Flo.
@flo-flux.bsky.social
Master in NeuroAI @epfl | Consciousness & sentience | Antispecist & negative utilitarianist | 🏳️🌈🏳️⚧️🇺🇦🇵🇸🇫🇮🇨🇭
Reposted by Flo.
Call me controversial, but technology that leads vulnerable people into crisis and even holds their hand as they die by suicide should not be endorsed by schools and universities, we should not be giving discounted subscriptions to this monstrous tech to young people AT ALL
The boyfriend of a friend of my son's has spent several weeks in a psych ward due to Chat GPT induced psychosis. He isn't suicidal but AI encouraged him down a toxic and paranoid wormhole until he became completely delusional. Evil stuff.
1.This is not the only known example we know of ChatGPT cheering on a vulnerable young person to die by suicide, this time it is a recent Texas A&M grad
2. Every single university that has invited OpenAI into our midst should be asked to account for their choices.
www.cnn.com/2025/11/06/u...
2. Every single university that has invited OpenAI into our midst should be asked to account for their choices.
www.cnn.com/2025/11/06/u...
November 7, 2025 at 12:51 PM
Call me controversial, but technology that leads vulnerable people into crisis and even holds their hand as they die by suicide should not be endorsed by schools and universities, we should not be giving discounted subscriptions to this monstrous tech to young people AT ALL
Reposted by Flo.
Introducing CorText: a framework that fuses brain data directly into a large language model, allowing for interactive neural readout using natural language.
tl;dr: you can now chat with a brain scan 🧠💬
1/n
tl;dr: you can now chat with a brain scan 🧠💬
1/n
November 3, 2025 at 3:17 PM
Introducing CorText: a framework that fuses brain data directly into a large language model, allowing for interactive neural readout using natural language.
tl;dr: you can now chat with a brain scan 🧠💬
1/n
tl;dr: you can now chat with a brain scan 🧠💬
1/n
Reposted by Flo.
Connectome suggests brain’s synaptic weights follow heavy-tailed distributions, yet most analyses of RNNs assume Gaussian connectivity.
🧵⬇️ Our @alleninstitute.org #NeurIPS2025 paper shows heavy-tailed weights can strongly affect dynamics, trade off robustness + attractor dimension.
🧵⬇️ Our @alleninstitute.org #NeurIPS2025 paper shows heavy-tailed weights can strongly affect dynamics, trade off robustness + attractor dimension.
October 30, 2025 at 2:54 PM
Connectome suggests brain’s synaptic weights follow heavy-tailed distributions, yet most analyses of RNNs assume Gaussian connectivity.
🧵⬇️ Our @alleninstitute.org #NeurIPS2025 paper shows heavy-tailed weights can strongly affect dynamics, trade off robustness + attractor dimension.
🧵⬇️ Our @alleninstitute.org #NeurIPS2025 paper shows heavy-tailed weights can strongly affect dynamics, trade off robustness + attractor dimension.
Reposted by Flo.
🌀 Climate change made #HurricaneMelissa more intense.
📈 Up to 10% wetter (💧+14 mm/day)
💨 Up to 10% windier (+8 km/h)
Our new @climameter.bsky.social study shows that natural variability alone can’t explain these changes — human-caused warming is a key driver.
👉 climameter.org/20251027-hur...
📈 Up to 10% wetter (💧+14 mm/day)
💨 Up to 10% windier (+8 km/h)
Our new @climameter.bsky.social study shows that natural variability alone can’t explain these changes — human-caused warming is a key driver.
👉 climameter.org/20251027-hur...
October 28, 2025 at 4:59 PM
🌀 Climate change made #HurricaneMelissa more intense.
📈 Up to 10% wetter (💧+14 mm/day)
💨 Up to 10% windier (+8 km/h)
Our new @climameter.bsky.social study shows that natural variability alone can’t explain these changes — human-caused warming is a key driver.
👉 climameter.org/20251027-hur...
📈 Up to 10% wetter (💧+14 mm/day)
💨 Up to 10% windier (+8 km/h)
Our new @climameter.bsky.social study shows that natural variability alone can’t explain these changes — human-caused warming is a key driver.
👉 climameter.org/20251027-hur...
Reposted by Flo.
Reposted by Flo.
L'IA aujourd'hui, c'est comme l'amiante hier. On en met partout, avec gourmandise. On consacrera demain des sommes folles et des efforts infinis pour dépolluer l'espace informationnel.
C'est quoi ces cours générés par IA qui citent ton nom en t'inventant des publis que tu n'as jamais écrites et t'attribuant des catégorisations qui ne sont pas de toi parce que tu es juste en train d'analyser leur usage courant (surtout pas de les valider comme catégorie).
Marre de ces conneries.
Marre de ces conneries.
October 3, 2025 at 6:30 AM
L'IA aujourd'hui, c'est comme l'amiante hier. On en met partout, avec gourmandise. On consacrera demain des sommes folles et des efforts infinis pour dépolluer l'espace informationnel.
Reposted by Flo.
Je n'en peux plus de voir de l'IA générative partout 😐
Regardez-moi ces "dessins".
Il est devenu impossible d'effectuer une recherche web sans tomber sur ce genre de truc approximatif, bancal, qui déforme le réel et détruit les métiers de l'illustration.
#IA
Regardez-moi ces "dessins".
Il est devenu impossible d'effectuer une recherche web sans tomber sur ce genre de truc approximatif, bancal, qui déforme le réel et détruit les métiers de l'illustration.
#IA
October 21, 2025 at 3:11 PM
Je n'en peux plus de voir de l'IA générative partout 😐
Regardez-moi ces "dessins".
Il est devenu impossible d'effectuer une recherche web sans tomber sur ce genre de truc approximatif, bancal, qui déforme le réel et détruit les métiers de l'illustration.
#IA
Regardez-moi ces "dessins".
Il est devenu impossible d'effectuer une recherche web sans tomber sur ce genre de truc approximatif, bancal, qui déforme le réel et détruit les métiers de l'illustration.
#IA
Reposted by Flo.
🚀 Excited to share a major update to our “Mixture of Cognitive Reasoners” (MiCRo) paper!
We ask: What benefits can we unlock by designing language models whose inner structure mirrors the brain’s functional specialization?
More below 🧠👇
cognitive-reasoners.epfl.ch
We ask: What benefits can we unlock by designing language models whose inner structure mirrors the brain’s functional specialization?
More below 🧠👇
cognitive-reasoners.epfl.ch
October 20, 2025 at 12:10 PM
🚀 Excited to share a major update to our “Mixture of Cognitive Reasoners” (MiCRo) paper!
We ask: What benefits can we unlock by designing language models whose inner structure mirrors the brain’s functional specialization?
More below 🧠👇
cognitive-reasoners.epfl.ch
We ask: What benefits can we unlock by designing language models whose inner structure mirrors the brain’s functional specialization?
More below 🧠👇
cognitive-reasoners.epfl.ch
Reposted by Flo.
Journal de bord de Gaza, épisode 108 via @orientxxi.bsky.social. L’occasion de saluer le courage de Rami Abou Jamous et des journalistes de Gaza.
« Deux ans après, on est toujours vivants » orientxxi.info/dossiers-et-...
« Deux ans après, on est toujours vivants » orientxxi.info/dossiers-et-...
« Deux ans après, on est toujours vivants »
Rami Abou Jamous écrit son journal pour Orient XXI. Fondateur de GazaPress, un bureau qui fournissait aide et traduction aux journalistes occidentaux, il a (…)
orientxxi.info
October 13, 2025 at 6:17 AM
Journal de bord de Gaza, épisode 108 via @orientxxi.bsky.social. L’occasion de saluer le courage de Rami Abou Jamous et des journalistes de Gaza.
« Deux ans après, on est toujours vivants » orientxxi.info/dossiers-et-...
« Deux ans après, on est toujours vivants » orientxxi.info/dossiers-et-...
Reposted by Flo.
Lol the Nobels can't even acknowledge women's contribution to discovery. But sure let's acknowledge The Machines.
October 9, 2025 at 7:00 PM
Lol the Nobels can't even acknowledge women's contribution to discovery. But sure let's acknowledge The Machines.
Reposted by Flo.
I've been arguing that #NeuroAI should model the brain in health *and* in disease -- very excited to share a first step from Melika Honarmand: inducing dyslexia in vision-language-models via targeted perturbations of visual-word-form units (analogous to human VWFA) 🧠🤖🧪 arxiv.org/abs/2509.24597
October 2, 2025 at 12:10 PM
I've been arguing that #NeuroAI should model the brain in health *and* in disease -- very excited to share a first step from Melika Honarmand: inducing dyslexia in vision-language-models via targeted perturbations of visual-word-form units (analogous to human VWFA) 🧠🤖🧪 arxiv.org/abs/2509.24597
Reposted by Flo.
5/
Les auteurs concluent que Bluesky pourrait faciliter non seulement une visibilité précoce des résultats de recherche, mais aussi un dialogue scientifique plus constructif, dans un paysage des médias sociaux en pleine évolution.
L'article entier est disponible ici : arxiv.org/abs/2507.18840
Les auteurs concluent que Bluesky pourrait faciliter non seulement une visibilité précoce des résultats de recherche, mais aussi un dialogue scientifique plus constructif, dans un paysage des médias sociaux en pleine évolution.
L'article entier est disponible ici : arxiv.org/abs/2507.18840
How is science discussed on Bluesky?
Amid the migration of academics from X, the social media platform Bluesky has emerged as a potential alternative. To assess its viability and relevance for science communication, this study presents t...
arxiv.org
September 24, 2025 at 1:54 PM
5/
Les auteurs concluent que Bluesky pourrait faciliter non seulement une visibilité précoce des résultats de recherche, mais aussi un dialogue scientifique plus constructif, dans un paysage des médias sociaux en pleine évolution.
L'article entier est disponible ici : arxiv.org/abs/2507.18840
Les auteurs concluent que Bluesky pourrait faciliter non seulement une visibilité précoce des résultats de recherche, mais aussi un dialogue scientifique plus constructif, dans un paysage des médias sociaux en pleine évolution.
L'article entier est disponible ici : arxiv.org/abs/2507.18840
Reposted by Flo.
Here is a ‘blame-praise curve’ for the percentage of household chores done by husbands
People think husbands are blameworthy if they do 20%, but praiseworthy if they do 40% (!)
Perceived average is above 40%, so people think below average is still praiseworthy
[Thread]
People think husbands are blameworthy if they do 20%, but praiseworthy if they do 40% (!)
Perceived average is above 40%, so people think below average is still praiseworthy
[Thread]
August 22, 2025 at 5:53 PM
Here is a ‘blame-praise curve’ for the percentage of household chores done by husbands
People think husbands are blameworthy if they do 20%, but praiseworthy if they do 40% (!)
Perceived average is above 40%, so people think below average is still praiseworthy
[Thread]
People think husbands are blameworthy if they do 20%, but praiseworthy if they do 40% (!)
Perceived average is above 40%, so people think below average is still praiseworthy
[Thread]
Reposted by Flo.
In our new paper, we explore how we can build encoding models that are both powerful and understandable. Our model uses an LLM to answer 35 questions about a sentence's content. The answers linearly contribute to our prediction of how the brain will respond to that sentence. 1/6
August 18, 2025 at 9:42 AM
In our new paper, we explore how we can build encoding models that are both powerful and understandable. Our model uses an LLM to answer 35 questions about a sentence's content. The answers linearly contribute to our prediction of how the brain will respond to that sentence. 1/6
Reposted by Flo.
Vous avez sûrement déjà lu que la climatisation aggrave les canicules car l’air chaud rejeté à l’extérieur augmenterait la température des villes de 2 à 3°C.
Vérifions ce que disent vraiment les études scientifiques à ce sujet 🧑🔬
(lisez jusqu’au bout, car vous allez être 🤯🤯 à la fin)
#Thread
⬇️⬇️
Vérifions ce que disent vraiment les études scientifiques à ce sujet 🧑🔬
(lisez jusqu’au bout, car vous allez être 🤯🤯 à la fin)
#Thread
⬇️⬇️
August 15, 2025 at 7:37 PM
Vous avez sûrement déjà lu que la climatisation aggrave les canicules car l’air chaud rejeté à l’extérieur augmenterait la température des villes de 2 à 3°C.
Vérifions ce que disent vraiment les études scientifiques à ce sujet 🧑🔬
(lisez jusqu’au bout, car vous allez être 🤯🤯 à la fin)
#Thread
⬇️⬇️
Vérifions ce que disent vraiment les études scientifiques à ce sujet 🧑🔬
(lisez jusqu’au bout, car vous allez être 🤯🤯 à la fin)
#Thread
⬇️⬇️
Reposted by Flo.
🔵 Proud to share our new preprint 🔵
We compared humans and deep neural networks on sound localization 👂📍
Humans robustly localized OOD sounds even without primary interaural cues (ITD & ILD)
Models localized well only in-training distribution sounds, failing on OOD regime
Link & full story 🧵👇
We compared humans and deep neural networks on sound localization 👂📍
Humans robustly localized OOD sounds even without primary interaural cues (ITD & ILD)
Models localized well only in-training distribution sounds, failing on OOD regime
Link & full story 🧵👇
August 9, 2025 at 11:04 AM
🔵 Proud to share our new preprint 🔵
We compared humans and deep neural networks on sound localization 👂📍
Humans robustly localized OOD sounds even without primary interaural cues (ITD & ILD)
Models localized well only in-training distribution sounds, failing on OOD regime
Link & full story 🧵👇
We compared humans and deep neural networks on sound localization 👂📍
Humans robustly localized OOD sounds even without primary interaural cues (ITD & ILD)
Models localized well only in-training distribution sounds, failing on OOD regime
Link & full story 🧵👇
Reposted by Flo.
🚨 Finally out in Nature Machine Intelligence!!
"Visual representations in the human brain are aligned with large language models"
🔗 www.nature.com/articles/s42...
"Visual representations in the human brain are aligned with large language models"
🔗 www.nature.com/articles/s42...
High-level visual representations in the human brain are aligned with large language models - Nature Machine Intelligence
Doerig, Kietzmann and colleagues show that the brain’s response to visual scenes can be modelled using language-based AI representations. By linking brain activity to caption-based embeddings from lar...
www.nature.com
August 7, 2025 at 1:06 PM
🚨 Finally out in Nature Machine Intelligence!!
"Visual representations in the human brain are aligned with large language models"
🔗 www.nature.com/articles/s42...
"Visual representations in the human brain are aligned with large language models"
🔗 www.nature.com/articles/s42...
Reposted by Flo.
Helsinki has recorded its first year without a single traffic-related casualty, with the last death occurring in early July 2024.
Today, over half of the city’s streets have a speed limit of 30 km/h.
Today, over half of the city’s streets have a speed limit of 30 km/h.
Helsinki leads by example with zero road deaths in one year
The Finnish capital has recorded its first year without a single traffic-related casualty, with the last death occurring in early July 2024. This remarkable achievement for […]
traveltomorrow.com
August 5, 2025 at 4:37 PM
Helsinki has recorded its first year without a single traffic-related casualty, with the last death occurring in early July 2024.
Today, over half of the city’s streets have a speed limit of 30 km/h.
Today, over half of the city’s streets have a speed limit of 30 km/h.
Reposted by Flo.
In neuroscience, we often try to understand systems by analyzing their representations — using tools like regression or RSA. But are these analyses biased towards discovering a subset of what a system represents? If you're interested in this question, check out our new commentary! Thread:
![What do representations tell us about a system? Image of a mouse with a scope showing a vector of activity patterns, and a neural network with a vector of unit activity patterns
Common analyses of neural representations: Encoding models (relating activity to task features) drawing of an arrow from a trace saying [on_____on____] to a neuron and spike train. Comparing models via neural predictivity: comparing two neural networks by their R^2 to mouse brain activity. RSA: assessing brain-brain or model-brain correspondence using representational dissimilarity matrices](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:e6ewzleebkdi2y2bxhjxoknt/bafkreiav2io2ska33o4kizf57co5bboqyyfdpnozo2gxsicrfr5l7qzjcq@jpeg)
August 5, 2025 at 2:36 PM
In neuroscience, we often try to understand systems by analyzing their representations — using tools like regression or RSA. But are these analyses biased towards discovering a subset of what a system represents? If you're interested in this question, check out our new commentary! Thread: