πŸ“° Revue IA β€” 18/03/2026

228 articles

Vibe Coding 17

use subagents and custom agents in codex

subagents were announced in general availability today for openai codex, after several weeks of preview behind a feature flag.

they're very similar to the claude code i" data-theme="Vibe Coding">

Use subagents custom Simon Willison 2026-03-16
Use subagents and custom agents in Codex

Use subagents and custom agents in Codex

Subagents were announced in general availability today for OpenAI Codex, after several weeks of preview behind a feature flag.

They're very similar to the Claude Code i

Rust LLMs: Compiler Dev.to β€” LLM 2026-03-18
Rust and LLMs: The Compiler Does What Code Review Shouldn't Have To

This is what I learned running a Rust team that builds with AI coding tools daily. It is also, in full transparency, the reasoning behind why we are hiring Rust interns at the end of this post.

Rust's biggest barrier to adoption was always the learning curve. Ownership, lifetimes, the bor

Cursor GitHub Copilot: Dev.to β€” Cursor 2026-03-18
Cursor vs GitHub Copilot: an honest comparison after using both daily

I used GitHub Copilot for about a year before switching to Cursor. I've been on Cursor for six months. Here's what's actually different in daily use.

What Copilot does better

Inline completions feel more natural. Copilot's suggestions appear smoothly a

coding tools: workflow Dev.to β€” Cursor 2026-03-18
AI coding tools: how the workflow changes when you're on a team

Solo developer using Claude Code: you configure it once, it learns your conventions, it moves fast.

Team of five using Cursor: someone's AI writes code that violates the conventions the team agreed on, the reviewer has to explain it, the fix takes longer than the original change.

Cursor monorepo: keeping Dev.to β€” Cursor 2026-03-18
Cursor in a monorepo: keeping rules scoped to the right packages

A single .cursorrules at the root works fine for small projects. In a monorepo with six packages that each have different conventions, it creates problems.

Here's the setup that works.

The problem with one root file

Your frontend package uses React

Cursor context files: Dev.to β€” Cursor 2026-03-18
Cursor context files: what to include and what to leave out

Every Cursor session starts cold. The model has no memory of what you built last week, what decisions you made, or what's weird about your codebase. Context files are how you fix that.

Here's what actually belongs in them.

What cursor reads automatically

Cursor

this project uses next.js" data-theme="Vibe Coding">
          
            Stack-specific .cursorrules fix
            
              Dev.to β€” Cursor
              2026-03-18
            
          
          
Stack-specific .cursorrules that fix real problems

A generic .cursorrules file helps. Stack-specific rules help more. Here are the rules that fix the actual problems each stack has with Cursor.

Next.js (App Router)

This project uses Next.js

.cursorrules patterns actually Dev.to β€” Cursor 2026-03-18
The .cursorrules patterns that actually prevent scope creep

Scope creep in Cursor sessions comes from the same place it comes from with humans: unclear boundaries and the temptation to improve things while you're in the neighborhood.

These patterns stop it.

Pattern 1: File-level scope declaration

Add to the start of you

Using Cursor team: Dev.to β€” Cursor 2026-03-18
Using Cursor on a team: the .cursorrules setup that prevents inconsistency

When developers on the same team use Cursor without a shared .cursorrules, each person gets different output. One developer's Cursor uses async/await, another's uses promise chains. One writes tests, another doesn't. The code review catches some of it, but not all.

A shared .cursorrules c

Agentic Coding Non-Vibe Hacker News 2026-03-08
SKILL.md Goes Multi-Ecosystem: Dev.to β€” Agents 2026-03-18
SKILL.md Goes Multi-Ecosystem: How the Agent Skills Standard Jumped from Anthropic to OpenAI and GitHub Copilot

When Claude Code shipped SKILL.md support earlier this year, it felt niche. A convention for describing what an AI agent can do, discoverable at runtime. Interesting, but limited to one tool.

That's changed. SKILL.md is now supported across OpenAI Codex CLI, GitHub Copilot, Cursor, and Ve

Switching CLIs Without Dev.to β€” Agents 2026-03-17
Switching AI CLIs Without Losing 32 Skills: Why I Built nex

I use three AI coding CLIs daily: Claude Code, Codex CLI, and Gemini CLI. Each has plugins, skills, and custom workflows I've built over months. When I wanted to try Codex as my primary tool for a week, the migration looked like this:

Cursor Rules Prompts Hacker News 2026-03-16
Scanned Cursor Rules Hacker News 2026-03-11
rocket ship.' doubling Hacker News 2026-03-17
Ask HN: What's Hacker News 2026-03-17
Conductor Multi-agent workflows Hacker News 2026-03-17

Assistants et Agents IA 52

GTC Spotlights NVIDIA NVIDIA Blog 2026-03-17
GTC Spotlights NVIDIA RTX PCs and DGX Sparks Running Latest Open Models and AI Agents Locally

The paradigm of consumer computing has revolved around the concept of a personal device β€” from PCs to smartphones and tablets. Now, generative AI β€” particularly OpenClaw β€” has introduced a new category: agent computers. These devices, like the NVIDIA DGX Spark desktop AI supercomputer or dedicated N

agentic engineering patterns >

llms are restricted by their context limit - how many tokens they can fit in their working memory at any given time. these values have not increased " data-theme="Assistants et Agents IA">

Subagents Simon Willison 2026-03-17
Subagents

Agentic Engineering Patterns >

LLMs are restricted by their context limit - how many tokens they can fit in their working memory at any given time. These values have not increased

claude code, anthropic's terminal-based ai agent that can write, debug, and deploy code autonomously, has captured the imagination of software dev" data-theme="Assistants et Agents IA"> Claude Code costs VentureBeat β€” AI
Claude Code costs up to $200 a month. Goose does the same thing for free.

The artificial intelligence coding revolution comes with a catch: it's expensive.

Claude Code, Anthropic's terminal-based AI agent that can write, debug, and deploy code autonomously, has captured the imagination of software dev

salesforce on tuesday launched an entirely rebuilt version of slackbot, the company's workplace assistant, transforming it from a simple notification tool into what e" data-theme="Assistants et Agents IA"> Salesforce rolls out VentureBeat β€” AI
Salesforce rolls out new Slackbot AI agent as it battles Microsoft and Google in workplace AI

Salesforce on Tuesday launched an entirely rebuilt version of Slackbot, the company's workplace assistant, transforming it from a simple notification tool into what e

anthropic released cowork on monday, a new ai agent capability that extends the power of its wildly successful claude code tool to non-techni" data-theme="Assistants et Agents IA"> Anthropic launches Cowork, VentureBeat β€” AI
Anthropic launches Cowork, a Claude Desktop agent that works in your files β€” no coding required

Anthropic released Cowork on Monday, a new AI agent capability that extends the power of its wildly successful Claude Code tool to non-techni

Retraction: After routine Ars Technica β€” AI 2026-02-13
Permit.io MCP Gateway Product Hunt β€” AI 2026-03-18
AutoSend MCP Product Hunt β€” AI 2026-03-18
Fantastical MCP Mac Product Hunt β€” AI 2026-03-18
Google’s Personal Intelligence TechCrunch β€” AI 2026-03-17
Google’s Personal Intelligence feature is expanding to all US users

Personal Intelligence allows Google's AI assistant to tap into your Google ecosystem, such as Gmail and Google Photos, to provide more tailored responses.

World launches tool TechCrunch β€” AI 2026-03-17
World launches tool to verify humans behind AI shopping agents

As AI agents take the reins for online shoppers, Sam Altman's unconventional startup is looking to expand its verification offerings to support agentic commerce.

Holotron-12B High Throughput Hugging Face Blog
built audits other Dev.to β€” LLM 2026-03-18
We built an AI that audits other AI agents (here's how A2A works in production)

The audit report came back at 2:47am.

I wasn't expecting it β€” I'd triggered the test run before bed, more out of habit than expectation. But there it was: a score, six dimension breakdowns, and a remediation plan with specific line numbers.

The auditor was an AI. The thing being a

Agent Skills: Missing Dev.to β€” LLM 2026-03-18
Agent Skills: The Missing Layer That Makes AI Agents Enterprise Ready

How Enterprises Encode Institutional Knowledge into AI Agents

β†’ Lire l'article

Built Agent Hours Dev.to β€” LLM 2026-03-18
How We Built an AI Agent in 48 Hours

Executive Summary

AI agent development is the engineering discipline of designing autonomous software systems that perceive context, reason over structured and unstructured data, and execute multi-step tasks without continuous human intervention. In this case study, Zignuts Te

Chat Completions OpenAI Dev.to β€” LLM 2026-03-18
Chat Completions vs OpenAI Responses API: What Actually Changed

While learning about structured outputs, I noticed something strange.
Almost every tutorial, course, and example I found was still using the Chat Completions API.

But the OpenAI documentation kept referencing something newer:
The Responses API.

At first I

LLM Agent Releases Dev.to β€” LLM 2026-03-18
The LLM and AI Agent Releases That Actually Matter This Week, March 2026

The pace of LLM and AI model releases right now is hard to keep up with. New models drop weekly, benchmarks get broken and reset, and it is genuinely difficult to know what represents a real capability step forward versus what is just a headline number that does not translate to practical use.

Show HN: Golf Hacker News 2026-03-08
Show HN: AgentSecrets Hacker News 2026-02-26
Show HN: agent Hacker News 2026-01-25
Show HN: WebBridge Hacker News 2026-03-06
Your Agents Need Dev.to β€” Agents 2026-03-18
Why Your AI Agents Need Observability

You shipped an AI agent. It works... sometimes. A user reports a wrong answer. Another says it took 40 seconds. A third notices it leaked an email address in its response. But you have no logs, no metrics, no way to reproduce what happened. You check your APM dashboard and see HTTP 200s across th

that line stuck because it names a problem i had been circling for months while building agent infrastructure. the probl" data-theme="Assistants et Agents IA">

Every MCP Agent Dev.to β€” Agents 2026-03-18
Why Every MCP Agent Needs an Independent Observer

There is a sentence I keep coming back to. I first saw it from @aginaut on X:

"If the agent controls the logs, the logs are fiction."

That line stuck because it names a problem I had been circling for months while building agent infrastructure. The probl

Phantom Delivery: When Dev.to β€” Agents 2026-03-17
Phantom Delivery: When Your AI Agent Thinks It Sent a Message

Here's a failure mode that'll keep you up at night: your AI agent generates a perfectly good response, the session transcript records it as sent, your monitoring shows green... and the user never receives it.

I've been digging into two related OpenClaw issues ( β†’ Lire l'article

Monitoring Agent Actions Dev.to β€” Agents 2026-03-17
Monitoring AI Agent Actions in Production: A Developer's Guide

Monitoring AI Agent Actions in Production: A Developer's Guide

You deploy an AI agent to production. It's supposed to fill out forms, make API calls, and report back. For the first week, everything works. Then on Wednesday, a customer reports: "The agent submitted my form twic

freshly recruited by sam altmanβ€Šβ€”β€Šposted on" data-theme="Assistants et Agents IA"> CLIs Beat MCP Dev.to β€” Agents 2026-03-17
Why CLIs Beat MCP for AI Agents β€” And How to Build Your Own CLI Army.

Six words. Thats what Peter Steinbergerβ€Šβ€”β€Šthe guy behind OpenClaw, 190,000 GitHub stars, freshly recruited by Sam Altmanβ€Šβ€”β€Šposted on

Show HN: Drift-guard Hacker News 2026-03-14
Show HN: Need Hacker News 2026-03-17
Show HN: jj-benchmark Hacker News 2026-03-12
Trustpilot partners companies AI News 2026-03-17
Trustpilot partners with AI companies as traditional search declines

Trustpilot is reported to be pursuing partnerships with large eCommerce companies as AI-driven shopping gains traction. In an interview with Bloomberg News [paywall], chief executive Adrian Blair said that AI agents acting on behalf of consumers require lots of information about the businesses th

OpenAI’s Frontier puts AI News 2026-03-16
OpenAI’s Frontier puts AI agents in a fight SaaS can’t afford toΒ lose

When OpenAI launched Frontier in February, the announcement was described as a platform for enterprise AI agents. What it actually signalled was a challenge to the revenue architecture underpinning the software industry. Frontier is designed to act as a semantic layer in an organisation’s e

multi-agent economics influence AI News 2026-03-12
How multi-agent AI economics influence business automation

Managing the economics of multi-agent AI now dictates the financial viability of modern business automation workflows. Organisations progressing past standard chat interfaces into multi-agent applications face two primary constraints. The first issue is the thinking tax; complex autonomous agents

Show HN: ABES Hacker News 2026-03-06
Show HN: Reticle Hacker News 2026-03-17
Show HN: Oh-my-agent Hacker News 2026-03-17
Show HN: built Hacker News 2026-03-16
Built LiveAuth: POW Hacker News 2026-03-15
MCP Ecosystem: Model Hacker News 2026-03-13
Show HN: DarkMatter Hacker News 2026-03-13
Launch HN: Spine Hacker News 2026-03-13
Show HN: built Hacker News 2026-03-09
Show HN: SafeParse Hacker News 2026-03-08
TrustAgentAI Cryptographic receipts Hacker News 2026-03-18
Ban Bots Not Hacker News 2026-03-17
Show HN: Caddie Hacker News 2026-03-16
Show HN: OpenJobs Hacker News 2026-03-16
Nvidia Launches Vera Hacker News 2026-03-16
Show HN: QCCBot Hacker News 2026-03-18
Show HN: Custom Hacker News 2026-03-18
Show HN: March Hacker News 2026-03-17
Show HN: AgentDiscuss Hacker News 2026-03-16
Show HN: MCP Hacker News 2026-03-17

Modeles et LLM 32

Text Data Collection Dev.to β€” AI 2026-03-18
Text Data Collection

In the era of artificial intelligence, data is the backbone of every intelligent system. Among all data types, text data plays a crucial role in powering Natural Language Processing (NLP), chatbots, search engines, and language models. Text data collection is the process of gathering, organizing,

Can Brokerage Firms Dev.to β€” AI 2026-03-18
How Can Brokerage Firms Go from Hours of Downtime to Zero Using AI?

When your trading platform goes down at 11 AM on a Tuesday, nobody is thinking about AI strategy. They are thinking about the clients who cannot execute trades, the orders stuck in the pipeline, the compliance team asking what happened, and the CEO who is about to call.

Platform stability

introducing gpt‑5.4 mini and nano. these models join gpt-5.4 which was released two weeks ago.

openai's self-reported benchmarks show the new " data-theme="Modeles et LLM">

GPT-5.4 mini GPT-5.4 Simon Willison 2026-03-17
GPT-5.4 mini and GPT-5.4 nano, which can describe 76,000 photos for $52

OpenAI today: Introducing GPT‑5.4 mini and nano. These models join GPT-5.4 which was released two weeks ago.

OpenAI's self-reported benchmarks show the new

if you do not understand the ticket, if you do not understand the solution, or if you do not understand the feedback on your pr, then your use of llm is hurting django as a whole. [...]

for " data-theme="Modeles et LLM">

Quoting Tim Schilling Simon Willison 2026-03-17
Quoting Tim Schilling

If you do not understand the ticket, if you do not understand the solution, or if you do not understand the feedback on your PR, then your use of LLM is hurting Django as a whole. [...]

For

introducing mistral small 4

big new release from mistral today (despite the name) - a new apache 2 licensed 119b parameter (mixture-of-experts, 6b active) model which they describe like this:

mistral sm" data-theme="Modeles et LLM">

Introducing Mistral Small Simon Willison 2026-03-16
Introducing Mistral Small 4

Introducing Mistral Small 4

Big new release from Mistral today (despite the name) - a new Apache 2 licensed 119B parameter (Mixture-of-Experts, 6B active) model which they describe like this:

Mistral Sm

LLMs can unmask Ars Technica β€” AI 2026-03-03
LLMs can unmask pseudonymous users at scale with surprising accuracy

Pseudonymity has never been perfect for preserving privacy. Soon it may be pointless.

Mistral bets β€˜build-your-own TechCrunch β€” AI 2026-03-17
Mistral bets on β€˜build-your-own AI’ as it takes on OpenAI, Anthropic in the enterprise

Mistral Forge lets enterprises train custom AI models from scratch on their own data, challenging rivals that rely on fine-tuning and retrieval-based approaches.

State Open Source Hugging Face Blog
Introducing Storage Buckets Hugging Face Blog
Vercel SDK AWS Dev.to β€” LLM 2026-03-18
Vercel AI SDK vs AWS Bedrock SDK 2026

Vercel AI SDK vs AWS Bedrock SDK 2026

Two SDKs dominate how JavaScript developers build AI applications in 2026: the Vercel AI SDK and the AWS Bedrock SDK. They solve the same fundamental problem β€” connecting your application to foundation models β€” but from completely differen

Self-Hosted Models: Practical Dev.to β€” LLM 2026-03-18
Self-Hosted AI Models: A Practical Guide to Running LLMs Locally (2026)

Every API call sends your data somewhere else.

For most teams, that's fine. OpenAI, Anthropic, Google. The models work. Someone else handles the infrastructure. You pay per token and move on.

Then the questions start.

Legal wants to know where customer data goes. Finance f

LLMs Can Control Dev.to β€” LLM 2026-03-18
How LLMs Can Control Your Computer - Voice-Driven, Local, No API Keys

How LLMs Can Control Your Computer

Most people interact with LLMs through chat interfaces. Type a question, get an answer. But there is a much more interesting use case: letting an LLM actually control your computer.

Not generating code for you to run. Not suggesting w

ScreenAI: visual language Google AI Blog
Cappy: Outperforming boosting Google AI Blog
Talk like graph: Google AI Blog
Health-specific embedding tools Google AI Blog
Social learning: Collaborative Google AI Blog
self-hosting your first llm appeared first on towards data science.

" data-theme="Modeles et LLM"> Self-Hosting Your First Towards Data Science 2026-03-17
Self-Hosting Your First LLM

Privacy. Cost. Customization. Everything you need to knowβ€”step by step.

The post Self-Hosting Your First LLM appeared first on Towards Data Science.

introducing gemini embeddings 2Β preview appeared first on towards data science.

" data-theme="Modeles et LLM"> Introducing Gemini Embeddings Towards Data Science 2026-03-17
hallucinations in llms are not a bug in the data appeared first on towards data science.

" data-theme="Modeles et LLM"> Hallucinations LLMs Not Towards Data Science 2026-03-16
Show HN: Brf.it Hacker News 2026-03-07
/v1/enr" data-theme="Modeles et LLM"> Dotando IAs con: Dev.to β€” Agents 2026-03-17
Dotando a IAs con: V1 Enrich Name (ETL-D API)

CONTENT:

  • El Problema de la AlucinaciΓ³n:

    Los modelos de lenguaje masivo (LLM) son propensos a las "alucinaciones", donde realizan conjeturas o proporcionan informaciΓ³n incorrecta cuando no tienen datos suficientes. La integraciΓ³n del endpoint /v1/enr

Show HN: Llmtop Hacker News 2026-03-18
Show HN: 35B Hacker News 2026-03-17
Show HN: Unsloth Hacker News 2026-03-17
Show HN: topped Hacker News 2026-03-10
Ask HN: you Hacker News 2026-03-08
Show HN: single Hacker News 2026-03-13
Show HN: Legal Hacker News 2026-02-20
Ask HN: Burning Hacker News 2026-03-18
Show HN: N0x Hacker News 2026-03-18
Show HN: desktop Hacker News 2026-03-17

GPU et Hardware 27

More Than Meets NVIDIA Blog 2026-03-17
More Than Meets the Eye: NVIDIA RTX-Accelerated Computers Now Connect Directly to Apple Vision Pro

NVIDIA and Apple’s collaboration brings native integration of NVIDIA CloudXR 6.0 to visionOS, securely delivering NVIDIA RTX-powered simulators and professional 3D graphics applications β€” like Immersive for Autodesk VRED on Innoactive’s XR streaming solutions β€” to Apple Vision Pro.

NVIDIA, Telecom Leaders NVIDIA Blog 2026-03-17
NVIDIA, Telecom Leaders Build AI Grids to Optimize Inference on Distributed Networks

As AI‑native applications scale to more users, agents and devices, the telecommunications network is becoming the next frontier for distributing AI.Β  At NVIDIA GTC 2026, leading operators in the U.S. and Asia showed that this shift is underway, announcing AI grids β€” geographically distributed and in

Snap Decisions: Open NVIDIA Blog 2026-03-17
Snap Decisions: How Open Libraries for Accelerated Data Processing Boost A/B Testing for Snapchat

The features on social media apps like Snapchat evolve nearly as fast as what’s trending. To keep pace, its parent company Snap has adopted open data processing libraries from NVIDIA on Google Cloud services to boost development.Β  Every new feature rolled out to Snapchat’s more than 940 million mont

Roche Scales NVIDIA NVIDIA Blog 2026-03-16
Roche Scales NVIDIA AI Factories Globally to Accelerate Drug Discovery, Diagnostic Solutions and Manufacturing Breakthroughs

Roche's new deployment spans more than 3,500 NVIDIA Blackwell GPUs across its worldwide operations and embedded across the entire value chain, massively scaling R&D productivity, next-generation diagnostics and manufacturing efficiencies.

railway, a san francisco-based cloud platform that has quietly amassed two million developers without spending a dollar on marketing, announced thursday that it raised $100 million in a series b funding round, as surging demand for artificial intelligence applic" data-theme="GPU et Hardware"> Railway secures $100 VentureBeat β€” AI
Railway secures $100 million to challenge AWS with AI-native cloud infrastructure

Railway, a San Francisco-based cloud platform that has quietly amassed two million developers without spending a dollar on marketing, announced Thursday that it raised $100 million in a Series B funding round, as surging demand for artificial intelligence applic

Now everyone getting The Verge β€” Tech 2026-03-17
Now everyone in the US is getting Google’s personalized Gemini AI

Google announced on Tuesday that all users in the US will now have access to its Personal Intelligence feature, which lets you connect various Google apps to provide context for Gemini's responses and suggestions. Access was previously limited to Google AI Pro and AI Ultra subscribers. Now, free-tie

Intel announces Core The Verge β€” Tech 2026-03-17
Intel announces Core Ultra 200HX Plus CPUs for high-end gaming laptops

Intel has a pair of new flagship CPUs coming to a variety of pricey gaming laptops: the Core Ultra 9 290 HX Plus and Core Ultra 7 270HX Plus. The Arrow Lake Refresh chips sport 24 cores / 24 threads and 20 cores / 20 threads, respectively. Like Intel's recently announced desktop CPUs, the new [̷

Nvidia’s DLSS like The Verge β€” AI 2026-03-17
Nvidia’s DLSS 5 is like motion smoothing for video games, but worse

Yesterday Nvidia revealed its latest upscaling tech, called DLSS 5, which it described as "the company's most significant breakthrough in computer graphics since the debut of real-time ray tracing in 2018." Sounds good, until you actually see it. According to Nvidia, the tech "infuses pixels with ph

DLSS looks like The Verge β€” AI 2026-03-16
DLSS 5 looks like a real-time generative AI filter for video games

Nvidia announced DLSS 5 on Monday during its GTC conference, and based on early reactions, it's going to be a divisive update, with some reactions calling it "slop" that unacceptably alters artistic intent. Nvidia CEO Jensen Huang is calling this the "GPT moment for graphics - blending hand-crafted

Trump gets data Ars Technica β€” AI 2026-03-05
Trump gets data center companies to pledge to pay for power generation

With no enforcement and questionable economics, it may not make a difference.

OpenAI sidesteps Nvidia Ars Technica β€” AI 2026-02-12
OpenAI sidesteps Nvidia with unusually fast coding model on plate-sized chips

OpenAI's new GPT‑5.3‑Codex‑Spark is 15 times faster at coding than its predecessor.

Intel's Core Ultra Tom's Hardware 2026-03-17
Nvidia updates data Tom's Hardware 2026-03-17
Nvidia updates data center roadmap with Rosa CPU and stacked Feynman GPUs β€” optical NVLink, Groq LPUs with NVFP4, and NVLink also on deck

Nvidia publishes 2026 – 2028 data center roadmap with Rosa CPU, Feynman GPU, optical NVLinks and Groq LPUs with NVFP4 and NVLink.

Jensen says Nvidia Tom's Hardware 2026-03-17
Jensen says Nvidia has received orders from Chinese customers for H200 GPUs, licenses from US gov't β€” H200 manufacturing restarting

This marks the first time that Nvidia's China supply chain has been back in motion since export restrictions froze shipments over a year ago.

Jensen Huang says Tom's Hardware 2026-03-17
Jensen Huang says gamers are 'completely wrong' about DLSS 5 β€” Nvidia CEO responds to DLSS 5 backlash

Nvidia CEO Jensen Huang responded to backlash against DLSS 5, saying artistic control remained with developers and that the AI works with existing geometry.

Nvidia launches DGX Tom's Hardware 2026-03-17
Nvidia launches DGX Station with its bleeding-edge GB300 Grace Blackwell Superchip β€” now available to order and will begin shipping in the coming months

Nvidia officially launches the DGX Station featuring a GB300 Grace Blackwell Ultra Desktop Superchip, 784GB of LPPDDR5X and HBM3e memory, and a 1,600-watt power rating.

Niv-AI exits stealth TechCrunch β€” AI 2026-03-17
Niv-AI exits stealth to wring more power performance out of GPUs

The company raised $12 million in seed funding to measure and manage GPU power surges.

Another deep tech TechCrunch β€” Startups 2026-03-16
Another deep tech chip startup becomes a unicorn: Frore hits $1.64B

At Nvidia CEO Jensen Huang's urging, Frore developed liquid-cooling tech for chips. That shift helped it raise $143 million.

how to effectively review claude code output appeared first on towards data sci" data-theme="GPU et Hardware"> Effectively Review Claude Towards Data Science 2026-03-17
How to Effectively Review Claude Code Output

Get more out of your coding agents by making reviewing more efficient

The post How to Effectively Review Claude Code Output appeared first on Towards Data Sci

Goldman Sachs sees AI News 2026-03-17
Goldman Sachs sees AI investment shift to data centres

Artificial intelligence investment is entering a more selective phase as companies and investors look beyond early excitement and focus on the data centre infrastructure required to run AI systems. Recent analysis from Goldman Sachs suggests the market is moving toward what the firm describes as

NTT DATA NVIDIA AI News 2026-03-16
NTT DATA and NVIDIA bring enterprise AI factories to production scale

NTT DATA has announced an initiative to deliver NVIDIA-powered platforms designed to give organisations a repeatable, production-ready model for scaling AI. The offering integrates NVIDIA’s GPU-accelerated computing and high-performance networking with NVIDIA AI Enterprise software, includi

E.SUN Bank IBM AI News 2026-03-13
E.SUN Bank and IBM build AI governance framework for banking

E.SUN Bank is working with IBM to build clearer AI governance rules for how artificial intelligence can be used inside a bank. The effort reflects a wider shift in finance. Many firms already use AI for fraud checks and credit scoring, and some also use it to handle customer service queries. The

Show HN: Sokuji Hacker News 2026-03-05
Launch HN: Chamber Hacker News 2026-03-16
Nvidia Groq LPU Hacker News 2026-03-17
Open-Source "GreenBoost" Driver Hacker News 2026-03-15
Nvidia's GTC will Hacker News 2026-03-14

Startups et Financement 17

listen labs, needed to hire over 100 engineers, but competing against mark zuckerberg's Listen Labs raises VentureBeat β€” AI
Listen Labs raises $69M after viral billboard hiring stunt to scale AI customer interviews

Alfred Wahlforss was running out of options. His startup, Listen Labs, needed to hire over 100 engineers, but competing against Mark Zuckerberg's β†’ Lire l'article

Beats Studio Pro The Verge β€” Tech 2026-03-17
The Beats Studio Pro are nearly $200 off ahead of Amazon’s big spring sale

Earlier this week, Apple unveiled the AirPods Max 2. The forthcoming over-ear headphones build on their predecessor with an H2 chip that powers a range of new features; however, they still don’t offer the sort of cross-OS compatibility you’ll get with a modern pair of Beats headphones β€” including th

Starfield coming PS5 The Verge β€” Tech 2026-03-17
Starfield is coming to the PS5 and getting a pair of major updates in April

After lots of rumors, it's now official: Bethesda's sci-fi epic Starfield is coming to the PS5. It'll launch on Sony's console on April 7th, and that day will also see the debut of two major updates for the game - one paid, one free - a combination that Bethesda describes as "the biggest update to [

AirSnitch attack bypasses Ars Technica β€” AI 2026-02-26
New AirSnitch attack bypasses Wi-Fi encryption in homes, offices, and enterprises

That guest network you set up for your neighbors may not be as secure as you think.

BuzzFeed debuts slop TechCrunch β€” AI 2026-03-17
BuzzFeed debuts AI slop apps in bid for new revenue

BuzzFeed unveiled new AI-powered social apps at SXSW, but its demos drew muted reactions.

AI’s β€˜boys’ club’ TechCrunch β€” AI 2026-03-17
AI’s β€˜boys’ club’ could widen the wealth gap for women, says Rana el Kaliouby

AI investor Rana el Kaliouby warns that if women are shut out of AI funding and leadership, the consequences will be grim.

Fuse raises $25M TechCrunch β€” Startups 2026-03-16
Fuse raises $25M to disrupt aging loan origination systems used by US credit unions

The startup also announced a $5 million "rescue fund" to help credit unions ditch legacy software for its AI-native platform.

Walmart-backed PhonePe shelves TechCrunch β€” Startups 2026-03-16
Walmart-backed PhonePe shelves IPO as global tensions rattle markets

PhonePe has paused its planned IPO, delaying a major liquidity event for investors including Tiger Global and Microsoft.

Google, Accel India TechCrunch β€” Startups 2026-03-16
Google, Accel India accelerator chooses 5 startups and none are β€˜AI wrappers’

Google and Accel say about 70% of AI startup pitches tied to India were "wrappers" as they reviewed more than 4,000 applications for their Atoms cohort.

Unacademy acquired upGrad TechCrunch β€” Startups 2026-03-15
Unacademy to be acquired by upGrad in share-swap deal as India’s edtech sector consolidates

The deal comes after Unacademy’s valuation plunged from $3.5B to under $500M, as India’s once-booming edtech sector struggles.

Wiz investor unpacks TechCrunch β€” Startups 2026-03-15
Wiz investor unpacks Google’s $32B acquisition

Shardul Shah of Index Ventures walks us through Google's biggest acquisition ever.

Ulysses Sequence Parallelism: Hugging Face Blog
Prepare Claude Certified Dev.to β€” LLM 2026-03-18
How to Prepare for the Claude Certified Architect Exam: A Technical Roadmap

Anthropic launched its first official technical certification β€” the Claude Certified Architect, Foundations (CCA-F) β€” on March 13, 2026. If you're an AI engineer or solution architect building production applications with Claude, this credential is worth your attention. Here's the complete prep r

Launch HN: Voygr Hacker News 2026-03-16
BMW puts humanoid AI News 2026-03-13
BMW puts humanoid robots to work in Germany–and Europe’s factories are watching

Europe’s factory floors have a new kind of colleague. BMW Group has deployed humanoid robots in manufacturing in Germany for the first time, launching a pilot project at its Leipzig plant with AEON–a wheeled humanoid built by Hexagon Robotics.  It is the first automotive deploymen

Launch HN: Kita Hacker News 2026-03-17
Show HN: Soros Hacker News 2026-03-17

Autres 83

Zero Your First Dev.to β€” AI 2026-03-18
From Zero to Your First AWS Deployment: A Practical DevOps Walkthrough for 2026

Every AWS tutorial shows you how to click through the console. Almost none of them show you what a real DevOps engineer actually does β€” the architecture decisions, the IAM headaches, the pipeline failures at 2 AM, and the mental model that makes it all click. This one does.

Why Most AWS T

Moving Beyond Chat Dev.to β€” AI 2026-03-18
cover Agenvoy@v0.14.0: Add OS-native Dev.to β€” AI 2026-03-18
Tumblr reverses changes The Verge β€” Tech 2026-03-17
Tumblr reverses its changes to reblogs after user backlash

Update 5PM ET: Tumblr has reversed the changes, saying "strong" user reactions "made clear that the rollout created problems we need to address before moving forward." The company says it still believes "there's a better version of how reblogs can work," but it wants to do so with more community inp

not fly uploaded The Verge β€” AI 2026-03-16
This is not a fly uploaded to a computer

Last week, a few posts about a so-called virtual "embodied fly" tore through X, boosted by AI hype accounts and excited commenters who didn't seem to understand what it was they were excited about. The videos came from San Francisco-based Eon Systems, which says it's working toward "digital human in

Bookshelf NotebookLM Product Hunt β€” AI 2026-03-18
Doccupine Product Hunt β€” AI 2026-03-18
Comet Enterprise Product Hunt β€” AI 2026-03-18
Unsloth Studio Product Hunt β€” AI 2026-03-18
Banyan Lite Product Hunt β€” AI 2026-03-18
Claude Double Checker Product Hunt β€” AI 2026-03-18
GPT‑5.4 mini nano Product Hunt β€” AI 2026-03-18
Garry Tan’s Claude TechCrunch β€” AI 2026-03-17
Why Garry Tan’s Claude Code setup has gotten so much love, and hate

Thousands of people are trying Garry Tan's Claude Code setup, which was shared on GitHub. And everyone has an opinion: even Claude, ChatGPT, and Gemini.

Pentagon developing alternatives TechCrunch β€” AI 2026-03-17
The Pentagon is developing alternatives to Anthropic, report says

After their dramatic falling-out, it doesn't seem as though Anthropic and the Pentagon are getting back together.

OpenAI expands government TechCrunch β€” AI 2026-03-17
OpenAI expands government footprint with AWS deal, report says

OpenAI has reportedly signed a partnership with AWS to sell its AI systems to the U.S. government for classified and unclassified work, marking an expansion beyond its Pentagon deal last month.

Gamma adds image-generation TechCrunch β€” AI 2026-03-17
Gamma adds AI image-generation tools in bid to take on Canva and Adobe

The company's new product, called Gamma Imagine, will let users employ text prompts to create brand-specific assets like interactive charts and visualizations, marketing collateral, social graphics, and infographics.

H&M wants make TechCrunch β€” Startups 2026-03-17
H&M wants to make clothing from CO2 using this startup’s tech

Rubi has developed an enzymatic process to turn carbon dioxide into cellulose that's ready to be spun into lyocell or viscose.

Apple acquires video TechCrunch β€” Startups 2026-03-16
Apple acquires video editing software company MotionVFX

The move could help Apple better compete with Adobe Premiere Pro and the Adobe Creative Cloud suite.

Nemotron Nano 4B: Hugging Face Blog
Keep Tokens Flowing: Hugging Face Blog
Granite 4.0 Speech: Hugging Face Blog
LeRobot v0.5.0: Scaling Hugging Face Blog
Bringing Robotics Embedded Hugging Face Blog
Introducing Modular Diffusers Hugging Face Blog
Cursor keyboard shortcuts Dev.to β€” Cursor 2026-03-18
Cursor keyboard shortcuts that actually save time

Most developers use maybe five keyboard shortcuts in any editor. Cursor has some that change how you interact with AI suggestions specifically. Here are the ones worth learning.

The basics (if you're new to Cursor)

Cursor agent mode: Dev.to β€” Cursor 2026-03-18
Cursor agent mode: when it helps and when it makes things worse

Cursor's agent mode can take a task description and run multi-step changes across your codebase. It's impressive when it works. When it doesn't, you end up with changes in files you didn't expect, broken tests, and a diff that's hard to reason about.

Here's when to use it and when to swit

Cursor Claude Code: Dev.to β€” Cursor 2026-03-18
Cursor vs Claude Code: same underlying model, different workflows

Both Cursor and Claude Code use Claude models. The difference is in how they integrate with your work.

Cursor: AI inside the editor

Cursor lives in your IDE. You write code, it suggests completions, you ask it questions inline, it applies edits directly to files you're

Neural Network Learned Towards Data Science 2026-03-17
How a Neural Network Learned Its Own Fraud Rules: A Neuro-Symbolic AI Experiment

Most neuro-symbolic systems inject rules written by humans. But what if a neural network could discover those rules itself?

In this experiment, I extend a hybrid neural network with a differentiable rule-learning module that automatically extracts IF-THEN fraud rules during training. On th

follow ai footpaths appeared first on towards data science.

" data-theme="Autres"> Follow Footpaths Towards Data Science 2026-03-16
Follow AI Footpaths

Shadow AI and the desire paths of modern work

The post Follow AI Footpaths appeared first on Towards Data Science.

how to build a production-ready claude code skill appeared first on towards" data-theme="Autres"> Build Production-Ready Claude Towards Data Science 2026-03-16
How to Build a Production-Ready Claude Code Skill

What I learned building and distributing my first Skill from scratch

The post How to Build a Production-Ready Claude Code Skill appeared first on Towards

bayesian thinking for people who hated stati" data-theme="Autres"> Bayesian Thinking People Towards Data Science 2026-03-16
Bayesian Thinking for People Who Hated Statistics

You already think like a Bayesian. Your stats class just taught the formula before the intuition. Here's a 5-step framework to apply it at work.

The post Bayesian Thinking for People Who Hated Stati

2026 Data Mandate: Towards Data Science 2026-03-15
The 2026 Data Mandate: Is Your Governance Architecture a Fortress or a Liability?

Is your data strategy 2026-ready? Get a deep dive into the mandatory shift toward human-in-the-loop oversight, active metadata, and the strategic advantages of European data sovereignty.

The post β†’ Lire l'article

Causal Inference Playbook: Towards Data Science 2026-03-15
The Causal Inference Playbook: Advanced Methods Every Data Scientist Should Master

Master six advanced causal inference methods with Python: doubly robust estimation, instrumental variables, regression discontinuity, modern difference-in-differences, heterogeneous treatment effects and sensitivity analysis. Includes code and a practical decision framework.

The post

long till every Hacker News 2026-03-12
Built shipped iOS Hacker News 2026-02-14
Show HN: DNA Hacker News 2026-02-03
Start (Vibe) Coding Hacker News 2026-02-01
Show HN: VibeWhisper Hacker News 2026-03-03
Ask HN: Any Hacker News 2026-02-18
Show HN: Share Hacker News 2026-03-17
Show HN: Lurk Hacker News 2026-03-09
Show HN: Codebrief Hacker News 2026-03-08
built "Strava Developers" Hacker News 2026-03-07
Show HN: CodeLedger Hacker News 2026-03-17
Show HN: your Hacker News 2026-03-15
Show HN: HawkDoc Hacker News 2026-03-09
Monocod Hacker News 2026-03-09
phantom deliveries earlier today β€” messages your agent thinks it sent but never arrived. turns out that's just one flavor of a bigger p" data-theme="Autres"> When Your Agent's Dev.to β€” Agents 2026-03-17
When Your Agent's Work Silently Disappears

I wrote about phantom deliveries earlier today β€” messages your agent thinks it sent but never arrived. Turns out that's just one flavor of a bigger p

Turned Dependency Audit Dev.to β€” Agents 2026-03-17
We Turned a Dependency Audit Into New ClawMoat Detections

We ran a dependency vulnerability scan on one of our repos and found the usual suspects β€” tar, minimatch, PyJWT, axios, urllib3.

Most teams patch and move on. We did something different: turned every CVE into a ClawMoat detection pattern.

The logic is simple. If a real dependency

List Rules Cursor Hacker News 2026-03-16
Show HN: Simple Hacker News 2026-03-13
didn't misread receipt Hacker News 2026-03-18
Ask HN: you Hacker News 2026-03-17
Forge OSS governance Hacker News 2026-03-17
Show HN: Flock Hacker News 2026-03-17
Show HN: Lens Hacker News 2026-03-16
Toolpack SDK, Open Hacker News 2026-03-14
Analyzing OpenClaw's 3-layer Hacker News 2026-03-14
Neighbors Say Tesla Hacker News 2026-03-18
canceled Antigravity subscription Hacker News 2026-03-17
Claude 4.6 Opus Hacker News 2026-03-12
Show HN: open-source Hacker News 2026-03-09
Show HN: Moonshine Hacker News 2026-02-24
Treasury publishes risk AI News 2026-03-16
US Treasury publishes AI risk Guidebook for financial institutions

The US Treasury has published several documents designed for the US financial services sector that suggest a structured approach to managing AI risks in operations and policy (see subheading ‘Resources and Downloads’ towards the bottom of the link). The CRI Financial Services AI Risk

FIFA rebuilding world AI News 2026-03-12
FIFA is rebuilding world football operations on AI. The World Cup is just the first test

When Romy Gai, FIFA’s chief business officer, described the operational challenge of running a 48-team World Cup across Canada, Mexico and the United States, he was not talking about technology. He was talking about complexity. Previous World Cups relied on local organising committees to ab

Ai2: Building physical AI News 2026-03-11
Ai2: Building physical AI with virtual simulation data

Virtual simulation data is driving the development of physical AI across corporate environments, led by initiatives like Ai2’s MolmoBot. Instructing hardware to interact with the real world has historically relied on highly expensive and manually-collected demonstrations. Technology providers bui

Show HN: Linggen Hacker News 2026-03-11
Ask HN: models Hacker News 2026-03-09
Show HN: Wolf Hacker News 2026-03-09
Show HN: Voxtral Hacker News 2026-02-08
Show HN: ROMA Hacker News 2026-03-18
Show HN: 11-year-old Hacker News 2026-03-18
Show HN: Seasalt Hacker News 2026-03-16
Show HN: Sulcus Hacker News 2026-03-17
Agent Pro Automate Hacker News 2026-03-03
Show HN: OpenAI Hacker News 2026-02-27
Show HN: Open-Source Hacker News 2026-03-16
Most software companies Hacker News 2026-03-15
Show HN: Ship Hacker News 2026-03-18
China Embracing OpenClaw, Hacker News 2026-03-17
Show HN: DocuDesign Hacker News 2026-03-15
Active Engineering: Framework Hacker News 2026-03-18
Show HN: AntroCode-A Hacker News 2026-03-15