Jean-Guillaume Dujardin

AI is forcing VCs to rethink what “venture-backable” means. That’s what we explore with VCs in the March edition of Beating Heart of French Tech. One year ago, AI was already flooding dealflow. Today, the question is no longer if it’s investable but what still is. On the menu this month: - AI’s weight in dealflow and in future allocations - Overhype vs reality - What VCs now expect in terms of defensibility - Elaia & Eurazeo sharing their take 💬 - The hottest AI themes right now Thanks to all the VCs who already shared their view 🙏 For the others, still time to weigh in: 👉 https://lnkd.in/dezy8jvc Full edition dropping soon -> subscribe to Beating Heart of French Tech ➡️ link in 1st comment From Paris, with love 💚 Chausson Partners

22

1 commentaire

A 𝘀𝗰𝗮𝗹𝗮𝗯𝗹𝗲 and 𝘀𝗲𝗰𝘂𝗿𝗲 system built in 𝟯 𝗺𝗼𝗻𝘁𝗵𝘀. Two developers. No message queues. No external caches. Now powering 2,500 secure, real-time messages per second for over 𝟭𝗠+ 𝗵𝗲𝗮𝗹𝘁𝗵𝗰𝗮𝗿𝗲 𝗽𝗿𝗼𝗳𝗲𝘀𝘀𝗶𝗼𝗻𝗮𝗹𝘀 with 𝟵𝟵.𝟵𝟵𝟵𝟵% availability. Akka gave Doctolib the foundation to deliver: ✓ Seamless clustering for instant scale ✓ Low latency via sharding and pub/sub messaging ✓ Redundancy across multiple servers ✓ Simple deployment and ongoing management “It just works!” -Jasper Aarts ⬇️ Read the story #Innovation #Technology #Engineering #Java #DistributedComputing

32

4 commentaires

As LG Chair of the ETSI Software Development Group for OpenSlice, I’m happy to share one of our most ambitious development cycles to date. The release fixes more than 40 issues and provide some nice enhancements : - enabling Agentic interactions with MCP authentication & improved tool exposure - invoking chains of internal/external agents during service orchestration - a lightweight Git-based controller - TMF Product Management - Nokia Network-as-Code (NaC) extension - packaged service templates (including an indicative 5G Core example) A huge thank you to all SDG OSL members and participants for the contributions, reviews, testing, and continuous feedback that shaped this release. Special thanks to our TSC Kostis Trantzas, Rafael Direito, Georgios P. Katsikas for the guidance, prioritization, and technical stewardship throughout the cycle and ETSI for the infrastructure support and hosting! One more point I strongly believe in: Open Source contributions are not “free.” They require sustained engineering effort: governance, testing, documentation, and operations and, very often, the sacrifice of our personal/free time. Especially in Europe, a significant part of what we deliver is often made possible through public funding and research projects - which are increasingly difficult to secure. That’s exactly why we believe research results should be contributed back to the community as operational open-source software, giving them the opportunity to be openly sustained, maintained, and evolved over time.

46

1 commentaire

AI isn’t just for big brands anymore. In this interview, Meta CMO Alex Schultz explains how AI will give “superpowers” to SMEs and push big ad agencies to reinvent themselves. https://lnkd.in/ep7gXS7X Automation will make high-quality creative and media work accessible to teams that never had the budget for a big agency – while agencies that do survive will be the ones mastering new AI tools instead of fighting them. For me, the key message is clear: Either we learn to work with AI in communication, or we slowly fall behind those who do. Alex especially recommends to use ChatGPT for creatives, Veo/Sora for the animation and also learn Vibes (Meta's own content edition tool). I recommend reading the article (French) with your favourite browser translation – and asking yourself: where do I personally need to level up with AI? Curious what you think after reading it 👇 #ai #marketing #communication #sme #adtech #futureofwork

5

Very insightful contribution 👌🏽. I believe Elon Musk’s approach makes sense, especially considering that any form of censorship in such an undertaking would likely be easy to detect. From my own experience, I’ve explored different AI models across various projects, and I’ve often observed how many of them are designed to “play it smart” right from the start. Placing greater emphasis on clarity of purpose and refinement in an AI project reflects a level of maturity and a genuine drive to do better and push progress forward. More broadly speaking, AI models tend to mirror the values and intentions of those behind them—be it engineers, individuals, or entire organizations. That’s something worth keeping in mind, particularly when engaging with large language models (LLMs). As for Elon Musk, his continued pursuit of AGI fits neatly into his broader tech-driven vision. Yet beyond the ambition, what stands out here is the lucidity of an entrepreneur firmly set on carving out a unique place within the AI landscape—especially through a niche-focused model like Grok. #ArtificialIntelligence #AGI #ElonMusk #Grok #AIethics #LLM #TechInnovation #DigitalForesight #LinkedInInsights #AIconversations

1

𝐏𝐫𝐨𝐮𝐝 𝐦𝐨𝐦𝐞𝐧𝐭 𝐚𝐭 𝐁𝐞𝐭𝐭𝐞𝐫𝐄𝐝𝐠𝐞. What our team has built over the past years is nothing short of groundbreaking: a new kind of sovereignty anchor powered by best‑practice Confidential Computing. We’ve implemented a full environment for easy generation and portability of 𝐓𝐫𝐮𝐬𝐭𝐞𝐝 𝐄𝐱𝐞𝐜𝐮𝐭𝐢𝐨𝐧 𝐄𝐧𝐯𝐢𝐫𝐨𝐧𝐦𝐞𝐧𝐭𝐬 that don’t just protect data in use — they automatically take care of secure encryption at rest as well. A seamless chain of trust, end to end. The video says it best: “𝐖𝐞 𝐚𝐫𝐞 𝐞𝐱𝐩𝐞𝐫𝐭𝐬 — 𝐞𝐱𝐩𝐞𝐫𝐭𝐬 𝐢𝐧 𝐜𝐨𝐧𝐟𝐢𝐝𝐞𝐧𝐭𝐢𝐚𝐥 𝐚𝐧𝐝 𝐝𝐞𝐩𝐞𝐧𝐝𝐚𝐛𝐥𝐞 𝐜𝐨𝐦𝐩𝐮𝐭𝐢𝐧𝐠.” And today, I couldn’t be prouder of the people behind that statement. #BetterEdge #ConfidentialComputing #SouvereigntyAnchor #DigitalSovereignty

17

1 commentaire

The End of an Era at Meta: Why LeCun’s Bet on "World Models" Matters Yann LeCun’s departure from Meta tells us something about the LLMs and that the "Scaling Laws" debate is far from settled. Reading this article in Le Monde today, I found myself nodding along with his core premise. After multiple implementations of AI initiatives in productions based on LLMs and ML ($150M+ annual savings), I’ve hit the same challenges LeCun describes: ↳ The Probabilistic Trap: LLMs are statistical engines, and still struggle with reasoning. You can refine the weights, but you can’t engineer away the inherent stochasticity. ↳ The Determinism Gap: For enterprise applications, "mostly right" may not be enough. We need systems that can plan, reason, and understand cause-and-effect with high accuracy. Autoregressive text prediction fundamentally struggles to deliver them. LeCun is betting that World Models (learning from physical reality rather than just text) are the path forward. Given the architectural ceilings we're seeing with current Transformers, I think he’s right. We need better architectures. #AI #MachineLearning #YannLeCun #WorldModels #LLM #TechNews #AIUnleashed #LeadingWithAIAgents

7

Legacy systems really are eating the world. Instead of driving innovation, a significant percentage of IT budgets and resources are consumed by maintaining, patching, and securing obsolete software. Such a waste 😭 Definitely worth listening to Fabrice Bernhard on how we are using AI to solve such a massive problem.

13

2 commentaires