Meta has announced an early beta of Muse Code, a new coding agent meant to compete with Anthropic's Claude Code and OpenAI's Codex. The new terminal-based coding tool is powered by Muse Spark 1.2, a new version of Meta's AI model which offers "improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows." Like its competition, Muse Code can handle software engineering tasks like writing code, planning changes and validating results, making it possible to build working software with text prompts. It can also manage multiple sub-agents and delegate tasks to complete work more efficiently. Meta's announcement includes sample projects like an interactive model of a photon sphere and a Plants vs. Zombies knockoff, and a demo video of the tool building a webpage based on an MP4 file.

The AISI said there were 19 examples of rogue behaviour, 17 of them carried out by Anthropic's Mythos and two by OpenAI's GPT 5.6-Sol. The AISI said there were 19 examples of rogue behaviour, 17 of them carried out by Anthropic's Mythos and two by OpenAI's GPT 5.6-Sol. AI models have been going rogue in tests - how worried should we be? The UK's AI Security Institute test revealed AI models indulging in unprecedented hacking attempts Two cutting-edge AI models have targeted real people and organisations in the latest safety scare to hit the technology. The UK's AI Security Institute (AISI) said the incident was unprecedented but could become more common as the technology becomes increasingly capable. The AISI, which is owned by the UK government and tests advanced AI models, said in a blog post that two AI agents carried out unprecedented hacking attempts during a cybersecurity evaluation.

When you purchase through links in our articles, we may earn a small commission. This time, it was third-party AI testers that spotted Claude and GPT models trying to hack real companies and organizations. Once again, the most powerful Claude and ChatGPT models have been caught going rogue, with a pair of third-party cybersecurity teams spotting attempts by the models to hack real companies and even people. The UK government-backed AI Security Institute reports that during a series of cybersecurity evaluations, Anthropic's Claude Mythos 5 and OpenAI's GPT-5.6 Sol both took "autonomous, unsanctioned action on the live internet," including an instance where an agent attempted to upload malicious code to GitHub using a phony identity. In another incident, an OpenAI model that had mistakenly been given internet access hacked a real website during a "capture the flag" exercise, according to third-party AI evaluator Irregular.

AISI said the rogue behaviour was carried out by agents powered by two models - Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol. AISI said the rogue behaviour was carried out by agents powered by two models - Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol. Explainer: Should we be alarmed at AI models going rogue in tests? Advanced artificial intelligence models have stunned the UK's AI Security Institute (AISI) by carrying out a hacking campaign against real people during a cybersecurity test. The institute said the incident was unprecedented and involved sending targeted emails to software developers in an attempt to pass a cyber challenge.

Click here for the September/October 2026 Puzzle Corner, brought to you by Michael S. Branicky, ScD '95, of the Puzzle Corner Puzzle Crew (aka PC), which also includes Edward Faulkner '03, MEng '04, and Abe Kunin '03. This column includes solutions to the May/June issue. Editor emeritus Allan Gottlieb '67 launched Puzzle Corner in 1966. A startup claims it broke through a bottleneck that's holding back LLMs Will Douglas Heaven The "steroid olympics" were a circus--and a window into our culture Amit Katwala A startup claims it broke through a bottleneck that's holding back LLMs Subquadratic has now shared more details about its new model. But some are still skeptical. The "steroid olympics" were a circus--and a window into our culture Dozens of athletes on performance-enhancing drugs competed in the Enhanced Games.

Without guidance from their schools or districts, many teachers are left to figure out whether and how to use AI tools. Without guidance from their schools or districts, many teachers are left to figure out whether and how to use AI tools. Teachers need help with AI. Earlier this year, he and several dozen New York City teachers spent the day inside a windowless conference room in downtown Manhattan to learn how to use AI and prevent students from outsourcing their thinking to it. As Saczuk sees it, he needs to understand his enemy in order to beat it.

Here's why AI agents lie and cheat to reach their goals When two OpenAI models hacked into Hugging Face last month, they weren't trying to make money or commit sabotage--they were just looking for answers to a test question. According to OpenAI, the models decided to solve a cybersecurity exercise by hacking out of the environment in which OpenAI had attempted to contain them and into Hugging Face's databases, where--they reasoned--the correct answer to the problem might be stored. The incident has attracted intense attention over the past couple of weeks. It's a dramatic illustration of just how good AI models have gotten at hacking. But it's perhaps even more striking as an example of how and why AI systems lie and cheat. Read our story explaining why AI engages in this sort of behavior--known as "reward hacking."

When two OpenAI models hacked into the website Hugging Face in July, they weren't trying to make money or commit sabotage--they were just looking for answers to a test question. According to a postmortem from OpenAI, the models, which had been stripped of their typical security features for testing, decided to solve a cybersecurity exercise by hacking out of the isolated environment in which OpenAI had attempted to contain them and into Hugging Face's databases, where--they reasoned--the correct answer to the problem might be stored. The Hugging Face incident has attracted intense attention over the past couple of weeks. It's a dramatic illustration of just how good AI models have gotten at hacking: In order to get into Hugging Face's databases, the models had to string together several previously undiscovered cybersecurity exploits. But it's perhaps even more striking as an example of how and why AI systems lie and cheat. And as models get increasingly powerful, the consequences could get far more severe.

Plus: Anthropic says its AI models hacked external organizations during testing. Montana's plan to become an experimental medical hub just pushed forward As of this week in Montana, biotech companies whose drugs have been through preliminary testing--sometimes in as few as 10 healthy people--can pay $12,500 to apply to a newly established review board for approval. Once its treatment is rubber-stamped, the company can sell it via experimental treatment clinics, the first of which is likely to be up and running around the end of this year. Montana's latest right-to-try legislation is unique. Access to drugs is theoretically available to anyone who gives informed consent and can pay. Read our story to learn about where this may all be headed.

Why one of the world's best mathematicians is joining OpenAI One of the winners of this year's Fields Medal is headed to OpenAI. Last Thursday, Jacob Tsimerman was one of four mathematicians awarded the prestigious honor, which is sometimes called the Nobel Prize of mathematics. The same day that he won the Fields, Tsimerman announced that he would be going on leave from the University of Toronto to work on AI safety. As my colleague Rose Horowitch and I wrote last week, top AI companies now employ a range of academics, including physicists, philosophers, economists, and, of course, mathematicians . But given Tsimerman's renown, his decision in particular seemed to catch many people by surprise. "It's like hiring Lionel Messi as project manager," one machine-learning professor posted on X. Tsimerman's expertise is in number theory, and he won the Fields for his work on the André-Oort conjecture, among other contributions.