Sign in

Romain Brette

@romainbrette.bsky.social
898 followers 134 following 458 posts

I think about the biology of cognition in brains and in protozoa. romainbrette.fr New book: The Brain, in Theory. press.princeton.edu/books/paperback…

PostsRepliesMedia
Romain Brette @romainbrette.bsky.social · 08/10/2026
Behavioral work is not that demanding. The main difficulty, I would say, is that the little critters never do what you want. Which is kind of the point, I guess. Actually there's still much computational work to do on video analysis (not currently solved by AI).
030
Romain Brette @romainbrette.bsky.social · 07/10/2026
Well for one, I'm pretty sure OpenAI doesn't care. Plus, even a theoretician can do experiments!
160
Romain Brette @romainbrette.bsky.social · 07/10/2026
If you want to move to unicellular neuroscience I will help :)
140
Reposted by Romain Brette
Alexandra Delbot @alexandradelbot.bsky.social · 07/10/2026
Le cerveau est un organe à la fois fascinant et complexe. Il n’est pas toujours aisé de comprendre son fonctionnement. Alors... on se laisse happer par des explications séduisantes et intuitives qui sont en réalité fausses - ce sont les neuromythes. Sommes-nous accros aux mythes sur le cerveau ? 🧠
radiofrance.fr
Sommes-nous accros aux mythes sur le cerveau ?
Dopamine détox, cerveau masculin ou féminin, cerveau droit ou gauche, les neuromythes ont la cote dans nos librairies. Quels discours politiques se cachent derrière ces fausses idées sur le cerveau ? ...
38126
Reposted by Romain Brette
Jérémie Naudé @jeremie-n.cpesr.fr · 07/10/2026
Neuro arnaque sort aujourd'hui chez Textuel. On nous explique les crise écologiques par le striatum, l'inégalité scolaire par les gènes, les préférences genrées par l'évolution... La neurobiologie réelle raconte une tout autre histoire. www.editionstextuel.com/livre/neuro_...
editionstextuel.com
Editions Textuel . Livre . Neuro arnaque
Photographie, archives, musique et danse sont les axes majeurs du catalogue beaux-livres des éditions Textuel fondées en 1995. Deux collections dédiées au débat d’idées nourrissent le catalogue côté s...
1211052
Reposted by Romain Brette
Marielle Debos @debos.bsky.social · 06/10/2026
Avant la manif, pensez à écrire à vos député·es. On a 48h pour les alerter - au moment de l'examen du Projet de Loi de Finances. J'ai écrit et envoyé ça:
Mail type pour alerter les députés du délabrement des universités et des conditions d'études déplorables offertes aux jeunes
1911
Reposted by Romain Brette
Thibaut Brunet @thibautbrunet.bsky.social · 05/10/2026
One reason (among many) to study protists: as a toolbox, nature is unbeatable
2459
Romain Brette @romainbrette.bsky.social · 05/10/2026
In that post, I pointed out that an LLM does not understand language because LLMs are trained on form, not meaning. There's a paper by @emilymbender.bsky.social and @akoller.bsky.social that makes this point in detail. From 2020, but still accurate. aclanthology.org/2020.acl-mai...
aclanthology.org
Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data
Emily M. Bender, Alexander Koller. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020.
042
Romain Brette @romainbrette.bsky.social · 05/10/2026
Great conversation with @tommyblanchard.bsky.social at "Cognitive Wonderland": cognitivewonderland.substack.com/p/the-brain-...
cognitivewonderland.substack.com
The Brain, in Theory
Does the brain compute? A conversation with Romain Brette
071
Reposted by Romain Brette
ASUR @agorasur.bsky.social · 04/10/2026
#VeilleESR #ESR Le premier rapport des Assises pour l'Université et pour la Recherche propose à la représentation nationale un plan de redressement budgétaire pour l'Université, la Recherche et la Vie Etudiante. A télécharger ici: raw.githubusercontent.com/cpesr/assise...
Le premier rapport des Assises pour l'Université et pour la Recherche propose à la représentation nationale un plan de redressement budgétaire pour l'Université, la Recherche et la Vie Etudiante.
02330
Romain Brette @romainbrette.bsky.social · 01/10/2026
Ah ben non dotation c'est un financement récurrent, non sujet à appel d'offre. Et bourse c'est pour des financements individuels. Financement de projet, ou à la limite subvention.
010
Romain Brette @romainbrette.bsky.social · 29/09/2026
Les titres sont faits par ChatGPT, ou bien c'est juste un titre scandaleux?
021
Reposted by Romain Brette
Emily M. Bender @emilymbender.bsky.social · 26/09/2026
I think there are two key lessons from Linguistics as we deal with LLMs/chatbots being imposed on us everywhere. Grateful for Undark Magazine giving space for me to lay these out: undark.org/2026/09/25/i...
undark.org
Interview: What Does the 'I' in AI Really Mean?
Five scientific experts discuss whether artificial intelligence is — or might one day become — truly intelligent.
49319
Romain Brette @romainbrette.bsky.social · 24/09/2026
++ Linking the unity of consciousness to the unity of metabolism Also, @anilseth.bsky.social makes an error by relating thermodynamic free energy to informational free energy (not a physical quantity). An organism imports (not minimizes) free energy in order to survive. The opposite basically.
120
Romain Brette @romainbrette.bsky.social · 24/09/2026
Let's celebrate the "beauty of microscopy" with fake microscopy images! Yay!
0210
Reposted by Romain Brette
Julio Alcántara @julioalcante.bsky.social · 22/09/2026
We are looking forward to tomorrow, on September 23rd, 2026 at 10h30 for the presentation of the book The Brain in Theory and the Q&A with Romain Brette at the Institute Pasteur. Here it's the link to follow the event online and be part of the Q&A. www.crowdcast.io/c/brain-in-t...
crowdcast.io
Reading club: The Brain in Theory by Romain Brette
Register now for Reading club: The Brain in Theory by Romain Brette on crowdcast, scheduled to go live on July 22, 2026, 04:00 PM GMT+2.
1124
Romain Brette @romainbrette.bsky.social · 22/09/2026
Isn't it 10:30am?
100
Romain Brette @romainbrette.bsky.social · 22/09/2026
Some background about Paramecium: Integrative Neuroscience of Paramecium, a “Swimming Neuron” www.eneuro.org/content/8/3/...
eneuro.org
Integrative Neuroscience of Paramecium, a “Swimming Neuron”
Paramecium is a unicellular organism that swims in fresh water by beating thousands of cilia. When it is stimulated (mechanically, chemically, optically, thermally…), it often swims backward then turns and swims forward again. This “avoiding reaction” is triggered by a calcium-based action potential. For this reason, some authors have called Paramecium a “swimming neuron.” This review summarizes current knowledge about the physiological basis of behavior of Paramecium .
020
Romain Brette @romainbrette.bsky.social · 22/09/2026
We think this is exciting because foraging is a complex ecologically relevant behavior potentially involving sensory processing, memory, decision making, and motor control, and now we can study this in a unicellular organism!
120
Romain Brette @romainbrette.bsky.social · 22/09/2026
What happens is that locomotor cilia stop, while cilia along its oral groove (its digestive tube) still beat, and this seems to confer an attraction to surfaces. Then at some point, locomotor cilia start again and Paramecium flees to greener pastures. (below: fluid motion around a paramecium)
Fluid motion along the oral groove of an immobilized cell, shown using particles.
130
Romain Brette @romainbrette.bsky.social · 22/09/2026
But it does not simply become inactive. In fact, it seems to attach to surfaces of any orientation. We even noticed that it can stop on top, at the interface between air and water. While immobile, its oral cilia are in fact very active: it feeds.
Paramecium feeding on bacteria, immobile on a glass slide.
120
Romain Brette @romainbrette.bsky.social · 22/09/2026
We found out that, in its normal life, Paramecium alternates between exploring its surroundings and stopping on surfaces where it feeds, similar to the foraging behavior of many animals. When it senses bacteria, a calcium influx triggers stopping.
120
Romain Brette @romainbrette.bsky.social · 22/09/2026
Paramecium is a unicellular swimming eukaryote that feeds on bacteria. It was once a model organism in neuroscience under the nickname “swimming neuron”, because it controls its motility with action potentials. (here a drawing of the avoiding reaction triggered by an action potential; Jennings 1906)
a drawing of the avoiding reaction triggered by an action potential; Jennings 1906
120
Romain Brette @romainbrette.bsky.social · 22/09/2026
New preprint! “Active control of surface immobilization in the foraging behavior of Paramecium”. More about the elementary cognition of these little critters! www.biorxiv.org/content/10.6...
biorxiv.org
1121
Reposted by Romain Brette
ASUR @agorasur.bsky.social · 22/09/2026
Six principles for supporting academic freedom in universities. www.nature.com/articles/d41...
nature.com
Six principles for supporting academic freedom in universities
Many academics are concerned about threats to independent enquiry in higher education. Here’s how it can be protected while taking legitimate concerns into account.
044
Romain Brette @romainbrette.bsky.social · 21/09/2026
Excellent article dans Le Monde sur mes bestioles favorites.
082
Reposted by Romain Brette
Julio Alcántara @julioalcante.bsky.social · 16/09/2026
We thank Ferney Manrique for his interesting presentation about chapters 4 and 9. Today's video recording: www.crowdcast.io/c/brain-in-t... This is the last session of the reading group and next week we will have the book presentation at the Institute Pasteur on September 23 at 10:30 CET.
crowdcast.io
Reading club: The Brain in Theory by Romain Brette
Register now for Reading club: The Brain in Theory by Romain Brette on crowdcast, scheduled to go live on July 22, 2026, 04:00 PM GMT+2.
153
Reposted by Romain Brette
Emily M. Bender @emilymbender.bsky.social · 18/09/2026
I have been yelling since 2021 about how using synthetic text extruding machines for information access is a TERRIBLE IDEA. And yet and yet and yet: they are marketing so heavily that way that people keep falling for it. edition.cnn.com/2026/09/18/p...
edition.cnn.com
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
519168
Romain Brette @romainbrette.bsky.social · 16/09/2026
This talk about "alignment" shows that the companies don't take the risks seriously. They basically say that they must be allowed to develop massive autonomous technologies with no restrictions whatsoever, and then pretend that they're working on making those technologies "nice persons".
071
Romain Brette @romainbrette.bsky.social · 16/09/2026
All the "alignment" rhetoric places the problem inside the "AI", seen as an anthropomorphised moral person, but obviously the problem is outside: it is with what we allow these algorithms to interfere with.
172
Romain Brette @romainbrette.bsky.social · 16/09/2026
Letting kids play with the red button is dangerous. But it's not because kids are so much smarter than adults. The point is not to prevent AI to get "too smart", but to have strict barriers in place. An LLM did not figure out how to send requests over internet: an engineer implemented the feature.
1102
Romain Brette @romainbrette.bsky.social · 14/09/2026
Time to repost this: "On the existential risks of artificial intelligence" romainbrette.fr/on-the-exist...
romainbrette.fr
On the existential risks of artificial intelligence
The impressive progresses in machine learning have revived the fear that humans might eventually be wiped out or enslaved by artificial superintelligences. This is hardly a new fear. For example, t…
193
Reposted by Romain Brette
Philip Ball @philipcball.bsky.social · 12/09/2026
It's not impossible that AI could present catastrophic risk if we do something deeply dumb like put it in charge of nuclear arsenals or enable rogue agents to use it for bioweapons design. Neither requires "superintelligence".
5337
Romain Brette @romainbrette.bsky.social · 11/09/2026
Several times a week! Clearly AI written.
030
Reposted by Romain Brette
Princeton University Press @princetonupress.bsky.social · 11/09/2026
@romainbrette.bsky.social is on the Converging Dialogues podcast to discuss his new book, The Brain, In Theory with host @xavierbonilla.bsky.social: buff.ly/PIcn3vB
convergingdialogues.substack.com
#505 - The Brain Is Not A Machine: A Dialogue with Romain Brette
Listen now | In this episode, Xavier Bonilla has a dialogue with Romain Brette about the brain and various theoretical concepts in neuroscience.
022
Romain Brette @romainbrette.bsky.social · 11/09/2026
Some sense in the "rogue agents" debacle. Anthropomorphic metaphors of machines are misleading, just like machine metaphors of the brain are misleading.
064
Reposted by Romain Brette
Julio Alcántara @julioalcante.bsky.social · 03/09/2026
For the next session, we will have Raoul Doyez from Sorbonne Université presenting Chapter 8 and a brief recap of Chapter 5: www.crowdcast.io/c/brain-in-t... The topics are neural codes and the organization of brain processes. Join us on September 9th, 2026 at 16h CET. @romainbrette.bsky.social
crowdcast.io
Reading club: The Brain in Theory by Romain Brette
Register now for Reading club: The Brain in Theory by Romain Brette on crowdcast, scheduled to go live on July 22, 2026, 04:00 PM GMT+2.
173
Romain Brette @romainbrette.bsky.social · 08/09/2026
more in my book, chapter 5 ;)
In the same way, in sensory tasks where input stimuli are varied along n di-
mensions, we would expect the activity of a large population of sensory neu-
rons driven by ­ those stimuli to also vary along n dimensions. This observation
derives from the structure of the task rather than from the way the ­ neurons deal
with the stimuli. Quite logically, it is indeed found that the dimensionality of
neural manifolds in sensory areas varies with the kind of stimulus ( Jazayeri and
Ostojic, 2021). It could not be other­ wise, since dimensions are defined relative
to a par­ tic­ u­ lar category of stimuli (e.g., the orientation of a grating, the size of
an object, ­ etc.), as we ­ will discuss further in the next section. In the standard
connectionist view, the structure of the neural manifold is determined by the
connectivity structure of the network (Langdon et al., 2023), but this does not
work if that structure changes ­ every time the animal does something ­ else.
051
Romain Brette @romainbrette.bsky.social · 08/09/2026
The issue, as with all coding approaches, is that the dimensional analysis is done by the observer, not the brain. You could film a pong game and find that the image lives in a 4D manifold. Does that mean the camera tracks the ball?
140
Reposted by Romain Brette
Melanie Mitchell @melaniemitchell.bsky.social · 08/09/2026
Hey @ericlipton.nytimes.com 👋 AI is not “evolving” on its own. *People* are choosing how to build it. *People* are choosing how to train it. *People* are choosing how to run it. Maybe the problem is that they are the *wrong people* to make these choices.
28636174
Romain Brette @romainbrette.bsky.social · 05/09/2026
La video: www.youtube.com/live/mDXmhOG...
youtube.com
les assises de la recherche 4 septembre
YouTube video by Politique Des Sciences
050
Reposted by Romain Brette
Nicole Rust @nicolecrust.bsky.social · 04/09/2026
The superposition of these two studies — both incredible — reflect a snapshot of "one big thing" systems neuroscience has yet to resolve. In mice: 1. Task-related signals are everywhere. 2. Absent large swaths of "everywhere," suprisingly sophisticated learning & behavior remains.
Post by Enny van Beest, "Brainwide representation of navigation"Post by Jieyu Zheng, "How fast can mice learn complex mazes?"
1011738
Romain Brette @romainbrette.bsky.social · 03/09/2026
A presentation by Farid Zahnoun of my chapter on anticipation (from The Brain, in Theory).
040
Romain Brette @romainbrette.bsky.social · 02/09/2026
Ma contribution aux assises pour l'Université et pour la Recherche (3-4/9/2026), concernant le financement de la recherche. romainbrette.fr/pour-un-fina...
romainbrette.fr
Pour un financement collégial, non bureaucratique de la recherche
Ce texte est ma courte contribution aux assises pour l’Université et pour la Recherche (3-4/9/2026), concernant le financement de la recherche. La recherche repose sur deux principes indissoc…
196
Romain Brette @romainbrette.bsky.social · 02/09/2026
Is it just me who sees the hype for humanoid robots as a fascination for slavery? Why would we need our technological aids to look like people? My dishwasher looks just fine as it is.
180
Reposted by Romain Brette
Julio Alcántara @julioalcante.bsky.social · 02/09/2026
We are looking forward for today's presentation of Chapter 7 by Farid Zahnoun on September 2nd, 2026 at 16h CET. The topic is anticipation in neuroscience and living beings. Join us! @r3rt0.bsky.social @k4tj4.bsky.social www.crowdcast.io/c/brain-in-t...
crowdcast.io
Reading club: The Brain in Theory by Romain Brette
Register now for Reading club: The Brain in Theory by Romain Brette on crowdcast, scheduled to go live on July 22, 2026, 04:00 PM GMT+2.
1114
Romain Brette @romainbrette.bsky.social · 01/09/2026
Great point. There's a machine producing a fiction, and we are asking whether the character in that fiction is conscious.
062
Reposted by Romain Brette
Philip Ball @philipcball.bsky.social · 31/08/2026
And it led me to this, which is also good. blog.genesmindsmachines.com/p/no-alphafo...
blog.genesmindsmachines.com
No, AlphaFold has not completely solved protein folding
Biology is hard. Yes, even for AI.
2159
Romain Brette @romainbrette.bsky.social · 31/08/2026
A critique of grant-based funding of research. Koppanyi, 1945. www.jstor.org/stable/40220...
The theories and ways of thinking of scientists are as diverse as
the world itself. How can we set up a court with the power to
pass on what is good and what is not good for science? It might
have the same effect as having a court of art, directing artists
to follow only the pet artistic theories of members of the board.
Although few people are aware of what every scientist knows,
science is as much an expression of the individual human being as
any of the arts.
0121
Romain Brette @romainbrette.bsky.social · 31/08/2026
(N.B.: Renforcer le pouvoir du Parlement, c'était précisément la motivation pour la gauche de voter la LOLF en 2001, qui a introduit la gestion par objectifs, ie le nouveau management public. Conséquences: points SIGAPS, COMP et ANR)
010