Lumma Stealer, a notorious infostealer that infected nearly 395,000 Windows computers last year, has resurfaced with new, hard-to-detect attacks, according to a recent report. The malware, also known as Lumma Stealer, first appeared in Russian-speaking cybercrime forums in 2022 and utilizes a cloud-based malware-as-a-service model to distribute itself through lure sites offering pirated content.
The original infrastructure of Lumma was hobbled by law enforcement authorities around the world last May. However, researchers have confirmed that the malware is back at scale, pilfering credentials and sensitive files. The malware's return highlights the persistent threat posed by cybercriminals and the evolving tactics they employ.
In other technology news, Anthropic released its Claude Cowork AI agent software for Windows on Monday. This move brings the file management and task automation tool to approximately 70 percent of the desktop computing market. The Windows launch offers "full feature parity" with the macOS version, including file access, multi-step task execution, plugins, and Model Context Protocol (MCP) connectors. Users can also set global and folder-specific instructions for Claude to follow in every session, a feature that developers on Reddit described as "a game-changer."
Meanwhile, researchers at MIT, the Improbable AI Lab, and ETH Zurich have developed a new technique to allow large language models (LLMs) to learn new skills without forgetting their existing capabilities. Their technique, called self-distillation fine-tuning (SDFT), enables models to learn directly from demonstrations and their own experiments. Experiments show that SDFT consistently outperforms traditional supervised fine-tuning (SFT) while addressing the limitations of reinforcement learning. This development could help enterprises fine-tune LLMs for new tasks without the risk of breaking what the models already know, which often forces companies to maintain separate models for every skill.
In the realm of open-source AI, NanoClaw, a lighter and more secure version of the AI assistant OpenClaw, has been released. OpenClaw, developed by Peter Steinberger, has seen rapid adoption in recent weeks, offering a powerful means of autonomously completing work and performing tasks across a user's computer or phone. However, its "permissionless" architecture raised security concerns. NanoClaw, which debuted under an open-source MIT license, addresses these issues.
Finally, in the world of sports, the Crave show "Heated Rivalry" is bringing new fans to hockey. The show, which follows closeted pro hockey-rivals-turned-lovers Shane Hollander and Ilya Rozanov, has seemingly sent the entire world into mass psychosis. HBO, which acquired the show for US distribution, is now playing it in well over a dozen countries and says viewership has more than doubled.
Discussion
AI Experts & Community
Be the first to comment