De Abreu Lopes Fabien

#MardiConseil | Une urgence ? Selon la situation, vous devez vous adresser… 📲 18 ➡️ Aux sapeurs-pompiers, si cela concerne une situation de péril ou un accident 📲 15 ➡️ Au SAMU, si cela concerne une urgence médicale 📲 112 ➡️ Seul numéro à connaître lors d’une situation d’urgence, où que vous soyez en Europe 📲 114 ➡️ Pour toutes personne présentant une déficience auditive ou une incapacité de parler à haute voix #Alerter #Secours #Prévenir #numérodurgence

#ChangementHeure | ⏰ Cette nuit, nous sommes passés de 2h à 3h… 1 heure de sommeil en moins, mais une excellente occasion de penser à votre sécurité 👇 🔥 Le changement d’heure est le moment idéal pour : ✔️ Tester votre détecteur avertisseur autonome de fumée (DAAF) ✔️ Remplacer les piles si nécessaire 🚨 Un détecteur en état de marche peut sauver des vies en vous alertant dès les premières fumées. 👉 Adoptez le bon réflexe : changement d’heure = vérification du détecteur #Prévention #Sécurité #SDIS #RisqueIncendie #pompiers

@sama We understand the compute costs & pricing challenge It’s amazing to see these new features coming! It would be even cooler if Plus subscribers could also get the chance to test some of these innovations — not only Pro users. Plus members are also big OpenAI fans.

Over the next few weeks, we are launching some new compute-intensive offerings. Because of the associated costs, some features will initially only be available to Pro subscribers, and some new products will have additional fees. Our intention remains to drive the cost of intelligence down as aggressively as we can and make our services widely available, and we are confident we will get there over time. But we also want to learn what's possible when we throw a lot of compute, at today's model costs, at interesting new ideas.

Des images qui me ramènent 2️⃣0️⃣ ans en arrière… 😅 lors de cette #FISPP🚒💪🏼 dans cette #EDSPY ! Des moments gravés… profitez bien et ne galvaudez rien ! 💪🏼👊🏼

@VraserX 🙌🏻😅

Ok, us EU peasants got lucky. I have access to GPT-5 now.🥰🥰🥰

@VraserX Hope full in EU too 🤞🏻🫣

Finally, GPT-5 is dropping tomorrow 🥰

@AravSrinivas 🤞🏻

a fresh batch of comet invites just went. check your email

@VraserX Hope that, conservatively, we tell ourselves that regulation is the best way to act, but we will only know this in several years. The truth is that technologically, the EU will be behind and, in fact, dependent.

The EU AI Act: Where Innovation Slows but Trust Grows Europe has passed the world’s first comprehensive AI regulation. It promises to be a landmark for global tech policy, but it also sends a strong message: if you want to play in the EU, you play by our rules. The problem? The rules might come with brakes. Let’s start with the tough part. The EU AI Act is not light-touch regulation. It categorizes AI systems by risk level, and once you hit the high-risk threshold, you enter a bureaucratic labyrinth. Think mandatory risk assessments, post-market surveillance, logging obligations, and documentation requirements that look more like a pharmaceutical approval process than a software rollout. For small companies or open-source projects, this could mean death by compliance. And it gets worse. Foundation models, including large language models like GPT, face an extra set of rules. These include data transparency mandates and adversarial testing, along with potential fines of up to seven percent of global turnover. Yes, global. For startups dreaming of European markets, this feels less like an opportunity and more like a legal minefield. Even worse is the uncertainty. The Act grants the European Commission power to set thresholds later, meaning developers today have no idea whether tomorrow’s AI model will be labeled a “systemic risk” subject to even stricter rules. That kind of unpredictability is poison for innovation. Now let’s talk about the flip side. Because there is one, and it’s powerful. The AI Act was built around the idea of trust. It bans things that should be banned, like real-time facial recognition in public spaces or AI systems that manipulate people subliminally. It gives users the right to know when they’re talking to a machine. It aligns with Europe’s strict data protection laws, creating a clearer framework for developers who want to build responsibly. By enforcing transparency in model training, the Act addresses a core issue: data provenance. If you’re building an AI on top of stolen or biased data, you’ll have to answer for it. That’s not a bug, that’s the beginning of real accountability. It also encourages what Europe does best: long-term thinking. While American companies might chase short-term market dominance, Europe is laying the groundwork for sustainable AI ecosystems where public trust is built in, not bolted on. And it’s not all rigid. There are sandboxes, exceptions for research, and carve-outs for low-risk tools. Artists, educators, game developers, they’re largely left alone. The Act aims its firepower at systems that can cause real harm in real lives. So what does this mean? The EU AI Act will likely slow things down. But that slowdown may not be failure. It may be the pause that builds something better. Europe is betting that trust, not just speed, is the real currency of the AI age. Whether that bet pays off remains to be seen but it’s one the world is now watching.

@unitedzt @VraserX C’est vrai ….

@VraserX The problem is AI accelerating very fast and if you slow innovation you will not lead AI revolution and will be always dependent on outside sources. Also what eu ai research will choose? Mostly American labs and this is brain draining from your country. It’s not good thing.

@VraserX Did you use a VPN ?

ChatGPT Agent Mode is a game-changer. I love it when my agent is working. 🥰

🔴 #Canicule | @meteofrance place 16 départements en #VigilanceRouge à partir de demain à 12h. Un épisode exceptionnel traverse le pays et les températures atteindront jusqu’à 41°. ⚠️ Ne vous exposez pas et restez extrêmement prudents. 📲 vigilance.meteofrance.fr

Pour l'Histoire 📸

👏🏆

@AravSrinivas 🙏🏻🤞🏻

expanding comet access to more folks as we near the finish line. reply here if interested. needs to be MacBook. Preferably Apple silicon but Intel is fine too. DM your email that you use for perplexity.

[12h30] 🔥 Important incendie à Versailles, rue Champ Lagarde. ⚠️ Merci de faciliter le passage des secours et d'éviter le secteur.

Etude de la Dre Pauline Zangl et du SDIS 38 sur l’exposition aux fumées lors des formations à feu réel. Des enseignements clés pour renforcer nos protocoles (EPI, décontamination, maintenance) et protéger la santé des formateurs comme des stagiaires ! dumas.ccsd.cnrs.fr/dumas-04279507…

🚨 #Sécurité | Attention aux Arnaques ! 🚨 La saison des calendriers des @pompiers78 a commencé ! Soyez vigilants pour éviter toute tentative de fraude. 🚒🗓️ Comment être sûr ? Vérifiez l'identité des sapeurs-pompiers qui se présentent à vous en suivant les indicateurs présents en image. En cas de doute, n'hésitez pas à appeler le 18 ou le 17 ! 📞 #SécuritéCivile #Prévention #AttentionArnaques #pompiers #sapeurspompiers

#Immersion | Cette semaine, les @pompiers78 sont honorés d'accueillir 4 lieutenants de la #FILT 115 de l’@ENSOSP ! 🚒 Dans le cadre de leur Formation d'Intégration de Lieutenant (FILT), ces officiers réalisent des séquences d'immersion professionnelle en dehors de leur territoire d’origine. Et c'est au Sdis des Yvelines qu'ils ont choisi de venir se perfectionner. 🤝 Nous accueillons ces 4 officiers venant du @Pompiers_13, du @PompiersHerault, du @SDIS44 et du @SDIS54 Durant la semaine, ils découvriront, entre autre, notre groupement prévention, nos logiciels métiers, notre service de santé, ainsi que notre patrimoine avec le château de Versailles. 🏰

@RozacMarion Bravo Marion !

18:20 : Accident de la circulation en cours Buchelay @MairieBuchelay autoroute A13 PK48 sens Province Paris. carambolage - autoroute bloquée - secours en cours - ne pas géner l'intervention des secours @StephaneMillot @Prefet78 @PoliceNat78 @Gendarmerie_078 @Les_Yvelines @wazeIDF