Today we announce collaborative banking
INNOVATION
-

Telecom Networks Evolving into AI-Driven Autonomous Systems
By
–
Telecom networks are evolving into AI-driven systems capable of sensing, learning, and adapting. That was one of my biggest takeaways from @IBM
’s telco discussions at #MWC26 Why autonomy is becoming operational intelligence. https://
linkedin.com/posts/haroldsi
nnott_mwc26-autonomousnetworks-ai-activity-7435253319603912705-RlWm
… #AI @MWCHub @GSMA -
Balancing specialization with experimentation in tool adoption strategy
By
–
Totally agree! Specializing definitely pays out. But there’s still a point where you need to keep some kind of experiment vs. Exploitation ratio for new tools, else we’d still be stuck with what we were using years ago. But that’s my usual take to wait a few months and see
-

Hugging Face Builders Paris Chapter: Lead Open AI Community
By
–
Hugging Face Builders is a global community program that puts local leaders at the center of the open-source AI movement If you're passionate about open AI and love bringing people together, this is your invitation to lead Apply for to build the Paris chapter today
-
Mechanics and Software Integration: The Real Industrial Challenge
By
–
Most people think we can code our way out of any industrial challenge. #sponsored by Siemens.
— Lucian Fogoros (@fogoros) 12 mars 2026
But here is the truth: If the mechanics don't fit, the software is stuck.
Rainer Brehm of @Siemens breaks down why the "interaction" between the physical and digital worlds is the true… pic.twitter.com/nMmAUnZy6GMost people think we can code our way out of any industrial challenge. #sponsored by Siemens.
But here is the truth: If the mechanics don't fit, the software is stuck.
Rainer Brehm of @Siemens breaks down why the "interaction" between the physical and digital worlds is the true -

LLM2Vec-Gen: Frozen LLMs Generate Better Embeddings Through Reasoning
By
–
LLM2Vec-Gen represents a major paradigm shift for embeddings/retrieval. Why encode the query when the LLM already knows what to look for and can directly produce an embedding for it? Best part: it’s self-supervised, and it does all of this while the LLM remains completely frozen. Think about it: "solve x² + 3x − 4 = 0" has zero reasoning in it. But the LLM's response does. By encoding the response, the embedding captures the reasoning — and the better the LLM reasons, the better the embedding. This is why our results scale with model size. As LLMs get smarter, our embeddings automatically get better. LLM2Vec-Gen is also the first demonstration of the promise of @ylecun's JEPA for text embeddings. The alignment loss is JEPA — predict in representation space, not token space. The reconstruction loss goes beyond — it keeps embeddings decodable. This paradigm shift opens new frontiers: 🔬 Can we build a full JEPA for language where the teacher and student are the same LLM? ⚡ Can LLMs reason in compressed space without ever generating text? 🤖 Can agents reason in compression tokens and carry that directly into retrieval? 💬 Can agents talk to each other in compression tokens instead of text — dense, fast, and still human-readable? LLM2Vec-Gen is a first step toward all four. Vaibhav Adlakha (@vaibhav_adlakha) Your LLM already knows the answer. Why is your embedding model still encoding the question? 🚨Introducing LLM2Vec-Gen: your frozen LLM generates the answer's embedding in a single forward pass — without ever generating the answer. Not only that, the frozen LLM can decode the embedding back into text. 🏆 SOTA self-supervised embeddings 🛡️ Free transfer of instruction-following, safety, and reasoning — https://nitter.net/vaibhav_adlakha/status/2032065008603951187#m
→ View original post on X — @hugo_larochelle, 2026-03-12 12:37 UTC
-
Technological Inversion: Military to Civil and Back with AI
By
–
Until roughly the 1970s, most civilian technological progress was downstream of military technological progress. The trend has since inverted. Today most of the modern (e.g. Ukraine) military playbook comes from consumer tech from the past two decades. With AI entering the
-

AI21 Labs to Showcase Enterprise AI Solutions at NVIDIA GTC26
By
–
We are heading to @NVIDIA #GTC26 in San Jose. If you’re attending, stop by Booth 3103 to meet the team and see what we’re building for enterprise AI. See you there.
-

New Agentic AI Engineering Course with Multi-Agent Projects
By
–
Our new course, "Agentic AI Engineering," contains 17 notebooks and a complete multi-agent project (because you can't learn to build production agents through notebooks only). Notebooks are really powerful for teaching the core concepts quickly and easily to anyone. Which we can
-

AI Monitors Real-Time Land Use Changes via Satellite Imagery
By
–
Can AI help governments track land use changes as they happen? Design AI systems that combine satellite imagery with ground-level data to monitor land use shifts across Andhra Pradesh in near real time, enabling smarter planning, better resource management, and more responsive