Wazafak — All Articles

The emergence of GLM 5.2, a high-quality open-weights model, threatens to disrupt the AI industry's profit margins. With costs potentially 50% lower than frontier labs, businesses may soon face a margin collapse.

Researchers at Anthropic have discovered a 'global workspace' in language models, which could have implications for AI safety and interpretability. The workspace, dubbed the 'J-space', is a small collection of internal neural patterns that play a special role in the model's processing.

A hacker's arrest reveals Microsoft can track users via a Windows device ID, sparking privacy concerns about Windows' surveillance capabilities.

The European Parliament has passed the Chat Control regulation, allowing for the scanning of private messages for child sexual abuse material. The move has been met with criticism from civil rights activists and IT security researchers, who warn of mass surveillance and unacceptably high error rates of AI scans.

A hidden authentication backdoor in Tenda router firmware grants admin access, potentially allowing attackers to gain control over devices. The issue, tracked as CVE-2026-11405, affects multiple firmware versions and can be exploited to bypass normal login checks.

Apple has sued OpenAI, accusing the company of stealing trade secrets. The lawsuit alleges that OpenAI's chief hardware officer, Tang Tan, and a former Apple employee, Chang Liu, misappropriated Apple's confidential information.

The QuadRF, a phased-array radio built around a Raspberry Pi 5 and an FPGA board, can spot drones and see WiFi through walls, raising concerns about security and surveillance. Its advanced signal processing and beamforming capabilities make it a powerful tool for both legitimate and malicious uses.

Noma Labs discovered a critical prompt injection vulnerability within GitHub's new Agentic Workflows, allowing an unauthenticated attacker to silently pull data from private repositories. The vulnerability, named GitLost, highlights the security challenges of agentic AI systems.

GPT-5.6 Sol Ultra has produced a proof of the Cycle Double Cover Conjecture, a 50-year-old problem in graph theory. The conjecture asserts that every bridgeless undirected graph has a collection of cycles that covers every edge exactly twice.

Boko Haram uses frontier AI to aid in combat and day-to-day operations, posing a growing threat to global security. The terrorist group has successfully circumvented some AI safeguards, utilizing tools like ChatGPT and Meta AI to plan attacks and design explosive devices.

A Uniqlo t-shirt features an obfuscated bash script that has been decoded, revealing a Peace for All message. The script uses base64 encoding and eval to print a happy Easter egg message.

A use-after-free vulnerability in OpenBSD allows local privilege escalation to root, posing a significant security risk. The vulnerability, identified as CVE-2026-57589, affects OpenBSD versions through 7.9.