Net movement of molecules or atoms from a region of high concentration (or high chemical potential) to a region of low concentration (or low chemical potential)
POPULARITY
Categories
(13) Jack Burnham discusses how Nvidia chips reach the Chinese military through loopholes in export controls and subsidiaries. He notes bureaucratic confusion over the "AI diffusion rule" allowed Chinese firms to stockpile high-end hardware. Burnham recommends stricter Commerce Department guidance to prevent further military modernization.
Cette semaine, Simon nous parle du coffret Blu-ray « Disco Italiana » de Severin Films et de ses visionnements des films de la série Emmanuelle. Pour sa part, Antoine aborde le director's cut de « The X-Files: I Want to Believe », une histoire de Russes ayant usurpé l'identité d'Angine de Poitrine, ainsi que le spectacle de lancement d'ICI Musique Rock aux Foufounes Électriques. De son côté, Benoit revient sur les récents mouvements chez Marvel Comics, qui pourraient avoir un impact sur le fonctionnement de l'éditeur, sur l'adaptation cinématographique des « Livres dont vous êtes le héros » par le collectif Radio Silence, ainsi que sur les bandes dessinées de Derek Laufman qu'il s'est procurées lors de la plus récente édition du Festival BD de Montréal. En dernière partie d'émission, nous discutons de « The Punisher: One Last Kill », un film de Reinaldo Marcus Green disponible sur la plateforme Disney+ mettant en vedette Jon Bernthal et Deborah Ann Woll. Consulter l'ensemble de nos archives: https://www.mysterieuxetonnants.com/category/podcasts/emission/ Laissez-nous un message vocal: https://www.speakpipe.com/mysterieuxe Devenez membre de la communauté Patreon: https://www.patreon.com/MysterieuxE Diffusion originale : 25 mai 2026 Site web : MysterieuxEtonnants.com © Les Mystérieux Étonnants. Tous droits réservés
Stewart Alsop interviews Nizar, CEO of Pixel Robotics, on the Crazy Wisdom Podcast to explore the intersection of AI, robotics, and perception. The conversation covers a wide range of technical topics including how transformers enable multimodal representation across text, images, and voice, the role of world models in predicting physical interactions, the advantages of diffusion models over traditional LLMs for certain applications, and the challenges of achieving real-time processing for robotics applications. Nizar explains Pixel Robotics' work on creating accurate 3D meshes from smartphone cameras for companies like L'Oréal, moving away from specialized sensors to make the technology more accessible through sophisticated algorithms, and discusses the future of robotics as closing the perception-action loop to enable robots to perform real tasks beyond simple demonstrations. To find out more visit Pixel Robotics' website.Timestamps00:00 Stewart welcomes Nizar, CEO of Pixel Robotics, discussing what a pixel is as the smallest visual unit on screens composed of red green and blue colors05:00 Discussion of perception systems and how logarithmic laws help compress signals in both human and artificial systems, exploring normalization layers and sigmoid functions in deep learning10:00 Exploring how transformers unified different data modalities including text voice and images, creating common representations through methods like contrastive learning15:00 Nizar explains transformers as brute force learning systems with room for improvement through focused attention mechanisms and knowledge graphs rather than processing everything20:00 Conversation about loss functions local minima versus global minima and how mixture of experts uses specialized small models instead of one massive generalist network25:00 Discussion of deterministic versus probabilistic systems and how explicitly defined task graphs often outperform orchestrator-based approaches in AI systems30:00 Exploring world models as predictive physics-based systems that learn environmental flows and transformations, complementing rather than replacing language models35:00 Nizar discusses real-time processing challenges for robotics requiring millisecond responses with small memory footprints using vision transformers for faster experimentation40:00 Pixel's work creating three d meshes from smartphone cameras for companies like L'Oreal, moving away from specialized sensors toward accessible software-based solutions45:00 Explanation of different three d representations including voxels point clouds and meshes, with meshes being optimal for manipulation and rendering in applications50:00 Future direction involves closing perception-action loops in robotics, moving beyond dancing toy robots toward practical multimodal systems that perform real tasks55:00 Pixel's goal is democratizing high-quality three d scanning through smartphones, making mesh creation accessible to unlock applications in gaming cinema and virtual showroomsKey Insights1. Pixel Robotics derives its name from combining perception and action in robotics, where the pixel represents the digital perception component and robotics represents the physical action component. The pixel serves as a metaphor for how robots must quantize and digitize continuous analog information from the real world into discrete units that computer systems can process, similar to how pixels are the fundamental building blocks of images on a screen. This quantization process is essential because numerical systems cannot work with truly continuous data and must convert reality into tractable digital representations that algorithms can manipulate.2. The transformer architecture has created a fundamental unification in how different types of data can be represented and processed across multiple modalities. Before transformers, researchers working on natural language processing, computer vision, and audio analysis used completely different approaches and methodologies. The breakthrough of transformers was establishing a common representational framework that could handle text, images, voice, and other data types using similar underlying mechanisms. This unification is what enabled the development of truly multimodal AI systems and represents one of the most significant advances beyond just the language modeling capabilities that initially gained public attention.3. Current transformer-based systems represent a brute force approach to learning that will likely be superseded or enhanced by more efficient algorithms. Despite claims that we have exhausted internet text data for training, significant improvements continue to emerge every few months through algorithmic innovations rather than simply adding more data. Future developments will likely involve more specialized attention mechanisms that focus on relevant information rather than correlating everything with everything, mixture of experts architectures with small specialized models, and approaches inspired by biological systems such as logarithmic compression laws and event-based processing that humans use naturally.4. Diffusion-based language models represent a promising alternative to standard next-token prediction that could produce more accurate outputs through an iterative refinement process. Unlike traditional language models that predict one token at a time and cannot revise earlier outputs, diffusion models treat text generation like image denoising, starting with a noisy representation and progressively refining the entire output across multiple steps. This holistic approach allows the model to reconsider and improve all parts of the response simultaneously, potentially leading to higher quality results, though it may be slower than current autoregressive methods. This represents an important direction for overcoming fundamental limitations in how language models currently generate text.5. For robotics applications, real-time performance and small model size are critical constraints that differ significantly from the requirements of large language models deployed in data centers. Vision transformers are being used as a testbed for developing efficient real-time algorithms because they require far fewer computational resources to train and test compared to large language models, making them more practical for rapid experimentation. The goal is to achieve millisecond-level response times with minimal memory footprint so that robots can react quickly to dynamic environments and run on affordable hardware that can be embedded in actual robotic systems rather than requiring expensive server infrastructure.6. Practical robotics implementation requires moving beyond specialized sensors to software solutions that work with ubiquitous devices like smartphones for tasks such as three-dimensional reconstruction. Pixel Robotics evolved from building specialized scanning hardware to focusing on algorithms that can generate high-quality mesh representations of environments using only smartphone cameras, making the technology far more accessible and practical for real-world deployment. This approach enables applications ranging from industrial robotic arm control to virtual showrooms, and more importantly, it allows anyone to capture three-dimensional data without expensive equipment, which can also help generate larger training datasets for future AI development.7. The next frontier in AI and robotics is closing the perception-action loop to enable robots to perform real practical tasks rather than remaining as demonstration systems or toys. While significant progress has been made in cognitive capabilities through language models and in robotic mobility through mechanical engineering advances, the critical challenge is integrating perception with action through systems like Vision-Language-Action models. The fundamental starting point for learning this integration is simple perception-action exercises, such as programming a camera mounted on servo motors to track and center a colored object, which demonstrates the basic principle of using sensory input to drive physical response that underlies all more sophisticated robotic behaviors.
From 2000: Listen to Adam Mark speak with Candice Webb about the Fijian Crested Iguana, The Australian Science and Technology Heritage Centre, and a Panel discussion about the Federal government attack on the ABC. Hosted by Ian Woolf Produced by Chris Brown and Deanna Coleman Panelled by anonymous Support Diffusion by making a contribution Support Diffusion by buying venus flytrap shirts
durée : 00:54:45 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:27:17 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
What happens when hidden truths slowly spread beneath the surface—unnoticed until they begin to change everything? In this intriguing and thought-provoking episode, Linda J. Brown explores In Secret Diffusion, a story and concept that delves into secrecy, influence, and the subtle ways ideas and events can shape lives from behind the scenes. Drawing from themes of mystery, psychology, and human behavior, Linda discusses how hidden information, personal experiences, and quiet influences can ripple outward in unexpected ways. She explores the tension between what is seen and unseen, known and unknown, and how secrecy can impact both individuals and society. This episode invites listeners to reflect on the nature of hidden forces and personal perception. How do concealed truths influence relationships and decision-making? Why are people drawn to mysteries and secrets? And what happens when information that has long remained hidden finally begins to emerge? Join us for a compelling and atmospheric conversation that journeys into the shadows of human experience—where secrecy meets revelation, and where even the quietest influence can create lasting change.Become a supporter of this podcast: https://www.spreaker.com/podcast/the-x-zone-radio-tv-show--1078348/support.Please note that all XZBN radio and/or television shows are Copyright © REL-MAR McConnell Meda Company, Niagara, Ontario, Canada – www.rel-mar.com. For more Episodes of this show and all shows produced, broadcasted and syndicated from REL-MAR McConell Media Company and The 'X' Zone Broadcast Network and the 'X' Zone TV Channell, visit www.xzbn.net. For programming, distribution, and syndication inquiries, email programming@xzbn.net.We are proud to announce the we have launched TWATNews.com, launched in August 2025.TWATNews.com is an independent online news platform dedicated to uncovering the truth about Donald Trump and his ongoing influence in politics, business, and society. Unlike mainstream outlets that often sanitize, soften, or ignore stories that challenge Trump and his allies, TWATNews digs deeper to deliver hard-hitting articles, investigative features, and sharp commentary that mainstream media won't touch.These are stories and articles that you will not read anywhere else.Our mission is simple: to expose corruption, lies, and authoritarian tendencies while giving voice to the perspectives and evidence that are often marginalized or buried by corporate-controlled media
Avec : Tristane Banon, journaliste et essayiste. Benjamin Amar, prof d'histoire-géo. Et Juliette Briens, journaliste à L'Incorrect. - Accompagnée de Charles Magnien et sa bande, Estelle Denis s'invite à la table des français pour traiter des sujets qui font leur quotidien. Société, conso, actualité, débats, coup de gueule, coups de cœurs… En simultané sur RMC Story.
Mardi politique reçoit Laetitia Saint-Paul, députée Horizons & Indépendants de Maine-et-Loire. Laetitia Saint-Paul est interviewée par : - Frédéric Rivière (RFI) - Roselyne Febvre (France 24) Diffusions : 18h10-18h30 sur France 24 21h10-21h30 sur RFI
Cette semaine, pour célébrer le 100e épisode du podcast, ce sont nos plus vaillants supporters les Bonnes Parsonnes qui prennent le contrôle du contenu! Au menu, des tops des plus flyés, y compris une liste de mon meilleur jeu pour chaque année de ma vie! En impressions, on discute de Rocket Knight Adventure Re-Sparked, une compilation des meilleures quêtes de Sparkster, le chevalier opposum et produite par Limited Run Games. Est-ce que ça vaut la peine de retourner dans cet univers 16-bits de Konami? Puis finalement en questions, les Bonnes Parsonnes veulent tout savoir sur l'évolution des manettes, mes meilleurs aniversaires, les chiens, technologie olfactive, Game Gear et bien plus encore! Bonne écoute et bon 100e épisode! Diffusion originale: le 18 mai 2026► Pour nous supporter et accéder à du contenu exclusif: https://ko-fi.com/fredgemus ► Vous pouvez aussi nous suivre et laisser une note positive: C'est gratuit et le meilleur encouragement! ► Écoutez nous en direct au https://www.twitch.tv/elgemusio ► Pour suivre tous mes projets et m'encourager, visitez le https://fredgemus.com/ ► Rejoignez notre communauté Le HUB sur Discord: https://discord.gg/qPDEaYqzXt ► Musique par Rubis Desrenards. Découvrez le ici: linktr.ee/rudyberhnard► Montage et diffusion en audio: Fanie Grégoire au www.fanie.ca
durée : 00:18:45 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:58:47 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
From 2002: Tim Baynes interviews Dr Virgina Shepherd about Fungi. Gina Sartore reports on research on methamphetamines and the immune system of cats, and the effect on the researcher, and then opens the microphone to discussion on the subject with Tim and Chris. Presented by Chris Stewart. Produced by Tim Baynes, and panelled by Gina Sartore. Edited by Ian Woolf Support Diffusion by making a contribution Support Diffusion by buying venus flytrap shirts
durée : 02:04:54 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
Je me suis rendue au Mans pour rencontrer Willy Mangin, entraîneur sportif et diététicien. Ayant moi-même été sportive de niveau national en judo et en roller acrobatique dans mes jeunes années, et tentant désormais de me maintenir en forme de manière plus qu'amatrice, j'avais beaucoup de questions à lui poser.Est-ce que les arts martiaux c'est du pipeau irlandais à la sauce orientale ? Sont-ils vraiment utiles pour se défendre en cas d'agression ? Peut-on pratiquer sérieusement le tai-chi sans croire au ch'i ?Michael Phelps a-t-il gagné les 200 m papillon aux Jeux olympiques de Rio grâce à l'application de ventouses sur son dos ? Novak Djokovic a-t-il raison de croire aux énergies ? Pourquoi Zinedine Zidane met-il toujours sa chaussette gauche en premier ? Je voulais aussi savoir : comment faire pour améliorer la performance ? Les compléments alimentaires (créatine, protéine, magnésium ou oméga-3) sont-ils utiles ? Peut-on être végétarien et sportif accompli ? Avoir des courbatures est-il un passage obligé ?Être accroché à ses statistiques de VO2 max est-il utile, ou même souhaitable ?Et puis : quels sont les bénéfices réels de la pratique sportive et quand peut-elle devenir néfaste ? Quels sont les secrets d'une bonne préparation mentale ? Comment enrayer les discriminations dans les milieux sportifs ? Quelle est la différence entre un bon entraîneur et un mauvais entraîneur ?Bref, j'ai occupé le terrain, et comme Willy Mangin est bavard, très bavard, on a papoté et bu du thé dans son salon pendant une journée entière. Me voilà donc de retour avec une série en 6 chapitres. Diffusion du chapitre 1, vendredi prochain à 18 h !•• SOUTENIR ••Méta de Choc est gratuit, indépendant et sans publicité. Vous pouvez vous aussi le soutenir en faisant un don ponctuel ou mensuel : https://soutenir.metadechoc.fr/.•• SUIVRE ••Facebook, Instagram, LinkedIn, TikTok, Bluesky, Mastodon, Threads, PeerTube, YouTube. Hébergé par Acast. Visitez acast.com/privacy pour plus d'informations.
From 2002: Presented by Tim Baynes. News by Tim. Gina Sartore reports from the Journal of Psychosomatic Medicine, about how life events affect health. Lachlan Whatmore gives a guided tour of the insides and outsides of sea sponges. Produced by Lachlan Whatmore with technical support by Gina Sartore. (Music originally broadcast in the show, removed for copyright reasons)
Jeudi 14 mai 1998 – Diffusion du dernier épisode de la serie culte new yorkaise : SeinfieldLe jeudi 14 mai 1998, plus de 76 millions d'américains sont devant leur télévision sur la chaîne NBC à l'occasion de la diffusion du tout dernier épisode d'une sitcom devenue culte : Seinfeld.Après neuf saisons et 180 épisodes, cette date marque la fin d'une série profondément new-yorkaise, véritable symbole, aux côtés de Friends, des années 90.Cet éphéméride est donc l'occasion de revenir sur ce fameux “show about nothing” — une « série qui ne parle de rien », comme ses propres personnages aiment la définir.N'hésitez pas à aller visiter notre site racontemoinewyork.com Retrouvez tous les liens des réseaux sociaux et des plateformes du podcast ici : https://linktr.ee/racontemoinewyorkHébergé par Ausha. Visitez ausha.co/politique-de-confidentialite pour plus d'informations.
Send us Fan MailIn this episode of Sidecar Sync, Amith Nagarajan and Mallory Mejias dig into two major shifts happening beneath the surface of AI: how enterprise software vendors are responding to the rise of AI agents, and why diffusion language models may be moving from research curiosity to real-world infrastructure faster than expected. They unpack Salesforce's open, agent-friendly “Headless 360” strategy, SAP's more restrictive API stance, and what these moves mean for associations trying to maintain control over their data. Then, they revisit diffusion LLMs through the lens of Inception Labs' Mercury 2, exploring why faster, cheaper models could matter for voice agents, enterprise search, taxonomy work, content classification, and the future of model flexibility.
durée : 00:59:30 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
Après l'exploit technologique des Jeux olympiques de Jeux olympiques de Paris 2024, diffusés en 4K HDR avec son immersif, on aurait pu penser que la télévision publique française poursuivrait sur cette lancée. Et pourtant, le signal envoyé aujourd'hui par France Télévisions est tout autre : un retour partiel en arrière sur la 4K native. Un choix qui peut surprendre… mais qui s'explique. Car derrière cette décision, il ne s'agit pas d'un recul technologique, mais d'un arbitrage économique. Depuis plusieurs mois, le groupe évolue sous forte contrainte budgétaire. Une commission d'enquête parlementaire lancée fin 2025, combinée aux alertes répétées de la Cour des comptes, a mis en lumière une situation financière tendue. Résultat : un plan d'économies d'environ 140 millions d'euros a été acté pour 2026.Dans ce contexte, certaines dépenses deviennent difficiles à défendre. La production en 4K native en fait partie. Produire une image en Ultra Haute Définition implique une chaîne technique plus lourde : caméras spécifiques, traitement des images plus complexe, besoins accrus en bande passante pour la diffusion… Autant de coûts supplémentaires, pour un bénéfice visuel qui n'est pas toujours évident pour tous les téléspectateurs.C'est là tout le paradoxe. La 4K promet une qualité d'image supérieure, mais son impact dépend fortement des conditions de visionnage : taille de l'écran, distance, qualité du signal. Sur un événement comme Roland-Garros, dominé par des plans larges, la différence avec une image bien optimisée peut rester discrète. La solution envisagée repose donc sur un compromis technique. Plutôt que filmer directement en 4K, France Télévisions pourrait capter en Full HD, puis utiliser un procédé appelé upscaling. Il s'agit d'un traitement algorithmique qui “reconstruit” une image en Ultra HD à partir d'une source plus basse définition, avec l'aide du HDR, une technologie qui améliore les contrastes et les couleurs.Résultat : une qualité visuelle jugée satisfaisante pour le grand public, à un coût bien inférieur. Ce choix marque un tournant. La 4K native ne disparaît pas, mais elle pourrait devenir l'exception, réservée aux grands événements, plutôt qu'un standard systématique. Un virage pragmatique, dicté par la réalité des finances publiques. Hébergé par Acast. Visitez acast.com/privacy pour plus d'informations.
durée : 01:25:55 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Rafik Zénine, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
Listen to the science of Andes hantavirus, how it spreads, and why its important not to confuse it with other hantaviruses.In Budget news, CSIRO gets more money but cuts more scientists and research. Produced and hosted by Ian Woolf Support Diffusion by making a contribution CC BY 4.0 via Wikimedia Commons Support Diffusion by buying venus flytrap shirts
durée : 00:57:53 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:16:41 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
Today on Scope Conditions: when the bombs don't go off, the war isn't over.We tend to think of peace as beginning when the bombs stop falling. But as our guest today shows us, this is only half the story. Over the course of the Vietnam War, the United States engaged in massive bombing in Cambodia. Between 1965 and 1973, the U.S. dropped 500,000 tons of explosives there — more than the combined weight of every man, woman, and child in the country. Dr. Erin Lin, an Associate Professor of Political Science at the Ohio State University, set out to understand the continued impacts of this cataclysmic bombing campaign on Cambodian society. A landmark 2011 study had given us a partial answer: it had concluded that US bombing had no measurable long-term effects on economic outcomes in Southeast Asia. For years, that finding set the terms of the debate.In her award-winning book, When the Bombs Stopped: The Legacy of War in Rural Cambodia, published by Princeton University Press, Erin pushes back. She argues that those analyses were looking at the wrong level — that district-level aggregates conceal devastating effects on individual households and farms. More than that, they were looking at only half the intervention. It's the bombs that didn't detonate — an estimated 26 million cluster munitions still embedded in the soil — that are shaping life today in rural Cambodia.Erin spent years farming alongside families, combing through declassified military records, and building some of the most granular data ever assembled on the American bombing campaign. Her creative multi-method research design allows her to trace the dramatic long-term consequences of unexploded ordinance for the economic livelihood of Cambodian farmers.We talk with Erin about the many ironies laced through her findings: that cluster munitions are most likely to fail in soft, fertile soil, meaning Cambodia's most agriculturally valuable land is also its most contaminated; that bomb contamination can paradoxically shield farmers from predatory land seizures by political elites; and that unexploded ordnance, rather than forging solidarity among those living with it, tends to deepen ethnic divisions within villages.We hope you learn from this conversation. To stay informed about future episodes, follow us on X and Bluesky @scopeconditions and check out our website, scopeconditionspodcast.com, where you can also find references to all the academic works we discuss. And if you like the show, please rate and review us on Apple Podcasts or Spotify.We note that we recorded this interview before the recent US-Israeli war with Iran. Now, here's our conversation with Erin Lin.Works cited in this episodeBiddle, Steven. 2004. Military Power: Explaining Victory and Defeat in Modern Battle. Princeton University Press.Brooks, Rosa. 2014. “Cross-Border Targeted Killings: ‘Lawful but Awful'?” Harvard Journal of Law and Public Policy 38:233–50.________. 2014. “Drones and the International Rule of Law.” Ethics & International Affairs 28(1):83–103. ________. 2016. How Everything Became War and the Military Became Everything: Tales from the Pentagon. Simon and Schuster.Horowitz, Michael C. 2010. The Diffusion of Military Power. Princeton University Press.Lyall, Jason, and Isaiah Wilson. 2009. “Rage against the Machines: Explaining Outcomes in Counterinsurgency Wars.” International Organization 63(1):67–106.Reiter, Dan, and Allan C. Stam. 2010. Democracies at War. Princeton University Press.Pape, Robert A. 2014. Bombing to Win: Air Power and Coercion in War. Cornell University Press.Schelling, Thomas. 2008. Arms and Influence. Yale University Press.Sheehan, Neil. 1971. “Should We Have War Crime Trials?” New York Times Book Review.
durée : 01:29:52 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
From 2007 in memory of Charles Willock and Darren Osborne: Identity Card Special Edition Presented by Darren Osborne, Synthetic interview with Anna Johnston about privacy concerns, Aras Vaichas speaks with Ian Woolf about RFID, Interview with Professor Graham Greenleaf about Access and Cyberlaw Produced by Ian Woolf and Charles Willock
Welcome back to Scaling Theory. In this episode, I speak with Matthew O. Jackson, the William D. Eberle Professor of Economics at Stanford University and an external faculty member at the Santa Fe Institute. Matthew is one of the founders of the modern economics of networks and the author of The Human Network and Social and Economic Networks.We talk about the friendship paradox, why homophily slows how fast a society learns the truth but helps niche ideas catch fire, and the gossip study where villagers in southern India proved remarkably good at naming the most central spreaders in their community. We then turn to AI agents as a different species: Turing tests on LLMs, the steerability of agent personas through system prompts, and what to make of Moltbook, the social network for AI agents.By the end, you will know why telling students how much their peers actually drink reduces binge drinking more than warning them about the dangers of alcohol, why the same network can spread a virus quickly and a belief slowly, and why AI agents change their behavior when asked to explain it.Papers and works referenced in the conversationBooksThe Human Network: How Your Social Position Determines Your Power, Beliefs, and Behaviors — Matthew O. Jackson (Pantheon, 2019). https://web.stanford.edu/~jacksonm/books.htmlSocial and Economic Networks — Matthew O. Jackson (Princeton University Press, 2008). https://web.stanford.edu/~jacksonm/books.htmlPart I — The scaling of human networks"Diffusion and Contagion in Networks with Heterogeneous Agents and Homophily" — Matthew O. Jackson and Dunia López-Pintado, Network Science 1(1), 2013. https://arxiv.org/abs/1111.0073"How Homophily Affects the Speed of Learning and Best-Response Dynamics" — Benjamin Golub and Matthew O. Jackson, Quarterly Journal of Economics 127(3), 2012. https://web.stanford.edu/~jacksonm/homophily.pdf"Using Gossips to Spread Information: Theory and Evidence from Two Randomized Controlled Trials" — Abhijit Banerjee, Arun G. Chandrasekhar, Esther Duflo, and Matthew O. Jackson, Review of Economic Studies 86(6), 2019. https://academic.oup.com/restud/article/86/6/2453/5345571"Empathy and Well-Being Correlate with Centrality in Different Social Networks" — Sylvia A. Morelli, Desmond C. Ong, Rucha Makati, Matthew O. Jackson, and Jamil Zaki, PNAS 114(37), 2017. https://www.pnas.org/doi/10.1073/pnas.1702155114Part II — The scaling of AI agents"Inequality's Economic and Social Roots: The Role of Social Networks and Homophily" — Matthew O. Jackson, in Advances in Economics and Econometrics: Twelfth World Congress of the Econometric Society (Cambridge University Press, 2025). https://arxiv.org/abs/2506.13016"AI Behavioral Science" — Jackson, Mei, Wang, Xie, Yuan, Benzell, Brynjolfsson, Camerer, Evans, Jabarian, Kleinberg, Meng, Mullainathan, Ozdaglar, Pfeiffer, Tennenholtz, Willer, Yang, and Ye, arXiv 2509.13323, 2025. https://arxiv.org/abs/2509.13323"A Turing Test of Whether AI Chatbots Are Behaviorally Similar to Humans" — Qiaozhu Mei, Yutong Xie, Walter Yuan, and Matthew O. Jackson, PNAS 121(9), 2024. https://www.pnas.org/doi/10.1073/pnas.2313925121
durée : 00:41:09 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:31:36 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:29:49 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
Listen to news of self-watering fungi, and a fruit fly brain copied into a computer, hooked up to a pretend body, living in a simulated world. Produced and hosted by Ian Woolf Support Diffusion by making a contribution Support Diffusion by buying venus flytrap shirts
durée : 03:00:01 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:54:08 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:46:32 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:44:39 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
From 2008: Listen to Nanotechnologist Michael Ford explain how to make matter do whatever you wish, Marc West goes on a pub crawl of science-related pubs in London, News of a PhD theses interprative dances competition, and make your brain grow. Presented by Nick Evershed, Produced by Ian Woolf. Support Diffusion by making a contribution Support Diffusion by buying venus flytrap shirts
Indie Lee launched her namesake beauty brand — a pioneer of the "clean" beauty movement — in 2010. Now owned by parent company American Exchange, the brand is embarking on a new chapter: a diffusion line, Indie Lee Botanicals, which launched at Whole Foods in February. The range launched with a tight edit including a cleanser, toning mist, serum and moisturizer, each priced $20-$25. On this week's episode of The Glossy Beauty Podcast, Lee joins co-host Sara Spruch-Feiner to explore why now was the right time to launch a lower-priced line, how she approached maintaining efficacy while cutting costs, and how this diffusion brand will grow and shape her core collection's future. The conversation also dives into how shifting consumer behavior, whether driven by economic pressure or interest in ingredient safety, is reshaping how and where people shop for beauty products. Read Glossy Beauty's coverage: Why ‘The Devil Wears Prada 2' is the collaborator fashion and beauty brands have been waiting for The beauty industry welcomes a flood of new peptide products as ‘peptide therapy' trends online
durée : 00:29:55 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:59:34 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 01:00:02 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 00:56:10 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
durée : 01:36:41 - Les Nuits de France Culture - par : Albane Penaranda - réalisation : Mathias Le Gargasson, Antoine Dhulster, Rafik Zénine, Vincent Abouchar, Emily Vallat, Hassane M'Béchour, INA Vous aimez ce podcast ? Pour écouter tous les épisodes sans limite, rendez-vous sur Radio France
Listen to news of 3D printed brain interfaces, Neural computers, brain-computer-interface speech, AI gets its own computer, and night-time solar power. Hosted and produced by Ian Woolf Support Diffusion by making a contribution Support Diffusion by buying venus flytrap shirts
durée : 00:52:31 - Les Nuits de France Culture - par : Philippe Garbit - Par Anaïs Kien - Avec Philippe Cousin, Jacques Bernard, François Leroux et David Cueco - réalisation : Virginie Mourthé - invités : François Leroux Baryton
durée : 00:54:50 - Les Nuits de France Culture - par : Antoine Dhulster - Par Roland Auguet - Avec Gilbert Rouget (ethnomusicologue français, directeur de recherche honoraire au CNRS) - Réalisation Christiane Mallarmé - réalisation : Virginie Mourthé
From 2008: Victoria Bond finds out if the Large Hadron Collider destroyed the world, Ian Woolf talks to Professor Mike Cortie about zapping gold nanoparticles to cure mind-control parasites, News by Patrick Rubie: King Tut's foetuses, solar power at night, smart bricks, dandelion rubber Hosted and produced by Ian Woolf Support Diffusion by making a contribution Support Diffusion by buying venus flytrap shirts
durée : 00:44:28 - Les Nuits de France Culture - par : Geneviève Huttin - -1 : Autobiogre km 75, -2 : Les relations des écrivains et des dictionnaires, 3 : Phanées les nuées (1ère diffusion : 01/03/1980, 28/03/1981 et 23/01/1982) Par Gérard-Julien Salvy - Avec Hubert Lucot (écrivain et rédacteur en chef des dictionnaires Hachette) - Réalisation André Mathieu - réalisation : David Jacubowiez
Free Mini course here for https://youtu.be/hzROwnG2BXs Gentle Night Soothing lauren-ostrowski-fenton-s-school.teachable.com/p/gentle-night-soothing-deep-sleep-meditation-reflection-pack Calm racing thoughts and soften your body before bed with my free Gentle Night Soothing mini-course (audio + 2-page PDF worksheets). Download it here: lauren-ostrowski-fenton-s-school.teachable.com/p/gentle-night-soothing-deep-sleep-meditation-reflection-pack Free worksheets Free video and audio
This Week in Machine Learning & Artificial Intelligence (AI) Podcast
Today, we're joined by Stefano Ermon, associate professor at Stanford University and CEO of Inception Labs to discuss diffusion language models. We dig into how diffusion approaches—traditionally used for images—are being adapted for text and code generation, the technical challenges of applying continuous methods to discrete token spaces, and how diffusion models compare to traditional autoregressive LLMs. Stefano introduces Mercury 2, a commercial-scale diffusion LLM that can generate multiple tokens simultaneously and achieve inference speeds 5-10x faster than small frontier models, paving the way for latency-sensitive applications like voice interactions and fast agentic loops. We also cover the open research challenges in diffusion LLM training, serving infrastructure requirements, and post-training for diffusion-based systems. Finally, Stefano shares his perspective on whether diffusion models can rival or surpass autoregressive LLMs at scale, the advantages for highly controllable generation, and what the future of multimodal diffusion models might look like. The complete show notes for this episode can be found at https://twimlai.com/go/764.