Anticipate the regulations, adapt your strategy and your models.

Exciting update from #aiPULSE2025 in Paris today: Ampere and @Scaleway are strengthening their collaboration with an expanded portfolio of Ampere's ARM-based CPU's. These CPUs will deliver high cost-efficient compute for sustained inference operations. https://t.co/k4jF80ORe5

Avec @Xavier75 et @ylecun, l’ai-PULSE, c’est des échanges denses sur les limites des LLMs, l’intérêt des World Models, la robotique sociale et les ambitions européennes en robotique généraliste et en voice AI, un aperçu précis de ce que préparent @EnchantedTools, @UMA_Robots et

Franchement ai-pulse vraiment top Des Keynotes ultra intéressantes et très optimistes, d’autres plutôt techniques et surtout du très beau monde pour faire un peu de réseau Bien joué @joinstationf et @Scaleway

Awesome panel on world models (and beyond) at @Scaleway'ai-Pulse. Delighted to collaborate with @gen_intuition and to have @ylecun as scientific advisor!

establishing "sandbox" environments to stress-test in real time, implementing a "Kill Switch" that ensures an immediate return to classical, deterministic rules if the model drifts cognitively, ready to audit the systemic stability of your models? Let’s talk.

that's why we deploy an Operational Resilience Framework for our clients, built on three non-negotiable pillars: defining strict tolerance thresholds under operational stress to guarantee maximum acceptable variance,

the risk is a cognitive divergence, the defining competitive advantage of the next decade will not be owning the fastest AI, but possessing the most robust infrastructure.

For European enterprises, this is a total paradigm shift, by entrusting the optimization of energy grids, the regulation of financial flows, or the coordination of supply chains to foundation models, we are outsourcing the very understanding of our systems,

Dear businesses, dear clients, as you know, we are undergoing a tectonic shift: the transition from a deterministic IT world to a probabilistic logic

The cognitive infrastructure shift called AI has already started, businesses are entering a new era. 1/6

✅ Turn regulation into your competitive advantage. 📞 Book your free discovery call today and start your compliance journey. 🌍 Coming soon: our Global Compliance Program — guiding your market entries in the U.S., India, and China.

💼 Pricing: project-based — no surprises, no retainers. For long-term peace of mind, choose our Annual Compliance Support: → Quarterly monitoring → Periodic audits → Continuous updates as new AI rules (like GPAI guidelines) emerge.

🛠️ What we offer: • A dedicated compliance expert for your project • Full risk classification under the EU AI Act • GDPR integration (privacy by design) • Technical audits & actionable compliance plans • 24/7 availability

We’re launching our new EU AI Compliance Service helping you master the EU AI Act from day one. 🧩 As of October 2025, the regulation’s rollout is accelerating: -High-risk AI enforcement in 2026 -Full application by 2027 If you deploy high-risk systems, now is the time to act.

Our services : 1️⃣An AI Act consultant fully dedicated to your project 2️⃣Unlimited legal advice from your dedicated consultant 3️⃣All your legal documents provided 4️⃣24/7 access to our customer service 5️⃣An "AI Act Compliant" label All consultants trained in GDPR compliance

Prediction: LLMs in their current form may not be able to do everything, but AI now has enough momentum that this won't matter. Beam engines couldn't do everything either, but they were enough to set off the Industrial Revolution.

How do they have European users ? VPN.

As things stand, Sora would struggle to pass European compliance tests, which require increased transparency, robust human oversight, and the protection of fundamental rights.

Could Sora 2 be released in Europe ? Spoiler : no. Sora would likely be classified as "high-risk" due to its potential for manipulation (biometric, behavioral) and disinformation.