Our paper "Beyond [cls]: Exploring the True Potential of Masked Image Modeling Representations" has been accepted to #ICCV2025! 🧵 TL;DR: Masked image models (like MAE) underperform not just because of weak features, but because they aggregate them poorly. [1/7]

bored? build your own git: wyag.thb.lt build your own database: build-your-own.org/database/ build your own redis: build-redis-from-scratch.dev/en/introduction build your own neural networks: sausheong.github.io/posts/how-to-b… build your own os: youtube.com/watch?v=ZjwvMc… I can keep going....

Applications are now open for the SWC Systems Neuroscience PhD Programme. Join us in London! 🧠 World-class neuroscience training 💰 Fully-funded 4-year programme 🖥️ Close links to Gatsby Computational Neuroscience Unit Apply by 3 Nov: sainsburywellcome.org/web/content/ph…

🚨 We’re honored to host Prof. Jürgen Schmidhuber (The Swiss AI Lab & KAUST) at the World Modeling Workshop 2026! ✨ A pioneer of modern AI, Prof. Jürgen Schmidhuber has made influential contributions that shaped the field — we’re thrilled to welcome him. 🌐 world-model-mila.github.io

Looking for a neuroscience postdoc opportunity with competitive salary + research funds? Apply for PNI's C.V. Starr Fellowship, which offers collaboration, mentorship, and up to 3 years of support! Applications due Oct 10. Learn more 🔗 pni.princeton.edu/apply/postdocs…

GPNs in the house! 2025.ccneuro.org/abstract_pdf/T… Go listen to Tim talk about it at Tubingen.

LLMs as a synthesis between symbolic and distributed approaches to language 💯 arxiv.org/abs/2502.11856

Poll time, just out of curiosity. ( This is in the context of our symposium, new talks and content updated every week: thoughtforms.life/symposium-on-t… . . ) How many of the professional mathematicians and physicists here think that the properties of mathematical objects are

Figuring out how the brain uses information from visual neurons may require new tools, writes Grace Lindsay. Hear from 10 experts in the field. thetransmitter.org/the-big-pictur…

Cool that ToI is picking up fundamental research! 😇

🕳️🐇Into the Rabbit Hull – Part I (Part II tomorrow) An interpretability deep dive into DINOv2, one of vision’s most important foundation models. And today is Part I, buckle up, we're exploring some of its most charming features.

Danielle Fong 🔆 there is a nice paper making precise a point similar to yours arxiv.org/abs/2110.09485 by Randall Balestriero et al

There is a simple but universal reality that more are beginning to appreciate: the physical world is rate limiting. Whether you believe that AGI is imminent, that digital tool-using agents will completely disrupt knowledge work, that abstract scientific and creative work will be

One obvious thing we **should** have done when moving to electronic publication was switch to rich citations that point directly to what you are citing in a paper - a data point, a figure, a method, a conclusion - and explain why you are citing it. Another thing publishers

the dream

To really understand a concept, you have to "invent" it yourself in some capacity. Understanding doesn't come from passive content consumption. It is always self-built. It is an active, high-agency, self-directed process of creating and debugging your own mental models.