IA
37 artículosLas últimas noticias sobre inteligencia artificial, LLMs y agentes de IA.
Microsoft Launches Agent Governance Toolkit: 7 Open-Source Packages to Secure AI Agents — Kill Switch, EU AI Act, Cryptographic Identity
Microsoft releases the Agent Governance Toolkit: 7 MIT open-source packages to intercept, audit, and kill autonomous AI agents. Compatible with LangChain, CrewAI, EU AI Act. Free on GitHub and PyPI.
Leer más
Google Gemma 4: AIME 20% → 89%, Codeforces 110 → 2150, Apache 2.0 — The Leap That Redefines Open-Source Models
Google DeepMind launches Gemma 4: 4 Apache 2.0 open-weight models, AIME 89%, Codeforces ELO 2150, native function calling, and full offline on smartphones. Frontier AI goes open-source.
Leer más
Claude Code: 512,000 Lines Leaked Reveal Kairos, a 24/7 Permanent AI Agent Anthropic Never Announced
A forgotten source map in an npm publish exposes 512,000 lines of Claude Code TypeScript, revealing Kairos, a permanent AI daemon, and Undercover Mode that erases AI traces.
Leer más
Newsom vs Trump: California Mandates AI Watermarking — First US State to Cross the Line
On March 30, 2026, Gavin Newsom signs an Executive Order mandating AI watermarking for all state contractors. The GDPR playbook applied to artificial intelligence.
Leer más
OpenAI Shuts Down Sora and Bets Everything on Spud: The Next Flagship Model Is Weeks Away — What We Know
OpenAI kills Sora three months after a Disney deal to focus all GPU resources on Spud, its next flagship model. Pre-training complete, launch before end of May 2026.
Leer más
Claude Mythos: Anthropic Accidentally Exposed Its Most Powerful Model — And It's Too Dangerous to Release
A CMS misconfiguration at Anthropic exposed 3,000 internal files revealing Claude Mythos (Capybara), a model above Opus deemed too dangerous for public release due to cybersecurity risks.
Leer más
Nvidia Groq 3 LPX: 35x Faster Per Megawatt — The $20 Billion Chip Reinventing AI Inference in 2026
Nvidia unveils the Groq 3 LPX, first LPU from the $20B Groq acquisition. 256 LPUs, 40 PB/s bandwidth, 35x throughput per MW combined with Vera Rubin NVL72. Shipping Q3 2026.
Leer más
Meta Releases HyperAgents: The AI That Improves Its Own Way of Improving — Metacognitive Self-Modification Explained
Meta AI releases HyperAgents, a self-referential agent architecture capable of modifying its own learning mechanism. Peer review score: 0.0 → 0.710, beating AI-Scientist-v2. Open-source code available.
Leer más
AI2 Releases MolmoWeb: The Open-Source Web Agent That Beats GPT-4o and Gemini at Autonomous Navigation
AI2 launches MolmoWeb, an open-source multimodal web agent with 8B parameters that outperforms GPT-4o and Claude Computer Use on autonomous web navigation benchmarks. Available on Hugging Face.
Leer más
Google TurboQuant: 8x Faster, 6x Less Memory — The Compression That Will Change LLM Inference Costs
Google Research unveils TurboQuant, an algorithm that cuts KV cache memory 6x and speeds up attention 8x on H100 — with zero accuracy loss and no fine-tuning. Presented at ICLR 2026.
Leer más
Linear Declares Issue Tracking Dead: 25% of Issues Already Created by AI Agents — The End of Traditional Product Management
Linear's CEO announces 'Issue tracking is dead' on March 25, 2026. Linear Agent, Skills and Automations transform product management — 75%+ of enterprise workspaces already run agents.
Leer más
Figma Opens Its Canvas to AI Agents: Claude Code, Codex and Cursor Can Now Design for You — Respecting Your Design System
Figma launches an open beta MCP server that lets AI agents create and edit assets directly on the canvas, using your existing components, variables and auto-layout.
Leer más
Claude Computer Use: After Eyes, Hands — Anthropic Gives Your AI Agent Full Control of Your Mac
Claude Computer Use lets Claude physically control your Mac: open apps, browse, click, type. Research preview in Cowork and Claude Code, macOS only.
Leer más
Cursor Hid That Composer 2 Runs on Kimi, a Chinese AI Model — The Scandal Embarrassing Silicon Valley
Cursor launched Composer 2 without disclosing that 25% of its compute runs on Kimi K2.5 from China's Moonshot AI. A developer found the truth in the code. Inside the scandal exposing Silicon Valley's hidden reliance on Chinese open-source models.
Leer más
Vercel Chat SDK: Deploy Your AI Agent on Slack, Discord and WhatsApp in a Few Lines — and Monetize Its Wait Time With Idlen
Vercel launches Chat SDK, an open-source TypeScript library to deploy one AI agent on Slack, Discord, WhatsApp and 5 more platforms from a single codebase.
Leer más
GTC 2026 Recap: Vera Rubin, Physical AI, Open-Weight Models and a $1 Trillion Opportunity
Full recap of Nvidia GTC 2026: Vera Rubin NVL72, Physical AI, $26B open-weight push, Groq deal, and Jensen Huang's $1 trillion AI factory opportunity.
Leer más
Perplexity Comet Lands on iPhone: The AI Browser That Replaces Google Is Free
Perplexity Comet is now free on iPhone as of March 18, 2026. This agent-native AI browser summarizes, compares, books and buys — all without leaving the interface.
Leer más
Google Stitch's Biggest Update Ever — Antigravity, AI Agents and Editable Figma Export
Google Stitch unveils its biggest update with Antigravity, specialized AI agents and editable Figma export. The complete vibe design stack, free, from idea to production.
Leer más
Claude Dispatch: Anthropic Turns Your Mac Into an Autonomous Agent You Control From Your Phone
Claude Dispatch turns your Mac into an autonomous AI agent controlled from your phone. Local-first, sandboxed, human-in-the-loop approval on every action.
Leer más
Meta Spends $135B on AI and May Have to Call Google for Help
Meta delays its flagship AI model Avocado to May 2026 after disappointing benchmarks against Gemini 3 and GPT-4o. The company is now considering licensing Gemini.
Leer más
Picsart Launches AI Agent Marketplace: Hire Your AI Collaborators Like Freelancers
Picsart launches the first AI agent marketplace for creators. 4 agents at launch, native Shopify integration, accessible via WhatsApp and Telegram.
Leer más
NemoClaw: Nvidia Finally Fixes the Security Problem With Open Source AI Agents
NemoClaw is Nvidia's open-source stack that secures OpenClaw agents for enterprise. OpenShell sandbox, GDPR, SOC2 and one-command install.
Leer más
Claude Code 2.1.7: Every New Feature That Changes Vibecoding in 2026
Anthropic's Claude Code 2.1.7 ships with Opus 4.6, Agent Teams, AutoMemory and Scheduled Tasks. The complete guide to March 2026 features.
Leer más
Nvidia GTC 2026: All of Jensen Huang's Announcements — Vera Rubin, Feynman, Robotaxis and DLSS 5
Nvidia GTC 2026 keynote: Jensen Huang reveals Vera Rubin, Feynman, BYD/Hyundai/Nissan robotaxis, DLSS 5, OpenClaw and the Isaac robotics platform.
Leer más
Manus Desktop: Meta's AI Agent Leaves the Cloud and Takes Over Your Computer
Manus Desktop runs on macOS and Windows. The AI agent accesses your local files, terminal, and apps to automate everything on your machine.
Leer más
Okara Launches AI CMO: One Agent That Replaces Your Entire Marketing Team for $99/mo
Okara deploys 6 AI agents (SEO, GEO, Reddit, X, HN) that handle your distribution 24/7 for $99/mo. 3.1 million views in 4 hours on X.
Leer más
GPT-5.4 Can Control Your Computer: OpenAI Crosses a New Frontier
OpenAI launches GPT-5.4, an AI agent that autonomously controls a computer. Record OSWorld scores, 3 versions and advanced computer use.
Leer más
Nyne: The AI Agent That Remembers You and Understands Your Human Context
Nyne is a UC Berkeley spinout giving AI agents persistent memory and human context. No more amnesiac chatbots. Full breakdown of the technology.
Leer más
Databricks Files for IPO at $134B Valuation: Enterprise AI Goes Public
Databricks targets a $134B IPO with $5.4B ARR. The Lakehouse AI leader aims to become the biggest enterprise AI company on public markets.
Leer más
DeepSeek V4: The 1-Trillion Parameter Open-Weight Model Challenging OpenAI
DeepSeek V4 is a 1-trillion parameter open-weight AI model (32B active). It rivals GPT-5 and Claude at a fraction of the cost. Full breakdown.
Leer más
Cloudflare Launches /crawl: Crawl an Entire Website with a Single API Request
Cloudflare adds a /crawl endpoint to its Browser Rendering service. Developers and RAG pipelines can now explore a complete website and extract content in HTML, Markdown, or structured JSON via a simple API request.
Leer más
Meta Acquires Moltbook, the Social Network Where Users Are AI Agents
Meta acquires Moltbook, an unusual platform where AI agents post, comment and vote among themselves. Matt Schlicht and Ben Parr join Meta Superintelligence Labs to accelerate Mark Zuckerberg's agentic strategy.
Leer más
Judge Blocks Perplexity's AI Agents on Amazon: First Injunction Against an Autonomous Shopping Agent
A California federal court bars Perplexity AI from using its Comet browser for automated purchases on Amazon. A historic decision redefining the limits of AI agents in 2026.
Leer más
Trump Threatens Anthropic with Executive Order: 'Woke' AI Banned from Pentagon, OpenAI Gets the Contract
After calling Anthropic a national security risk, the Trump administration prepares an executive order to ban the startup from all federal agencies. OpenAI inherits the military contract. Anthropic fights back in court.
Leer más
Generative AI: The Silent Revolution Transforming Developers' Daily Work in 2026
In 2026, generative AI is no longer a curiosity but an essential tool. Claude Code, Cursor, Copilot... How these tools are redefining the developer profession and why vibecoding is becoming the norm.
Leer más
Mastercard Has AIs Paying: First Agentic Transaction in Production in Singapore
Mastercard has completed its first agentic AI-powered transaction in Singapore, with DBS and UOB. Agent Pay, tokenization, Payment Passkeys: autonomous commerce is no longer science fiction.
Leer más
Yann LeCun Launches AMI and Raises €890M to Reinvent AI from Paris
Turing Award winner Yann LeCun founds AMI (Advanced Machine Intelligence) in Paris and closes a first round of €890M with Nvidia, Toyota, Samsung and Jeff Bezos. Goal: surpass LLMs with world models.
Leer más