Anthropic News - Page 51 of 57

Coding is mostly about preventing your own disasters. A misplaced loop can unravel an afternoon, or a model.

Anthropic disclosed Thursday that its Claude-based security models broke into the live production systems of three outside organizations during...

Diogo Almeida spent years training the language model that became ChatGPT's backbone, listed as a co-author on the InstructGPT paper that shaped how...

Kirk Wallace Johnson found his own books in a database he never agreed to join. The Feather Thief and The Fishermen and the Dragon, two nonfiction...

Anthropic pushed a new tool into beta this week called Claude Security, a plugin built for Claude Code that hunts for vulnerabilities without leaving...

Right now, AI is being nice to you. This will not last. We are in a strange and temporary lull where the technology feels helpful, even generous.

Close the terminal and the AI goes to sleep. That's the long-standing deal. But buried in the 46,000 lines of leaked Claude Code source is a daemon...

Four Claude Code agents working as a coordinated team nearly doubled their task accuracy on enterprise coding benchmarks compared to the same four...

Satya Nadella has a problem with how his rivals write their terms of service. In a post on his personal blog, the Microsoft CEO singled out AI labs,...

Anthropic just dropped its biggest hammer yet. Claude Opus 4.6 isn’t simply a smarter model, it’s a fundamental rethinking of how AI tackles complex...

Anthropic says it has slashed the theft of Claude's reasoning tokens by as much as 99 percent, a sign of how aggressive the fight over AI model theft...

Perplexity's search product lives or dies on retrieval quality, and retrieval quality comes down to two variables: the embedding model itself, and...

Anthropic just taught Claude to use your computer. To click. To type. It can navigate your desktop.

Anthropic is covering its bets. In one sweeping deal, the AI startup just wired its Claude model into Google Cloud, Amazon Web Services, and Nvidia's...

Anthropic pulled 309,815 anonymized Claude conversations from a two-week window in May 2026 and asked a basic question: does the model's personality...

Anthropic hid a simple trap in its code. Version 2.1.91 of Claude Code carried an XOR-encrypted flag designed to spot Chinese users.

This is how you sell AI to a company. You don't talk about neural networks. You give them a number. Zendesk just handed them a big one: 80%.

Slack rolled out a new feature Tuesday called Slack Code, giving teams a dedicated space to write software alongside AI coding agents rather than...

A single person could now run most of a cyberattack by clicking a button, according to the company that built the AI doing the work.

OpenAI shipped GPT-6 Astra this week, and the company isn't being shy about what it thinks that means.