A series of recordings from the second rendition of the best conference on theoretical AI alignment (ILIAD) just dropped! youtube.com/playlist?lis...

1/🧵 Out today in @frontiersin.bsky.social : “Consciousness science: where are we, where are we going, and what if we get there?” By @axc.bsky.social @liadmudrik.bsky.social, & me (a CIFAR & @erc.europa.eu production 🧠). www.frontiersin.org/journals/sci...

🚨The Formalism-Implementation Gap in RL research🚨 Lots of progress in RL research over last 10 years, but too much performance-driven => overfitting to benchmarks (like the ALE). 1⃣ Let's advance science of RL 2⃣ Let's be explicit about how benchmarks map to formalism 1/X

It makes sense that evolution selected neural circuits that make us to “want” certain things, as that ensures robust proactive behaviour… But why that wanting has to be experienced at all, instead of acting simply as an unconscious compulsion? (I think of this as the “little hard problem”)

Is the the neuroimaging community (broadly conceived) submitting to 𝗡𝗲𝘂𝗿𝗼𝗶𝗺𝗮𝗴𝗲 again????? Not to be publication police but the shift had seemed major 1-2 years ago. Looking at the editorial board it seems to have serious scientists once more. Curious about thoughts of the community.

🥲

I deeply sympathise with every word in this post 👇🏽

If you love theory and want to make a difference for AI alignment, please apply and join us! pibbss.ai/pibbssxiliad... (Please share around!)

🚨New paper: arxiv.org/abs/2510.05786 *Shapley values beyond game theory* We show that Shapley values aren’t just about dividing payoffs--they are the right way to project down any higher-order structure. We generalise them, and Möbius inversions, in important ways: 🧵

After a long time in the making, my paper on nonequilibrium thermodynamics of Darwinian evolution has been published in Philosophical Transactions B. Updated version: arxiv.org/abs/2112.02809

The usual explanation for great music is that it's predictable but not too predictable. But there's more to it, said @frosas.bsky.social at #ALIFE2025. We prefer music when we perceive agency behind it. Improvisation trumps a prepared performance. See paper at www.frontiersin.org/journals/psy.... ⚙️🧫

It’s not the method that makes you causal it’s the assumptions

It is often argued that the origin of life must be related to "phase transitions" This concept is often misused, or even abused. In a new paper with @ricardsole.bsky.social we discuss which types of transitions might be involved and some relevant models. royalsocietypublishing.org/doi/full/10....

Nearly 4,000 y ago, Egyptians finished their pyramids Meanwhile, across the ocean, termites in NE Brazil were building something even larger: a hidden empire of 200M mounds, over an area the size of UK 🌍🐜 Still active today, a masterpiece of non-human engineering! 🧪🌐 HT @ricardsole.bsky.social

Finally published: “Toward a unified taxonomy of information dynamics via Integrated Information Decomposition” www.pnas.org/doi/10.1073/... An information-theoretic framework to analyse complex multivariate dynamical processes ✨

Postdoc job! I expect to have an opening at Johns Hopkins for a postdoctoral researcher working somewhere in the broad realms of physics, philosophy, and complexity. Apply at Academic Jobs Online: academicjobsonline.org/ajo/jobs/30496

𝗪𝗵𝗮𝘁'𝘀 𝗰𝗼𝗴𝗻𝗶𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗶𝘀 𝗶𝘁 𝗲𝗺𝗲𝗿𝗴𝗲𝗻𝘁? Don't miss the Neuroscience and Philosophy Salon. Earl Miller and team will discuss recent paper and we'll have plenty of discussion. Open to all. Sept 12, noon EST-US umd.zoom.us/meeting/regi... #neuroskyence @earlkmiller.bsky.social

Finally published: “Characterising high-order interdependence via entropic conjugation” rdcu.be/eB3k3 On how a “conjugation” of information illuminates the space possible information-theoretic metrics of high-order phenomena!

Our department is seeking applicants for an Assistant Professor position with a focus on affective science. Please apply and/or pass this along to anyone who might be interested! facultypositions.stanford.edu/en-us/job/49...

Preprint time: “Compatibilist emergence for the science of consciousness” osf.io/preprints/ps... On how a multifaceted understanding of emergence can illuminate the link between consciousness and life!

The first one is some crazy work by @monmitabhar.bsky.social, Tanumoy, @hari-flies.bsky.social and Kamal Kishore. They eavesdropped on worm conversations and found the exciting phenomenon of movement of memory from one worm to another 2/n www.biorxiv.org/content/10.1...

Third edition of the emergence workshop is almost here! You can register using the QR code on the poster Any questions, you can message me here or send an email!

What is emergence? This is one of the most central ingredients of complexity, and a challenging one to formalize. Here's a paper by @seanmcarroll.bsky.social & Achyth Parola that attempts to classify different forms of emergence. @manlius.bsky.social @sfiscience.bsky.social arxiv.org/pdf/2410.15468

🗺️ Pathfinding: a neurodynamical account of intuition 📔 We #review the neuroscience of #intuition, and present it as a predictive embodied process, emergent from hippocampal replay & brain metastability 🤝 with S. Kotler, M. Mannino, K. Friston, G. Buzsáki & J.A.S Kelso 🔗 nature.com/articles/s42...

Amazing opportunity 👇🏽

Are you interested in our work on complex physical systems and applications (to neural systems)? Wanna join us in Sydney? 🐨 🦘 We have two open positions: 1. Postdoc position: usyd.wd105.myworkdayjobs.com/USYD_EXTERNA... 2. Fully-funded PhD position: www.sydney.edu.au/scholarships...

𝗪𝗵𝗮𝘁 𝗺𝗼𝘁𝗶𝘃𝗮𝘁𝗲𝘀 𝗮𝗻𝗶𝗺𝗮𝗹𝘀 𝘁𝗼 𝗮𝗰𝘁, 𝘄𝗵𝗮𝘁 𝗶𝘀 𝗺𝗼𝘁𝗶𝘃𝗮𝘁𝗶𝗼𝗻? What are the most developed frameworks of the brain basis of motivation? Older ideas, new ideas, etc. #neuroskyence

Applications for next year’s Guggenheim Fellowships are now open! @guggfellows.bsky.social www.gf.org/program/how-...

New bogpost about our paper on world models for safe and interpretable AI sandboxing: www.lesswrong.com/posts/L6Z6K8... Including funky ideas such as what makes a world model, the metaphysical and empirical components of Bayesian beliefs, and the mathematics of the Kantian noumenon…

This is absolutely insane: a LLM of worlds! 🤯🤯 deepmind.google/discover/blo...

𝐖𝐡𝐚𝐭 𝐰𝐞 𝐚𝐫𝐞 𝐫𝐞𝐚𝐝𝐢𝐧𝐠 “Top-down and bottom-up neuroscience: overcoming the clash of research cultures,” by Fernando Rosas @frosas.bsky.social 𝑒𝑡 𝑎𝑙. in 𝑁𝑎𝑡𝑢𝑟𝑒 𝑅𝑒𝑣𝑖𝑒𝑤𝑠 𝑁𝑒𝑢𝑟𝑜𝑠𝑐𝑖𝑒𝑛𝑐𝑒 www.nature.com/articles/s41... #neuroskyence

Great news article from the Basque Center for Applied Mathematics about our recent work: www.bcamath.org/en/news-even...

Finally published: “Explosive neural networks via higher-order interactions in curved statistical manifolds” nature.com/articles/s4146… Enhancing the capabilities of recurrent neural networks by deforming their geometry!

Explosive research with an explosive title 😎🌟👇🏼

I'm curious if any senior nonhuman primate Neuro-AI researchers would be interested in joining Queen's University if we were to obtain a research chair position (full professor level)? Could you please send me a confidential message to indicate your interest? [email protected]

You know how RL is that whole big thing nowadays? Present your work at the first-ever New York Reinforcement Learning Workshop (NYRL), co-organized by Amazon, Columbia Business School & NYU Tandon School of Engineering. ny-rl.com!

Very sad to learn of the death on 18th July of Margaret (Maggie) Boden, a titan of cognitive science and AI. I met her many times, and respected her greatly. www.theargus.co.uk/memorials/de...

Finally published: “Explosive neural networks via higher-order interactions in curved statistical manifolds” nature.com/articles/s4146… Enhancing the capabilities of recurrent neural networks by deforming their geometry!

Trying to train RNNs in a biol plausible (local) way? Well, try our new method using predictive alignment. Paper just out in Nat. Com. Toshitake Asabuki deserves all the credit! www.nature.com/articles/s41...

This was such a fun read and really resonates with my philosophy on the balance of the two perspectives (and the power of modeling as a tool to integrate them).

"Concretely, computational modelling provides one fruitful avenue for achieving this synthesis (although certainly not the only avenue)" 👏

Hear, hear! “We advocate for… a more constructive mindset in which different dimensions of progress in neuroscience are explicitly acknowledged and their value is recognized - to temper their respective limitations and build on each other’s strengths.”

Beautifully written and important piece here:

cool piece on the meaning of scientific progress and "cultural" differences among neuroscientists "...towards a more constructive mindset in which different dimensions of progress in neuroscience are explicitly acknowledged and their value is recognized" 💯 👇 www.nature.com/articles/s41...

Finally published: “Top-down and bottom-up neuroscience: overcoming the clash of research cultures” www.nature.com/articles/s41... Looking for ways to better understand different neuroscientific perspectives and enable productive collaborations

This preprint looks awesome… another step towards setting cybernetic dreams (internal model principle, good regulator theorem, etc) into a rigorous form ✨

Is it time to bring back natural philosophy? ABC Radio chats with Sandra Mitchell, Alan Guth, Melanie Mitchell @melaniemitchell.bsky.social, Anil Seth @anilseth.bsky.social, David Albert and me, as produced by @danfalk.bsky.social. www.abc.net.au/listen/progr...

Ever since I started thinking seriously about AI value alignment in 2016-7, I've been frustrated by the inadequacy of utility+RL theory to account for the richness of human values. Glad to be part of a larger team now moving beyond those thin theories towards thicker ones.