hall 3 + Hall 2B to learn more.

📢 Sabíeu que aquest any celebrem el cinquè aniversari del Cicle OEIAC? Doncs sí! És per això que aprofitem l'avinentesa per compatir alguns dels seminaris més vistos d'aquests 5 anys de transferència de coneixement sobre usos ètics i capacitats (reals) de la IA. Disfruteu-ho! 🧵

My PhD advisor, Moritz Hardt, has just released the first half of his new book, The Emerging Science of Machine Learning Benchmarks. It’s freely available and highly recommended: mlbenchmarks.org

🎓 ESTALMAT Catalunya – Promoció 2025–2027: Programa gratuït per a nois i noies amb interès especial per les matemàtiques: 📍 Dissabtes lectius (2025–2027) | 10:00–13:00 📌 Facultat de Matemàtiques i Estadística (UPC), Barcelona, Organitzat per: Societat Catalana de Matemàtiques & Feemcat.

Very excited to join the KnightHennessy community as I begin my PhD journey at Stanford University this fall! I am very grateful for their support of my research in theoretical computer science to ensure algorithms are used responsibly

The computer scientist Ryan Williams recently devised a mathematical procedure for transforming any algorithm — no matter what it does — into a form that uses much less computing space. “I just thought I was losing my mind.” quantamagazine.org/for-algorithms…

In a recent talk at the Simons Institute, Somesh Jha made a case for applying a security and cryptography mindset to evaluating the trustworthiness of machine learning systems, particularly in adversarial and privacy-sensitive contexts. bit.ly/43Xe0AP

New preprint with my supervisor, Varun! We show that padding the input of a Transformer with blank "pause" tokens strictly increases expressivity (in the finite-precision case), enabling it to compute everything in AC0.

Agradecida de poder divulgar la investigación científica que se realiza hoy en día desde la informática para el uso responsable de los algoritmos en sociedad. Podéis leer mi artículo en The Conversation ES y Revista Telos. ¡Gracias Laura G. De Rivera por el trabajo editorial!

Do you know how much text your LLM can generate in an hour? With different data requirements? With quantization? Go check out our results from benchmarking vLLM, LMSYS Org's SGLang, and @NVIDIA's TensorRT-LLM on DeepSeek, Llama, Qwen, and more! modal.com/llm-almanac/ad…

🚨 Ever wondered how much you can ace popular MCQ benchmarks without even looking at the questions? 🤯 Turns out, you can often get significant accuracy just from the choices alone. This is true even on recent benchmarks with 10 choices (like MMLU-Pro) and their vision

We (Moritz Hardt, Olawale Salaudeen,Joaquin Vanschoren) are organizing the Workshop on the Science of Benchmarking & Evaluating AI EurIPS Conference 2025 in Copenhagen! 📢 Call for Posters: rb.gy/kyid4f 📅 Deadline: Oct 10, 2025 (AoE) 🔗 More Info: rebrand.ly/bg931sf

See more about the upcoming Blackwell documentary, and donate to help support it, at gameofgeniusfilm.com For those who don't know, Blackwell also happens to be the first black full professor in the entire University of California system (1955, UC Berkeley Statistics).

Thinking of doing a PhD/research visit? Consider applying to/visiting ELLIS Institute Tübingen Intelligent Systems. We have PIs doing cool work, great funding, and most importantly, one of the best academic compute: 50+ GB200s, 250+ H100s, and many more A100 80GBs. Come join us!