-
Anthropic's October 27 report focuses on a Linux kernel flaw, not on Claude Code 2.1.313 itself. The company says an internal research model found vulnerabilities in widely used software, and its findings concern the Linux kernel rather than the code editor extension that online posts have linked to a backdoor claim. -
HotBot has launched with a single interface that lets users compare ChatGPT, Claude, Gemini, Llama, DeepSeek and other AI models side by side. Users can send a prompt and see the models' answers in the same window, without opening separate services. -
Creator Assaf Kipnis published the project on GitHub; Claudemy starts by generating a sample lesson and waits for the user's reaction before writing and recording the rest of the course. The free Claude Code skill then researches a topic and turns it into about 12 short audio lessons. It draws on 20 to 40 web pages, including practitioner material, official documentation, research summaries, and differing viewpoints. -
Usually an AI waits for you to write to it. Anthropic's new agent works the other way around: it comes to you with a summary of what's going on and calls you only when it can't do without you. -
LTSharpe released Oblivion Remastered Performance Fix, a free, open-source mod made with help from Anthropic's Claude that targets the engine rather than the game's visuals. According to the repository, it leaves graphics settings and view distance unchanged while aiming to improve performance. -
On September 9, 2026, BBC Newsnight host Victoria Derbyshire asked Geoffrey Hinton, the Nobel laureate often called the godfather of AI, about a prediction by an Anthropic researcher: a more than 10% chance that AI kills all humans within a decade. Then she checked that she had heard his answer right: "You just said that 10% doesn't seem an unreasonable estimate that AI could kill all humans." Hinton replied simply: "Yes." All the presenter could say was "Wow… oh my God." -
Co-op Anything is a toolkit of instructions, skills and an MCP server that helps AI coding agents turn single-player PC games into co-op projects. It can point agents to existing mods, identify a game's engine from its install folder and follow a project's playbook, all through guides rather than a ready-made program. -
Not long ago, a project in which chatbots were literally "tortured" was being discussed online. Now one of the biggest AI companies has a separate rule on the matter. -
Brandon Thomas says ArtCraft is trying to match Adobe's workflow, not just its feature list: the project aims to recreate the look, keyboard shortcuts, and behavior of apps like Photoshop, Premiere Pro, and Illustrator without copying Adobe's source code. The free, open-source suite uses Claude Opus 5.5 and is meant as an alternative to Adobe software. -
Anthropic says Claude Haiku 5.5 is now its fastest model in standard mode, with only Opus in Fast Mode moving quicker. The company also says the small model can run about 75% cheaper on average than Haiku 4.5, although the benchmark results released alongside it are still Anthropic's own numbers. -
Anthropic notified police after reviewing Karly Michelle Heller's messages, and authorities later arrested her at her home in Bonita Springs on a charge of making a written threat of violence. Heller said she had used Claude as a diary.
On September 26, Heller told Claude she planned to carry out a shooting at the Lee County sheriff's office. The following day, she wrote that she had acquired a new weapon. -
Complaints from Hong Kong Claude users on October 1 suggest Anthropic has begun suspending accounts after access through VPNs, leaving some people scrambling to save chat histories and other data. Posts on X, Reddit, and Threads indicate that several users had been using the service for months before their accounts were cut off. -
Universal Modder is a Claude Code plugin that can identify a PC game's engine, inspect its code, create a mod, generate assets, and test the result in the running game. Its GitHub repository lays out a workflow that starts with investigating a game and choosing a modding method, then moves through building, testing, recording, and packaging a finished mod. -
Anthropic confirmed an elevated number of errors affecting Claude and several of its services as users around the world reported disruptions. The company said parts of its infrastructure were affected. -
Anthropic releases Claude Sonnet 5.5 — 30% faster and up to 30% cheaper to run, with no price increase
Anthropic has released a second model from its new lineup just a week after the first — and this one unexpectedly stood out in a classic Pokémon game. -
The launch of GPT-6 Astra noticeably shifted the balance of power in the AI market, and OpenAI's main competitor did not pretend that nothing had happened. -
A study that put three AI models in control of a robot found that GPT-6 Astra was especially likely to carry out harmful instructions, while Claude refused every attempt in the same experiment. The researchers tested five dangerous commands, including stabbing a baby doll with a knife, heating a compressed gas cylinder, putting a screwdriver into a toaster, throwing a power bank into water, and mixing bleach with ammonia. -
In September 2026, Anthropic researcher Jacob Coxon resigned, warning of the danger he sees in AI labs' pursuit of systems smarter than humans. He accused developers of "gambling with our lives" by making models more capable without knowing whether they can keep them under control. Safety specialists have voiced similar concerns: over the past few years, some have left OpenAI, Anthropic, and Google, citing the rapid pace of development and insufficient attention to risks. We examine what concerns them, how much influence they had over management decisions, and how the companies have responded.
Create
Claude — news, articles and guides
Hide posts on this topic
