All news
11343
Claude Cowork and Chrome: 10 ways to hand off browser busywork to an AI agent
A practical Habr guide shows how Claude Cowork with a Chrome extension can take over data lookup, forms, price monitoring, QA and other repetitive browser tasks.

Microsoft and NVIDIA Pitch Physical AI as Industry's Next Edge
MIT Technology Review describes a new phase of automation: physical AI helps factories not just cut costs but grow despite labor shortages and rising production…

Anthropic explains how businesses can adopt agentic systems without needless complexity
Anthropic says businesses should start with simple workflows, not autonomous agents: they are cheaper, more reliable and often more useful for marketplaces.

OpenAI, SoftBank and Oracle Clash Over Stargate as the Data-Center Megaproject Stalls

AWS and vLLM roll out P-EAGLE, speeding up large LLM inference by up to 1.69x

Elon Musk shakes up xAI: layoffs, consultants from Tesla and SpaceX, and a bet on AI data centers

Garry Tan releases gstack, a workflow system for Claude Code with QA, review and release modes

US Drops Plan to Tie AI Chip Sales to Mandatory Data Center Construction
Washington has not backed the idea of selling American accelerators only to countries and companies willing to build data centers in the US, leaving the market more…

Google overhauls Maps: Gemini answers questions, and navigation goes 3D
Google has shipped its biggest Maps update in a decade: Gemini-powered Ask Maps handles complex everyday queries, while Immersive Navigation shows your route in…

Seldon Vault Turns Asimov's Psychohistory into a Multi-Agent AI Forecasting Service

The Lancet Psychiatry warns: AI chatbots can amplify delusions in vulnerable people

Elon Musk Rebuilds xAI Again: Co-Founders Leave, Layoffs Begin and Grok Falls Behind

Anthropic and ETH Zurich: A Long CLAUDE.md Hurts Agent Performance and Raises Costs

AMD RX580 runs LLMs at home: beating ROCm and Ollama to get real GPU inference
An engineer got an old AMD RX580 running LLMs reliably via ROCm and Ollama, unpacking false GPU signals, hipMemGetInfo failures, and Vulkan's role in diagnostics.

Why Claude 4.6 Can't Save You Without Context: The Biggest Blind Spot in LLM Development
Even a strong model like Claude 4.6 loses effectiveness without systematically assembled context: a knowledge base, links between services and environment rules…









