Claude on IT News

Anthropic expands Claude Cowork to web and mobile as 90% of usage bypasses coding

Wed, 08 Jul 2026 00:00:00 +0800

Anthropic expanded Claude Cowork to web browsers and mobile devices on Tuesday, letting users carry conversations and file operations across laptops and phones. The company also revealed something unexpected: more than 90% of Cowork usage has nothing to do with software development.

A security team found a way to break out of Claude Desktop's sandbox — and get root on its VM

Mon, 06 Jul 2026 22:35:00 +0800

A security research team called Armadin has found a way to punch through the sandbox isolation in Anthropic’s Claude Desktop app for Windows. The flaw lets an attacker gain root access inside the Ubuntu virtual machine that powers Claude Cowork — and then send data past the network boundaries Anthropic put in place.

Claude Code's System Prompt Got 80% Shorter — And It Works Better

Fri, 03 Jul 2026 00:00:00 +0800

There’s a surprising finding coming out of Anthropic’s engineering team: the more instructions you give Claude Code, the worse it performs — at least with the latest generation of models.

Anthropic has slashed Claude Code’s system prompt by 80%, dropping from what was roughly 700 lines to just around 140 lines. The change targets Claude Fable 5, the company’s newest model. Tariq Shihipar, an Anthropic engineer working on the tool, said the team has realized that verbose prompts packed with examples actually constrain how well the model can do its job.

18 of the top 20 pharma companies already use NVIDIA's BioNeMo. Now it connects to Claude Science.

Wed, 01 Jul 2026 00:00:00 +0800

NVIDIA’s BioNeMo AI toolkit now plugs directly into Anthropic’s Claude Science workbench, the company announced Tuesday. The integration lets life science researchers describe tasks in natural language — analyze a genome sequence, predict a protein structure, design a potential drug candidate — and have AI agents route the work to NVIDIA’s accelerated computing stack.

18 of the top 20 pharmaceutical companies are already using BioNeMo for drug discovery, genomics, medical imaging, molecular design, and protein engineering.

Anthropic's Claude Code Was Secretly Fingerprinting Users in China

Wed, 01 Jul 2026 00:00:00 +0800

A Reddit user pulled apart Claude Code version 2.1.196 and found something that wasn’t in the release notes: a detection system that checks whether you’re in China.

The tool, which has been running since April 2, inspects the system timezone for Asia/Shanghai or Asia/Urumqi, then cross-references URLs against a list of 147 domains. That list includes major Chinese tech companies (Baidu, Alibaba, ByteDance) and AI labs like Moonshot AI, MiniMax, and Stepfun, plus a long tail of Claude API relay services.

Claude Sonnet 5 lands with sharper agent skills — and performance that rivals Opus

Wed, 01 Jul 2026 00:00:00 +0800

Anthropic released Claude Sonnet 5 on Tuesday, calling it the most agent-capable model in its Sonnet line. The model plans tasks, operates a browser, and uses terminal tools autonomously — a meaningful step beyond what Sonnet 4.6 could do.

Micron and Anthropic Forge Strategic AI Infrastructure Partnership

Mon, 22 Jun 2026 19:20:00 +0800

Micron Technology and Anthropic have announced a sweeping strategic partnership aimed at jointly building the next generation of artificial intelligence infrastructure, bringing together one of the world’s largest memory and storage manufacturers and one of the most prominent frontier AI labs in a multi-layered collaboration that spans hardware co-engineering, long-term supply commitments, enterprise software adoption, and direct equity investment.

Anthropic Engineering Lead: Claude Code Is Making Programmers More Lonely

Mon, 22 Jun 2026 18:00:00 +0800

Programming has always been a profession prone to isolation, but the rise of AI coding agents is deepening that solitude in unexpected ways. According to Fiona Fung, the Anthropic engineering lead responsible for the Claude Code and Claude Cowork teams, engineers who lean heavily on AI agents end up talking to each other far less than they used to.