🛠️ Vous développez un agent IA ? Le prototype n’est que le début. Optimiser le coût, la rapidité et la pertinence représente 90 % du temps. Découvrez comment les outils d’évaluation peuvent simplifier ce processus dans le dernier talk de @francoisz.bsky.social!👇 @afup.org youtu.be/NjrbQ8aWDv8?...

When using a backend-as-a-service like Supabase, you can only access the database through an HTTP API. Yet it is still possible to group queries into a transaction, without compromising on security. Read on to learn more. #database #security #NodeJS #supabase marmelab.com/blog/2025/12...

Generative UI is about letting agentic apps create user interfaces on the fly. A new specification called MCP Apps emerges, backed by Anthropic & OpenAI, to standardize it.It's very early stage ; to get an idea of the potential, try the popular MCP UI. blog.modelcontextprotocol.io/posts/2025-1...

Have you come across Spec-Driven Development yet? One AI writes the specs, another writes the code. Sounds smart, but it's basically waterfall in disguise. Here’s why we should stay agile instead. marmelab.com/blog/2025/11...

Etsy's Web Performance team implemented the Speculation Rules API (SRA) to prefetch product listing pages from search results, resulting in significant performance gains (20-24% improvement in metrics like TTFB, FCP, and LCP). #WPO www.etsy.com/codeascraft/...

"The ability to phrase our intent in natural language and receive working code does not replace the deeper understanding that comes from learning each language's design, constraints, and trade-offs." This analysis of #CodingAgents on #SoftwareDevelopment Is spot on. martinfowler.com/articles/llm...

Do #React Server Components Improve #Performance? When compared to a pure SPA, they reduce the LCP but don't suppress the initial interactivity gap. The main performance benefit comes from completely rewriting data fetching to be Server Components-first. www.developerway.com/posts/react-...

Using AI coding agents? You’re probably missing one thing: specs. I built vibe-spec, a CLI that extracts requirements from your coding agent logs and generates functional specs automatically. #AI #DevTools #OSS #softwareengineering marmelab.com/blog/2025/10...

Time flies #Symfony

Agentic Search surpasses Retrieval Augmented Generation (RAG): Teaching an agent to use grep is far more efficient than chunking, embedding and similarity search. #AI www.nicolasbustamante.com/p/the-rag-ob...

How to add metadata to 100M product descriptions? Etsy Engineering shares their journey with #LLMs, eval and monitoring. Great insights about a real-life #GenAI use case. www.etsy.com/codeascraft/...

Every developer hates AI hallucinations. But what if those “mistakes” are what make AI seem intelligent? Neuroscience says: noise isn’t a bug — it’s the engine of adaptability. #AIEngineering #AIAgents #LLMs #SoftwareArchitecture marmelab.com/blog/2025/10...

Scrum was revolutionary for human teams—but AI doesn't need standups, sprints, or sticky notes. The future of work requires dual-track agility: a slow lane optimized for human collaboration & meaning, and a fast lane for AI precision & speed. #agile substack.jurgenappelo.com/p/scrum-is-d...

Demain, c’est le Forum PHP organisé par l’ @afup.org ! Et devinez quoi ? @francoisz.bsky.social, CEO de marmelab, sera présent pour vous donner des conseils sur l'évaluation des agents IA ! 🤖 ⏰ 12h40 à 13h00 📅 9 octobre 2025 📍 Disneyland Paris 🔗 event.afup.org #ForumPHP #PHP #Symfony

MCP servers make LLM-powered agents super powerful, but also super vulnerable to all kinds of new attacks. The core vulnerability, prompt injection, is still way too easy to trigger. #Security #AI simonwillison.net/2025/Apr/9/m...

Is your frontend team making 10+ API calls per screen and waiting weeks for backend changes? New article: When you need a Backend-for-Frontend (and when you don't) 🧵 Real examples from Netflix, SoundCloud + implementation strategies marmelab.com/blog/2025/10...

When teams produce code faster than they can understand it, it creates what I’ve been calling “comprehension debt”. [...] When we have to edit the code ourselves, this debt is the extra time it’s going to take us to understand it first. #AI #Coding codemanship.wordpress.com/2025/09/30/c...

Engineers at CloudKitchens are saving 3 hours weekly with #GenAI coding tools, with no decline in code quality. They measured it through careful benchmarks, not industry hype. #DX #DevEx techblog.cloudkitchens.com/p/study-and-...

⏰ Plus qu’un mois avant le #ForumPHP2025 ! Rejoignez-nous à Disneyland Paris pour célébrer non seulement les 30 ans de #PHP mais aussi les 25 ans de l' @afup.org, les 20 ans de @symfony.com & les 10 ans d' @api-platform.com ! 🎂 On a hâte de vous y retrouver ! 🔗 event.afup.org

Will augmented coding lead to fewer or more programmers? Nobody knows, but some skills matter no matter the outcome: understanding, integration, and judgment. Inspiring thoughts by the great @kentbeck.com. #AI tidyfirst.substack.com/p/programmin...

ARTE 🤝 Marmelab : Depuis 2016, nous avons la chance de collaborer avec ARTE pour développer des briques essentielles de la plateforme ARTE. tv. Aujourd'hui, nous vous emmenons dans les coulisses de ce partenariat ! 😉👇 marmelab.com/blog/2025/09...

📣 Big news: Atomic CRM just got a makeover! It now uses @shadcn.com as part of its updated tech stack. Why @shadcn.com? ✅ You own your code ✅ 100% customizable ✅ Semantic design tokens 👉 Perfect for modern React devs 🔗 Check out Atomic CRM’s new look: marmelab.com/atomic-crm/ #React #OSS #shadcn

The word got out: Shadcn Admin Kit is an open-source #React framework for building single-page-apps with Supabase, react-query, react-hook-form, react-router, and shadcn. #OSS

Friday rant: Ctrl+Q is way too close to Ctrl+S. Nothing like trying to save your work and accidentally rage-quitting your entire IDE instead. Who approved this keyboard layout? Show yourself. 😤

Fascinating read about all the low-level programming techniques used by Bun to outperform (by an order of magnitude) every other #NodeJS package manager. I learned a lot! bun.com/blog/behind-...

Big O notation for beginners, with examples and diagrams. A good reminder of the basics of algorithmic complexity. samwho.dev/big-o/ samwho.dev/big-o/

LLMs can write code, but building software requires more than that. #AI #Programming zed.dev/blog/why-llm...

Are AI coding tools going to replace developers, or supercharge them? In my opinion, AI-augmented assistants like Claude Code and Copilot Agent aren't a threat—they're an opportunity. #AI #SoftwareDevelopment #Programming marmelab.com/blog/2025/09...

Gentle reminder: Never use floating-point data types for storing money amounts #SQL cardinalby.github.io/blog/post/be...

Young professionals between the ages of 22 and 25 in AI-intensive occupations—such as software development or customer service—have experienced job losses of 6 to 13 percent since the end of 2022. #Paper #AI digitaleconomy.stanford.edu/wp-content/u...

Les IA sont au cœur des préoccupations des devs, et ont donc une bonne place au sein du programme du Forum PHP 2025 : @francoisz.bsky.social nous parlera des outils et techniques visant perfectionner les agents IA , et nous apporte un premier éclairage dans l'interview du jour. buff.ly/Dz5q6HJ

Scammers just got superpowers: With Agentic AI browsers like Edge Copilot or Comet, the old phishing tricks fool the agent in no time. Don't us ethem for anything remotely personal. #Security #AI guard.io/labs/scamlex...

I find the new Trello #UI awful. It seems I'm not the only one. Atlassian is engaged in the classic enshittification process that has plagued small apps bought by giant editors in the past 20 years. So sad. www.theregister.com/2025/08/06/t...

[Paper] Google just published a carbon footprint analysis of its flagship #AI assistant, Gemini. It's super low. They report 0.24Wh / req, about the same as my estimate (0.3Wh). In terms of CO2 equivalent, the footprint of an average Gemini request is reported at 0.03g. arxiv.org/html/2508.15...

From zero to 💯! Shadcn Admin Kit just crossed 100 GitHub stars. ⭐ Thanks for the love — we're just getting started! github.com/marmelab/sha...

Mistral AI emitted 20,000 metric tons of CO2 equivalent to train Mistral Large 2, a 128B parameters model. That's a first: No third-party study on the carbon footprint of a state-of-the-art LLM has ever been published before. #AI #CarbonFootprint #Sustainability www.lemonde.fr/economie/art...

If you're building a #React SPA by connecting Tailwind CSS, shadcn/ui, TanStack Query, React Hook Form, and React-Router, you're reinventing the wheel. marmelab.com/shadcn-admin...

Imagine a phone that draws its own interface, just for you. We built a proof-of-concept OS that ditches app silos for a truly generative UI. It's controlled by voice or text and handles links, images, and navigation on the fly. marmelab.com/blog/2025/07...

AI makes experienced open-source devs… slower?!🐢 We might feel more productive with AI—but the data says otherwise: A new METR study found that when devs use AI tools, they take 19% longer than without. Maybe AI’s real superpower is making us believe it’s a superpower. 🔗 metr.org/blog/2025-07...

"Agents are overhyped and often overused. In most real-world applications, simple patterns and direct API calls work better than complex agent frameworks." #AI decodingml.substack.com/p/stop-build...

🆕 New video! We've recently introduced <DataTable> — the powerful new go-to component for rendering tables in react-admin. 💪 📹 And we just dropped a short tutorial to show you how to make the most of it! 👉 Watch now: youtu.be/IHjBd9WrVcI?... #react #ReactAdmin #frontend #opensource

GenAI is probably not ❌ the path to AGI. Why❓ Lack of generalization & hallucinations remain key blockers. Those who pretend otherwise (like OpenAI) have something to sell. Great review of a recent paper by Apple about the limitations of #LLMs👇 🔗 garymarcus.substack.com/p/seven-repl...

[Paper] How common are hallucinations in LLM responses ? A new benchmark named HalluLens tries to give a factual response - and the findings are concerning. #GenAI arxiv.org/html/2504.17...

Do you plan on using shadcn to build your next admin app? Try shadcn-admin-kit, the new open-source #React framework for admins SPAs. It's powered by shadcn ui, Tailwind CSS, React, TypeScript, react-hook-form, TanStack Query, react-router, and react-admin! github.com/marmelab/sha...

Check out another fantastic #CSS guide by Josh Comeau. Ever wonder why height="100%" sometimes doesn't work? What's the best approach to ensure an HTML element always fills the full height? www.joshwcomeau.com/css/height-e...

The new DataTable component in #ReactAdmin offers a better Developer Experience than the previous Datagrid, thanks to the unique strenghts of client-side components. #React marmelab.com/blog/2025/05...

🧵 react-admin 5.8.0 just landed — and it’s packed with goodies! 🎁 New components: DataTable & InPlaceEditor 🔐 AuthLayout: A common layout for auth related pages 🎨 Better theme support ✨ Added support for MUI v7 & Css variables ⏳ Configurable Loading timeout 🔍 Changelog: github.com/marmelab/rea...

@danabra.mov Love your recent blog series about RSC, thanks. I still wonder how React handles what HTTP used to take care of, namely Authentication & Cache. Could you please share your insights on this?

Did you know it only takes 2 lines to change the look of your react-admin app?🤯 Want to see it in action? We just dropped an article showing you how to give your app a Shadcn UI makeover with our new B&W theme. Check it out & give your app a fresh look in no time!😉 🔗 marmelab.com/blog/2025/04...

Reading this article is like reading my own mind… ankursethi.com/blog/the-too...