Samy Lahbabi

New paper from myself and Anne Laure Thibaud (Thieullent) on the topic of how the rise of operational AI is going to fundamentally alter the way that organizations need to look at data. Key points: 1) GenAI (and Agentic) will enable digital employees at operational speeds 2) Digital Employees need data that reflects the operational reality 3) You need to establish a digital operating model 4) You'll need to plan for people managing AI as well as people 5) The Business needs to take control of its digital reality When we started writing this paper it felt like this world was still a while away, since then we've had an explosion in applications embedding GenAI and Agentic solutions. So now it more feels like this is how you need to plan for 2025 rather than a vision for the medium term. Myriam Chave, Mark Oost, Sergey P., Franck Greverie, Eric REICH, Alex Marandon https://lnkd.in/gFw2w9aJ

56

7 commentaires

When to Invest in Long Context In-Context Learning ? 🗼 As language models continue to grow in size and capability, the way we use them is also evolving. A recent study by Bertsch et al. explores the potential of in-context learning (ICL) with long-context models, pushing the boundaries of how many examples can be provided as demonstrations. The results are surprising and thought-provoking, suggesting that in certain data regimes, long-context ICL could be a powerful alternative to traditional fine-tuning and retrieval methods. Traditional few-shot ICL, where only a handful of examples are given, has already seen wide adoption due to its simplicity and reasonable performance. However, it can fall short when dealing with complex tasks or large label spaces. On the other hand, fine-tuning the entire model on a large dataset has been the go-to approach for maximal performance, but comes with higher training costs and reduced flexibility. Long-context ICL offers a middle ground, allowing the model to attend to many more relevant examples without expensive fine-tuning. The study highlights that as the number of demonstrations grows, the importance of careful example selection diminishes, making it possible to reuse a single set of examples for all inputs. It is prudent to already state that LLMops system design will be a careful balance between retrieval with few shot examples (with knowledge graph or vector database), in context learning with long context model, and fine tuning to handle reasoning and context.

27

Super excited to speak at BIG DATA & AI PARIS on October 15th! Looking forward to sharing insights on how no code BI tools like Domo can help foster a data-driven culture in startups like Hive. Check out the details here: https://lnkd.in/e6e-CtpR. #BigData #AI #DataDrivenCulture #BI #Startups #BigDataParis

23

One year ago, we released our first mapping of the French Gen AI ecosystem. At the time, the landscape was just beginning to take shape—full of promise but still in its infancy. Today, that promise has been realized in a big way. In just one year, the number of GenAI companies has more than doubled —rising from 73 in 2023 to 150 this year. Together, these companies have raised an impressive $1.4b over the past year. Foundational models have expanded beyond language and now exist on images, videos, biology, chemistry, and more (the top 4 largest gen AI fundraising in France are foundational model). We are also seeing more and more verticals being disrupted by genAI applications. GenAI has been a core focus for Resonance reflected in our recent investments (Tomorro, Zeliq, Animaj, Flagcat) and more exciting rounds yet to be announced. This map is the result of our team's dedicated work, but the ecosystem is vast and evolving. If we missed any key players, please reach out—we’d be glad to include you. If you want to get access to the full airtable, see in the first comment ⬇ If you're building here, please reach out! 👋 cc Mathieu Sacchi Théophile Heymans

658

90 commentaires

EU-Startups: 100 start-ups created every year in laboratories under CNRS supervisory authority #startups #frenchtech #entrepreneurship #market #impactyse #4u2win!

10

How can we make AI resilient and robust? I was part of a panel discussion at the InCyber Conference, held in Lille, France, that debated the issue of "Resilient AI" - a challenge given how AI continues to rapidly evolve, especially in the context of arresting cybersecurity threats. A luminary panel of executives from SNCF, Oracle, Salesforce and EazyML discussed a variety of questions: What is a critical system and how is it characterized; How can AI show points of failure; What are the recipes and best practices for creating resilient AI, capable of adapting to unforeseen situations and ensuring their proper functioning even in the event of critical incidents; How can we increase the robustness, reliability and transparency of AI algorithms to meet resilience requirements in strategic areas such as health, security and critical infrastructures? Here's the recording of the session that discussed possible answers. Would welcome your thoughts. Thanks. https://lnkd.in/ehqKbhv4 #machinelearning #bigdata #ai #cybersecurity

14

1 commentaire

The #AEC world is increasingly digitalised. It can be beneficial to learn from what other business domains have spent decades researching - ontologies, semiotics, semantics, representation. The reposted paper is quite advanced, but also explains key concepts in this world well. Reposted paper is best enjoyed with a quite large cup of tea, and an appreciation of that digesting this paper takes some effort for those not heavily schooled in Digital. Thanks Dr Nicolas Figay.

1

2 commentaires

How Public Administration can benefit from AI today. Olivier Perard from Oracle presenting a PoC done together with Novoshore Europe S.L. using #orclAPEX and Generative AI to automatically generate pre resolution to urban planning applications in Junta de Extremadura #orclAPEX #Oracle #ai #PotencialDigital2024

25

📣 👉 Amazon to invest $1.3 bln in France ! "Amazon Web Services (AWS) investment would increase cloud infrastructure in the Paris area to support France's flourishing #generativeAI opportunity" If you ever had a doubt of how much French business is embracing #hybridCloud and #GenAI, do not doubt any more ! As we've already demonstrated as a prime AWS partner, IBM Consulting is the right partner to support efficient and at scale #GenAI uses-cases in #France. https://lnkd.in/dPdsxZ6w

31

1 commentaire

Generative AI Futures : Chip To Code ? What do you think about the growing trend where companies like NVIDIA are entering the generative AI space with a full-stack approach? They're not only designing GPUs but also actively participating in the development of AI technologies to improve efficiency at scale. Is this a trend we should keep an eye on? Will leaders in generative AI space - Azure AI ,OpenAI ,Databricks , Google Cloud,Hugging Face, LlamaIndex , Microsoft Copilot have to watch this out as a trend for them to embed into their core , as opposed to seeing these as forward and backward integration partnerships Need to pay attention to this approach, or is it just an outlier? #generativeai #genaistrategy #genai #backwarditegration Note When I wrote this post, I was not aware of another post published a few months ago that clearly indicates a trend in chip activity. Updating, given its significant as an update https://lnkd.in/gD4Fuybh

7

1 commentaire

Cool opening at the French (and European) top AI conference called AI-Pulse that anybody can watch. A lot of things happen in AI these days, and many companies (including SalesOMMO) are moving fast into applying AI to the day-to-day life. Major industry players, academics and Governments from EU and US are in the same race that was going on in #tech for tens of years - building better, faster, cheaper, now also more regulated systems. Some do things more open, some keep more secrecy about what they do, but all build APIs and seek funding or developers to make it happen. Many American companies represented, but also more Europeans want to build things that matter, and are showing them off - really impressive #FrenchTech on display, more details in the video and a content summary below. A lot of competition going on in the field, a lot of money go into those systems, for sure not all of these companies will be here in 10 years from now, but some will. As an angel investor I can only say that is better to make your bets now and invest in #early stages in some.of them (build a portfolio, look around for them also in Romania 🇷🇴 and EU 🇪🇺), not when they are already at the top and flipped to US 🇺🇸. #AIpulse 》》》《《《 Damien Lucas - Scaleway - offering up to 5,000 GPU powered with green energy and quantum computing l, switching from closed models (i.e. OpenAI) to open source models on their infrastructure Michael Dell - Dell Technology - good hardware is still needed (from notebooks to servers and storage), and AI is going to the edge where data is. Renee James - Ampere Computing - ex-Intel Xeon team building next generation CPUs with optimized power consumption, getting back AI to core (instead of offloading to GPUs) Brian K - head of Nvidia applied deep learning research, pushing for the limits with clusters of 100,000 GPUs, building systems that self optimize for faster inference, and ready to revolutionize everything Renen Hallak - Vast Data - building architectures for data and computing power to optimize both fast access and vast (of course, as company name) amounts of data in all formats needed by AI Patrick Perez - Kyutai - building their open mode called Moshi, and multi-modal foundations for AI to understand better voice, video, and anything else in a natural way, and also moving some of the new and powerful models to notebooks (the final edge) Charles Cantour - "H" company - visual models and multimodal systems, just had a 200M investment, looking for partners that want to develop action agents to automate things in various verticals Clara Chappaz - Secretary of State for AI and digital affairs - acceleration of AI requires people, infrastructure, and funding, plus more uniform and practical regulations across EU - we can all call her if we see regulations go rouge, she seems to be on the bright side for European tech. 》》》《《《 https://lnkd.in/d6hUHk79

11

Finitism in #AI datasets can be mitigated if a more significant portion of the worldwide internet populace can create data, and if there were a method to incentivize these interested contributors to create such data. A big drum roll for Pertin-ant 🐜here... #AI #AGI #LLMs #trainingdata

6

🌍 Exciting times ahead for digital health in France! I had the honor of speaking at the roundtable on "Digital Medical Devices: Where do we stand in their assessment and what healthtech stakeholders need to know?" organized by the Agence de l'innovation en santé (Lise ALTER, Camille Schurtz 👏 ) at the "Next generation of clinical research" event in Lille. The discussion was rich with insights on the current landscape of research and the future of digital therapeutics in France, especially. Special thanks to our moderator, Line FARAH (Ministère de la Santé), and my fellow speakers: Jérôme Fabiano (EIT Health France), Corinne Collignon (Haute Autorité de Santé), Sarah Zohar (INSERM), Dr. Joris Giai (CHU Grenoble Alpes) and Louisa Stüwe (Ministère de la Santé). 💡 One of the key insights: YES, foreign clinical data can be used to apply for reimbursed of a digital medical device in France! Let's connect and explore opportunities together! 🚀

95

1 commentaire

Join Michael Mansard from Zuora, Mélanie Septe from Cegid and Steven Forth from Ibbaka to get real world insights into AI monetization. Wednesday, October 30 2024 10:00 AM PT | 1:00 PM ET Generative AI is revolutionizing SaaS, but monetizing it sustainably is key. This session covers four key approaches for packaging AI, usage-based pricing, and agile strategies to maintain profitability.   Key takeaways: - GenAI's Impact: How it's reshaping SaaS and creating new opportunities. - Smart Pricing: Usage-based models aligned with customer needs. - Packaging for Profit: Strategies to maximize ROI. - Agile Monetization: Flexible strategies to adapt to AI advancements. - Industry Frameworks: Actionable monetization strategies from experts. https://lnkd.in/gmRP2eCa #AI #genAI #generativeAI #monetization #pricing #casestudy #SaaS #verticalAI

8

1 commentaire

We made predictive analytics algorithm for the French legislative elections - 2nd round, using an #AI model home made by our IPFCconline1's team. Here are the key takeways: - The far-right Rassemblement National is on the verge of power of power in France - President Macron and his center-right supporters ( ENSEMBLE ) have been disavowed, losing half their representatives in parliament - The future French National Assembly remains in a state of uncertainty, however: Will the far-right reach the fateful threshold of 289 elected members of parliament? , which would mean an absolute majority, leaving the door wide open for a far-right government who could develop its policies, without an absolute majority, France could find itself institutionally "blocked", with 3 distinct political blocs (Fart Right, Centre Right and Left wing) unable to pass any legislation... And let's end with the international press view of this French election: in Switzerland, the major German-language center-left newspaper 'Tages-Anzeiger ' headlines that the Le Pen wave is wiping out Macron's power, and deplores that "the country of the Enlightenment, human rights and cosmopolitanism is drifting further to the right than ever - and perhaps towards darkness, isolation and xenophobia". -- Cc Karin Lakhoune Mia Shah-Dand Theodora Lau Eric Thienot Yann Marchand Jean-Baptiste Lefevre Christine Soto Antonio Vieira Santos Antonio Grasso Nicolas Babin Dominique Crochu Imtiaz Adam donald callahan Evan Kirstel B2B TechFluencer gregory valancius Brian Ahier Anthony Rochand Jean Cayeux IT Mallys NM Eric GAUBERT Sally Eaves Sabine VanderLinden Tamara McCleary Harold Sinnott Mario Pawlowski 🇵🇱🇺🇦🇺🇸🇪🇺 Nadia Bahhar-Alves Bruce McTague Chris Garrod Hessie Jones

20

2 commentaires

Excited to share that we’re sponsoring this year’s Apigee Summit, themed “No AI without APIs”! Join us as we dive into the future of AI-driven innovation and discover how APIs are the backbone of successful AI integration. I’ll be at the event along with our top Subject Matter Experts in AI and API management, ready to explore how your business can unlock the full potential of AI. Whether you’re curious about enabling LLM workloads, securing your AI endpoints, or scaling your AI initiatives, we’ve got the expertise to help you succeed. Don’t miss the chance to stop by our booth—let’s discuss how we can take your AI strategy to the next level. See you there! Omar Taher Mehdi Dahmani Bashar Khalil Mohammad M. Jean Dulau Matt Feigal Kevin Bouwmeester Julien Rateau Lisa Martinez Gomis Paul Rohan Nils Swart Scott Haaland Pramodh Ramesh Vallanur Aly Kuly Khan Faeza Asad Khan Maroof Muhammad Fatima Asad-Said #ApigeeSummit #AI #APIs #Innovation #DigitalTransformation #GoogleCloud #Technology #Paris

71

11 commentaires

How will 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 (gen AI) impact the 𝘄𝗼𝗿𝗸𝗳𝗼𝗿𝗰𝗲? Let’s explore this question back to Vivatech with Corine De Bilbao (Microsoft France), Gabriel Hubert (Dust.ai), and Jérôme Nanty (Carrefour Group), Anaïs PIERRE (PwC France) and Matthieu Aubusson de Cavarlay (PwC France) 𝗛𝗼𝘁 𝗧𝗮𝗸𝗲𝘀 🔥 01:05 - 🔥 𝗥𝗲𝗮𝗹𝗶𝘁𝘆 𝘃𝘀. 𝗠𝘆𝘁𝗵: Jérôme Nanty discusses how AI is already integrated into their business, providing examples of data management and customer service enhancements through chatbots and AI tools. 03:00 - 🔥 𝗘𝗺𝗽𝗹𝗼𝘆𝗲𝗲 𝗔𝗱𝗮𝗽𝘁𝗮𝘁𝗶𝗼𝗻: Corine De Bilbao emphasizes the rapid adoption of generative AI among employees and how businesses must adapt to this shift, highlighting a significant increase in AI usage among knowledge workers. 05:57 - 🔥 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗮𝗻𝗱 𝗪𝗼𝗿𝗸 𝗣𝗿𝗼𝗰𝗲𝘀𝘀 𝗖𝗵𝗮𝗻𝗴𝗲𝘀: Gabriel Hubert explains how AI can transform product experiences and internal workflows by structuring unstructured data and breaking down information silos within organizations. 08:28 - 🔥 𝗦𝗸𝗶𝗹𝗹𝗶𝗻𝗴 𝗮𝗻𝗱 𝗪𝗼𝗿𝗸𝗳𝗼𝗿𝗰𝗲 𝗘𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻: Corinne De Bilbao talks about the necessity of upskilling employees to handle more creative tasks, reducing time spent on administrative duties through the use of virtual assistants. 13:08 - 🔥 𝗖𝘂𝗹𝘁𝘂𝗿𝗮𝗹 𝗮𝗻𝗱 𝗢𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗖𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀: Jérôme highlights the main challenge of transitioning from legacy systems to new AI-driven processes, emphasizing the need for cultural transformation within large organizations. 19:02 - 🔥 𝗧𝗼𝗽-𝗗𝗼𝘄𝗻 𝗮𝗻𝗱 𝗕𝗼𝘁𝘁𝗼𝗺-𝗨𝗽 𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵𝗲𝘀: Gabriel Hubert and Corinne de Bilbao discuss the importance of combining top-down leadership with bottom-up innovation to successfully integrate AI, making it fun and engaging for employees. 22:28 - 🔥 𝗘𝘁𝗵𝗶𝗰𝗮𝗹 𝗮𝗻𝗱 𝗛𝘂𝗺𝗮𝗻-𝗖𝗲𝗻𝘁𝗿𝗶𝗰 𝗔𝗜: The panel concludes on the need for AI to be ethical and human-centric, with a focus on building trust and creating environments where employees can experiment and share insights. For more details, watch the full discussion on 𝗬𝗼𝘂𝗧𝘂𝗯𝗲 : https://lnkd.in/exadzEKw

9

1 commentaire

Exciting news in the French public sector cloud market! The French government is taking significant steps to accelerate cloud adoption, awarding a major €600 million contract to Crayon France and SMLB. This initiative aims to simplify the transition to cloud services for public administrations, offering a wide range of IaaS and PaaS solutions. The importance of managed services in this transition cannot be overstated. With the complexity of cloud infrastructures and the need for tailored solutions, expert guidance and support are crucial. The selected providers, Crayon France and SMLB, specialize in sourcing, procurement, distribution, and customer/supplier relations, ensuring a smooth and efficient cloud integration process. This strategic move aligns with the broader trend of recognizing the value of managed services in navigating the evolving IT landscape. By leveraging the expertise of specialized providers, organizations can optimize their cloud strategies, enhance operational efficiency, and achieve their digital transformation goals. Stay tuned for the launch of AltCloud by AllEdge, a managed cloud services marketplace on September 2024! #CloudComputing #DigitalTransformation #ManagedServices #PublicSector

3