Robin Costé

🗞 More new developments from the Wikimedia Enterprise Team! 👏 🤗 We just released a new Wikipedia dataset on Hugging Face!   • Full English and French Wikipedia articles   • Structured JSON format for easy ML integration   • Includes article metadata, abstracts, infoboxes, and parsed sections   • Ideal for NLP tasks, knowledge base construction, and more   • Open access under CC BY-SA 4.0 and GFDL licenses Big news and exciting times for developers and researchers! If you want to explore our datasets you can do so here 👉 https://lnkd.in/eBtGh2M4 Need higher thresholds, interested in becoming a testing partner, or have questions? DM me to chat. Let's unlock the power of Wikipedia data and pave the way to smarter data access! 🔑 #Wikipedia #WikimediaFoundation #WikimediaEnterprise #HuggingFace #FreeAPI #StructuredContents #SnapshotAPI #KnowledgeSharing #DemocratizingKnowledge #OpenData #AI #MachineLearning #NLP #AIResearch #DeveloperTools #WikipediaData #StructuredData #KnowledgeGraph

2

Firecrawl (YC S22) is an all-in-one developer platform for crawling & scraping web data for AI applications. While founders Eric Ciarla, Caleb Peffer, and Nicolas Silberstein Camara were working on Mendable.ai, an early RAG platform used by companies like Coinbase, Snap, and MongoDB, they quickly realized that while web data was a valuable resource for AI, its quality was essential for success. Facing numerous challenges in building a reliable solution for various URLs, they found existing tools insufficient. They envisioned an API capable of crawling web pages, handling edge cases, and delivering up-to-date, user-friendly markdown. This led them to build Firecrawl. With Firecrawl, developers at companies like Amazon, Nvidia, and Zapier are delegating scraping to us so they can focus on their core tasks - be it RAG, agents, or data processing. It's an open source API that transforms any web data into a clean, LLM-ready format for RAG, agentic tasks, or training. Since launching in April, Firecrawl has already gained 8000 stars on GitHub ⭐️ 🚀 https://lnkd.in/gD9i3GPk

683

25 commentaires

📣 👉 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

CoderSchool (IT W21) saved $1000+/mo by switching to Lark Suite from Google Workspace, but there's even something more mindblowing about Lark...are you ready for this 🤯? There is a 🧋 Milk Tea Group Order🧋template in their default gallery. In all seriousness, while I originally switched as a cost-savings measure, I'm finding Lark Suite surprisingly delightful to use. It's fast, fluid, and unlike the Google tools, all the tools work together seamlessly, centralized around a "chat-first" philosophy. If people are interested, or even if they're not, I'll continue posting about some of the steps involved in our migration from 9+ years on Google Workspace. It wasn't that hard, although we're not through all of it yet. I'd encourage all startups to take a look - Lark is free for up to 50 users, so most small co's can replace Gmail/Notion/Slack/Sheets/Airbase/Drive for $0/mo. And you can still keep "Sign in With Google" on your old domain. Thanks to Mark Dembitz for the presentation about Lark. I was lucky to be sitting front and center wearing my bright red CoderSchool T-Shirt while you had the slides about why red is a bad choice for a brand color.

90

27 commentaires

#DataOrchestration: Kestra Secures $8M to Bridge Orchestration Gaps Kestra, the open-source data orchestration platform, has raised $8 million in seed funding led by Alven, with participation from industry leaders like Datadog's Olivier Pomel and Hugging Face's Clement Delangue. Kestra aims to unify fragmented orchestration tools, offering a comprehensive platform that simplifies data workflow management. With over 500 integrations, including AWS, Google Cloud, and Azure, Kestra empowers both technical and non-technical users to streamline operations swiftly. How might Kestra's expansion influence the data orchestration landscape? #Kestra #DataOrchestration #OpenSource #TechFunding #Saasverse

16

Frankly I find ` OOPS generally excessively complicated . The simplest ` object is an " address space " an ordered list . I've commented to Alan Kay long ago , eg: https://lnkd.in/gTSi-fRy https://lnkd.in/gJiuCyFQ , that that what ` objects miss is the ` geometric attribute that APLs , including K , naturally retain . In CoSy.com , evolved from K as an open vocabulary in Forth , the 3 cell header for an ` object is simply `( Type Count refCount )` ( where ` refCount handles memory management , free on return to 0 ) . Upon that , any more complex Types can be constructed ` inheriting the entire general list manipulating vocabulary , again , evolved & greatly simplified from APL via K .

🚀 Exciting news! My close collaborators, Alain Papazoglou and Philippe Perou from the Kredeum team, were finalists in the Scaffold-eth-2 Hackathon with their OnChainAI extension! 🌟 What is OnChainAI about? The OnChainAI extension, built on Scaffold-eth-2, uses Chainlink Functions to enable on-chain requests to OpenAI, facilitating real-time AI-powered interactions directly on the blockchain. If your smart contract needs to make prompts to OpenAI, don’t hesitate to use our extension. It’s also an open-source project, so feel free to contribute by sending your PRs for new functionalities. 🌟 Yeah but, what is Scaffold-eth 2? 🔍 Scaffold-eth is an open-source development framework created by Austin Griffith that simplifies building Ethereum-based applications with pre-configured components and tools. Ideal for rapidly creating and deploying dApps. Haven’t heard about it yet? Stay tuned for more information! 🌟 What is Kredeum about? 🔍 Kredeum is best known for their Kredeum NFTs Factory, a versatile platform that simplifies NFT creation and trading across multiple blockchains like Ethereum and Polygon. With tools like: 🏦 NFT Wallet 📜 Smart Contracts Library 🛠️ WordPress Plugin The platform enables easy customization and cross-chain support. 🌐 Now, with the introduction of AI through OnChainAI, Kredeum is taking a significant step forward, further solidifying their impact on the Web3 ecosystem. Proud to be connected with such minds ! 🌟 And remember, we are open to collaborating on your project! 🤝 🚀 Let's build the futur together! 🔗 To use the OnChainAI extension on Base Sepolia Testnet : website: https://lnkd.in/eU3eNJzd github: https://lnkd.in/eVM8TsQX 🔗 Discover Kredeum website: https://www.kredeum.com/ github: https://github.com/Kredeum discord: https://lnkd.in/eymZJS7W #Blockchain #AI #Web3 #Innovationt #WordPress #OpenAI #Kredeum

9

3 commentaires

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

𝐓𝐡𝐞 𝐅𝐫𝐞𝐧𝐜𝐡 𝐆𝐞𝐧𝐀𝐈 𝐄𝐜𝐨𝐬𝐲𝐬𝐭𝐞𝐦 - 𝐖𝐡𝐚𝐭’𝐬 𝐧𝐞𝐰 𝐢𝐧 𝐆𝐞𝐧𝐀𝐈? Significant activity across multiple layers with the French GenAI Ecosystem: 🖥️ Major tech companies, like Meta, are scaling their compute infrastructures, with plans to have 600K H100 GPUs for internal use. Other are going to 10,000+. 🚀 Companies such as OpenAI are reporting rapid revenue growth, reaching a $4B annual run rate in just a few months. Everyone wants cheap Intelligence, everywhere! 🧠 Open-source models, including Llama 3, are increasingly popular, challenging the dominance of closed-source models. 🛠️ Foundational model providers are focusing on creating seamless developer platforms to retain users and build long-term value. ⚠️ Startups in the application space face growing competition from LLM providers. 🧪 GenAI startups focusing on material discovery are gaining momentum. 🎥 AI-driven visual content creation, particularly for ads and UGC-style videos, is on the rise. MarTech multi-modalities are here. 📞 Startups applying AI to business process outsourcing (BPO) are emerging, with a focus on improving efficiency in customer support and call centers. AI x BPO. 🧬 Science & Development: Companies like Iktos, Owkin, and Altrove are advancing GenAI in healthtech and cybersecurity. 🎮 Application Layer: Startups like Flim, Taplio, and Animaj are innovating in multimedia, gaming, and marketing. 🗄️ Data & Enterprise Platforms: Pathway, Octopize, and Dataiku are powering the data and enterprise infrastructure behind AI apps. 🛠️ Dev Tools & Infrastructure: Companies like Hugging Face, FlexAI, and Giskard are providing crucial development tools and safety frameworks for GenAI scalability. 🔧 Foundational Layer: Leaders like Mistral AI and Pleias are building foundational models and driving open-source GenAI initiatives. The french GenAI ecosystem demonstrates a balance between innovation at the foundational model level as well as infrastructure and applications. Source: https://lnkd.in/ggHGvbxf #AI #GenAI #LLM #MLLM #AIInfrastructure #OpenSourceAI #Startups #LLM #MaterialDiscovery #BPO #AIforGood #HealthTech #Cybersecurity #FoundationalModels #GenAIDevelopment #AITrends #MultimodalAI

4

I’m so proud to be part of Pruna AI’s journey to redefine AI accessibility and sustainability! With $6.5M raised, Pruna AI is transforming the AI landscape by compressing models to make them faster⚡, cheaper💲, smaller💾, and greener🌳. I am inspired every day by the dedication of the co-founding team – Bertrand Charpentier, Rayan Nait Mazi, John Rachwan, Stephan Günnemann – to making efficient and accessible AI a reality. Follow Pruna AI for more updates! #EfficientAI #GreenAI #SmallAI #FastAI #PrunaAI

34

1 commentaire

Hugging Face x LangChain : A new partner package in LangChain - We are thrilled to announce the launch of langchain_huggingface, a partner package in LangChain jointly maintained by Hugging Face and LangChain. This new Python package is designed to bring the power of the latest development of Hugging Face into LangChain and keep it up to date. langchain-huggingface integrates seamlessly with LangChain, providing an efficient and effective way to utilize Hugging Face models within the LangChain ecosystem. This partnership is not just about sharing technology but also about a joint commitment to maintain and continually improve this integration. Getting started with langchain-huggingface is straightforward. Here’s how you can install and begin using the package: pip install langchain-huggingface #llm #huggingface #langchain #genai #generativeai #artificialintelligence

25

Le gratin du machine learning...

2

Discover my new personal web site : Free prediction contest (without money) : Linkedin has delete my precedent post since he think betting : no no money free and respect country legislation ! https://lnkd.in/ea2AD6Dc Discover our new web site : >> Done > Place prediction > Register with captcha v2 login and connect with google. Front with https://bolt.new and backend from sratch in NodeJS. >Matchs : France ligue1 matchs are automatically injected in database via a free api in json >Admin module to manage users, tournaments bets matchs >Have a libretranslate docker to translate website more than 30 languages >forgot password >captcha front >>Incoming >See results scores >Add provisioning of all ligue provided by the free api : more than 100 >add top entity service ID code to manage yourself your tournament and invite your friends but quote calculation will stay global >Show Google analytics results called by API >Simple model for quote calculation first >See fictiv rewards. >captcha back >update Register with captcha Google v3 >Quote calculation with Monte Carlo Simulations >Improve speed translation by pre stored all sentence traducted Please test and give me your reviews. ^^ Article here : https://lnkd.in/egU_S3he Lien vers le site : https://lnkd.in/e3hiKVAP

2

2 commentaires

Another open source model, Moshi!! #llm #genai #ai

7

1 commentaire

Today is the first OpenAI DevDay for this autumn. No new models are expected, but OpenAI has promised to introduce some long waited features to APIs. This is the schedule (in San Francisco timezone): - 10:00 - Welcome + kickoff by Olivier Godement & Romain Huet - 11:00-11:45 - Topic to be revealed during keynote - 12:00-12:45 - Topic to be revealed during keynote - 12:45-12:00 - Lunch - 2:00-2:45 - Topic to be revealed during keynote - 3:00-3:45 - Topic to be revealed during keynote - 4:00-4:45 - Fireside chat with OpenAI leaders - 4:45-7:00 - Closing reception So reading this, there could be up to four new announcements coming. Here is my wishlist in priority order: - Audio/voice multimodal API for gpt-4o (so advanced voice mode API) - Function call webhooks, so that AI could call app function in mids of voice-call or o1 reasoning sessions. - New gpt-4o image generation API - Prompt caching - Web search enabled api (ala Perplexity or Command-r) But I hope they leave the best for London in end of month, where I'm going 😂

12

1 commentaire

Your Contribution for tonight ✨: Airbyte ⬇️ Co-founded by Michel Tricot and Jean (John) Lafleur, Airbyte is a data integration platform for ETL / ELT data pipelines. You can integrate from APIs, databases, data warehouses, data lakes and data lakehouses. Airbyte is available self-hosted and cloud-hosted. 🧪 Here’s why you should contribute to this repo: 1️⃣ Their tech leverages modern practices to help solve data engineering challenges, affecting hundreds of companies and data professionals. If you are a Python or Java developer and want to further contribute to a powerhouse repo (with over 13,000 stars!), this is your opportunity. 🦾 2️⃣ The repo "really" welcomes new contributors. Many projects have a contributing.md file but rarely accept new contributors. Airbyte brought on their repo 12 new contributors just this last month! If you are looking for a welcoming repo, Airbyte may be what you have been looking for. 😄 3️⃣ Nearly ~45% of contributors submit a second PR. This means the repo has a supportive environment for building upon past contributions. If you enjoy working on this repo, chances are you can continue to work on it! ⚡️ If you are interested in contributing to building the only tool devs need to move data, check more of Airbyte's stats on Quira (https://lnkd.in/eM7etquW) or access the GitHub page directly here: https://lnkd.in/e_4Gczrk

13

A useful price comparison of LLMs.

7

This is how you can parse unstructured text *locally* and *cheaply* (no GPU required) in less than 5 min using the latest models from NuMind (YC S22) Many thanks to your team Etienne Bernard 👏 How to (5 min read) : https://lnkd.in/e2sbGnzn

3

Grateful to have been part of AI-Pulse! 🌟 A huge thank you to the organizers : Scaleway, Constance Morales, Frédéric Bardolle, and STATION F for hosting such an impactful gathering of minds. I was thrilled to contribute as a workshop speaker on "Hands-On Integration: Simulating Embedding Models for Structured Data." In our hands-on session, we explored the limitation of using language models for number representation. It’s clear that while LLMs excel in many areas, they are not the right tools to handle structured formats—such as tabular data and numeric representations within strings. And this is why Neuralk AI is building AI embedding models that are specifically designed for structured data. It was a pleasure sharing ideas and experimenting together with such a passionate group. Looking forward to continuing the conversation on optimizing embeddings for structured data in future forums! Workshop Code: https://lnkd.in/ek83-Mgk #AI #MachineLearning #StructuredData #EmbeddingModels

30

3 commentaires