A kick-ass chart by Matt McLean in @sondreus.bsky.social's fantastic retrospective of 250 years of American power www.economist.com/interactive/...

when you start seeing your learning rules research everywhere

Banger Dáil exchange where Paul Murphy is told not to call Elon Musk a "psychopath" in line with Dáil standing orders

It’s astounding how much the functioning of universities depends on a significant fraction of faculty being workaholic people pleasers with poor boundaries and ADHD

Absolutely this. Bell Labs was very productive when I was a kid. But they, as a monopoly, basically had the ability to tax the entire US population. Moreover, while they published or otherwise disseminated some of their findings, they kept some just for A T & T.

I somehow got Spain in our work world cup sweepstakes, £150 if they win the tournament, vamos lads!! 🇪🇸⚽️

Our new preprint is out!! The human brain runs on ~20 watts. Computers and modern AI systems require vastly higher energy. This gap raises a fundamental question: how does the brain compute so efficiently? In our new study, we take a critical step toward measuring and understanding this directly 1/3

It's *almost* as if Chris Olah is a spokesperson for a for-profit company that currently has no profitable business model and requires a continued suspension of disbelief from investors to keep the lights on

www.nature.com/articles/s42... banger piece in Nature Reviews Physics from Hiranya V. Peiris. 'fewer things better' is a fake idea.

Spatial synaptic regularization stabilizes learning across biological and artificial neural networks https://www.biorxiv.org/content/10.64898/2026.06.29.735142v1

How can the brain learn the hidden hierarchical structure from high dimensional data? In our latest work, we use synthetic datasets to analyze two classes of bio-plausible learning rules: variants of Direct Feedback Alignment, and local self-supervised learning. We find only the latter succeeds.

Anyone else finally starting to get to their long-neglected research-related list now that the students have dispersed? There are items that have been composting at the bottom for so long that I’d forgotten they existed. All except for one… ICYMI, my latest blog #AcademicSky tinyurl.com/4jhuk7nz

(apart from the important point that if we let AI do all our thinking, we are doomed) generally speaking this is why I have come to view presentations and paper writing as *doing science*, just as much as lab work or running computer simulations or whatever. Communication and discussion are central

An absolutely fascinating read!! tldr: the hard and important part of math was always building the scaffolding, while proofs (results) were a byproduct to socially capitalize on this work. AI disrupts that, as it produces proofs without re-injecting useful scaffolding into the system!

“As an informal experiment, I submitted the QED white paper to QED itself. The system assigned it a QED score of 46 and identified several methodological concerns.” x.com/sinabooeshag...

Avner Gvaryahu is a former IDF officer. The article is powerful. Human systems perform atrocities, but with friction in the form of internal objections and critique, whistleblowers, trauma, and those who show courage in ways that Gvaryahu has. The purpose of automation is to remove that friction.

The average life expectancy of a new Russian recruit—from arrival at a training ground to death in a combat zone—lies somewhere between 10 days and three weeks. Once sent onto the battlefield, they survive an average of 20 to 35 minutes. @peterfrankopan.bsky.social foreignpolicy.com/2026/06/25/r...

Hippocampal-cortical coupling dynamics drive system consolidation of remote memory https://www.biorxiv.org/content/10.64898/2026.06.22.733680v1

This paper is an absolutely beautiful combination of a deep set of cognitive ideas, a super creative task, a surprisingly impressive behaviour, and an elegant modelling framework. One of my favourite things in the last year.

This is remarkable.

I picked one paper among the 1% and asked Gemini to review it and after that review the analysis and gaps that QED had identified. Gemini was deeply impressed by QED. So based on that sample, the analysis tool seems to work very well. But the paper had 100% QED Score despite many major gaps.

Foundational! "The scientist doing basic research may not be at all interested in the practical applications of his work, yet the further progress of industrial development would eventually stagnate if basic scientific research were long neglected." From Science: The Endless Frontier, Bush, 1945

Agree there is no one best assessment metric but worse, a paper's quality is not even objective. Different people can get different utility from one paper Eg a synaptic physiologist vs a genomicist vs a pharma researcher could pull different insights from a paper linking synaptic genes to autism

Who needs sleep when you can just slow wave during wake? www.nature.com/articles/s41... #sleeppeeps

🔥 Published today, and just in time for summer conferences, here are your guidelines for presenting badly www.nature.com/articles/s41... rdcu.be/fqa92

I’m very excited to share that our paper is now out in Neuron! 🎉

“[HHMI] can just change their mind at any moment,” says a scientist [there]. Yep. It’s great Janelia is making a big, risky bet on one project: transparent fish. That kind of risk used to be hedged by the US govt’s long-term, “let 1000 flowers bloom” basic science approach. But now that’s dying.1/

Episode #42 in #TheoreticalNeurosciencePodcast: On neuronal identity and representational drift – with Timothy O'Leary @timothyoleary.bsky.social theoreticalneuroscience.no/thn42 How do neural and network functions stay the same while things change on the molecular and circuit levels?

🎯

"Right now, in some underfunded lab, a scientist is chasing a question that sounds pointless...a century of experience says that funding people to do this work is a good bet. If we stop making that bet, we will pay for it in cures no one will know to miss."

The AI posters for local community events are starting to annoy me I think it's the uniformity of the style. And The Style is some kind of saccharine kaleidoscope avatar-future I don't want to be part of

This is the stuff

More evidence, from a large-scale study in China, that using AI hurts learning if it undermines mental effort. When homework time drops due to AI use, so do test scores. Across studies, there is a clear theme: AI tutoring in support of classes is good, using AI to "help" with homework is bad.

A fundamental study of individual variation and contribution of learning to behavioural individuality. Conclusions are supported by compelling, rigorous analysis across a number of experiments in thousands of individuals across genotypes and conditions.

'Speaking this week at an event held by the Foundation for Science and Technology, Ian Chapman...was asked by former Conservative minister for universities and research, David Willetts, about a “nightmare scenario” for UKRI, where research grant applications are mass-produced with AI tools.' 1/4

"To achieve its incredibly tiny size, Danionella uses an evolutionary trick called paedomorphosis—meaning it essentially stops developing and retains larval (infant) features into adulthood. For example, it literally lacks a skull roof, lacks certain scales, and has a highly simplified skeleton"

"This will be the summer of paper writing", he thought optimistically, some said naively.

one anecdote: I was searching for a link to one of my old papers on Google, and the automated Gemini summary attributed all of my lab's work to my husband

Excited to share our first red ACh sensor! 🔴 We developed GRAB_rACh1h, the first genetically encoded red fluorescent sensor for recording acetylcholine dynamics in vivo. (1/3)

up and coming computational neuroscientists, now is a ideal time to pivot your efforts toward modelling danionella

Reading and writing and thinking hard are like going to the gym for your brain. Of course if we offload this work to GenAI we will lose cognitive skills in the same way it hurts me to run or lift weights without training

There is complete industry consensus on this. Every reputable cryptographer on the planet agrees: Signal’s cryptographic protocols are the most secure in class. Period. Just because your father's brother's nephew's cousin's former roommate told you otherwise doesn’t make it true

Job klaxon 📣 UCL is hiring two NeuroAI / computational neuroscience specialists (Lecturer or Associate Prof) — a big part of the new NeuroAI centre we're growing in Biosciences. Do consider applying if it's relevant to you. JD & details: https://www.jobs.ac.uk/job/DRX021 Happy to take questions.

Our latest review is out. We take a deep dive into the human mutations in microtubule motors that drive devastating neurodevelopmental and neurodegenerative diseases. Taking in motor protein biophysics, animal models of disease as well as patient phenotypes. onlinelibrary.wiley.com/doi/full/10....

new paper, #NeuroAI 📣📜 Can measured cortical organization be used as an inductive bias for artificial recurrent neural networks? In this work, we ask whether cortical geometry, wiring, and function can push RNNs learn. Not as metaphor, but as measurable structure! 1/n🧵👇 arxiv.org/abs/2606.14975

No. What I am saying is that independent of whether I read a PI's papers, what is written in a proposal is the only thing that should be considered for it's evaluation. As reviewer, it's not my job to read some papers to understand what the proposer is saying, the proposal needs to stand on it's own

Heliconius butterflies are among the longest-lived butterflies ever documented. @katherinejwu.com reports on the researchers trying to crack the butterflies’ secret to defying aging:

Interesting Transmitter article by Brian DePasquale on what will be the effect of the increasingly widespread use of LLMs to code up research ideas in neuroscience. Includes a somewhat sceptical commentary from me. 😉

Can I also add that James won best poster award! Really cool work on dynamics of linear RNNs, we are writing up soon for a preprint server near you

Appreciation post for my incredible supervisor Cian O'Donnell @cianodonnell.bsky.social. Such an inspiration & wonderful mentor as Supervisor & School Research Director. He organised a brilliant ISRC Discovery Day today @ulsteruni.bsky.social, showcasing the breadth & depth of our research.