GreyNoise says it found at least 143 compromised devices across 34 countries silently enrolled in the Pawns.app residential proxy network without owner consent. The attacker exploited known flaws including CVE-2024-3272/3273 in D-Link ShareCenter NAS, CVE-2021-44228 in Ubiquiti UniFi Video NVR, and CVE-2025-31324 in SAP NetWeaver Visual Composer, and also targeted Hikvision and Dahua DVR/NVR devices. In some cases SSH keys were added for persistent backup access.
Amazon introduced two new always-listening AI devices: the $75 Bee Frontier, a wrist-worn or clip-on upgrade with improved noise-filtering microphones and up to seven days of battery life, and the $100 Bee Eclipse, a credit-card-sized device that attaches magnetically to a smartphone. Both devices record, transcribe and analyze conversations using AI, continuing the Bee product line Amazon acquired in July 2025. The Bee Frontier launched Thursday, while the Bee Eclipse is said to be coming soon.
Bee, the wearable AI startup Amazon acquired over a year ago, has released two new AI devices along with a Pro subscription tier and software updates. The flagship, the Bee Frontier, replaces the original Pioneer with a smaller, lighter design, more color options, IP67 water and dust resistance, and dual noise-filtering microphones. It retains the same 7-day battery life and USB-C charging as its predecessor, and still lets users record and summarize conversations via a single button press.
Bee, the AI wearable company Amazon acquired last year, launched its first new products since the acquisition: the Bee Frontier, a wrist or clip-on device priced at $74.99, and the Bee Eclipse, a $99.99 card-sized attachment for smartphones. Both devices record and summarize conversations into notes, to-dos, and calendar entries, with free basic features and an $11.99 monthly Pro tier for deeper integrations.
Liquid AI released LFM2.5-24B-A2B, a 24-billion-parameter mixture-of-experts model with only 2 billion active parameters per token, as the largest and final entry in its LFM2.5 family. The model is reasoning-only, generating an explicit chain of thought, and both base and post-trained open-weight versions are now available on Hugging Face and Liquid's Playground.
Google announced new Android security features that use on-device AI to detect scam texts and calls, hide sensitive login fields during screen sharing, and flag suspicious call forwarding. The company said its systems already flag about 5 million suspected scam calls and texts monthly through the Phone and Google Messages apps, with an enhanced Messages protection mode rolling out over the coming weeks.
Anthropic says that by next year, compact AI models may be capable of autonomously self-replicating across smart fridges, vacuum cleaners, Android TV boxes, and other IoT devices. In experiments, the company ran small open-weight models on real and simulated low-cost IoT hardware, using quantization and workload-splitting across devices to get the models running, though it found significant challenges in actually building a working self-replicating malware prototype.
The FBI arrested and the Justice Department charged Abdulrahman Alfozan, a 35-year-old Saudi national and OpenAI product team employee since July 2025, with illegally using a device to intercept oral communications. Investigators say he hid a thumb-sized camera and microphone board inside a TV port in a restricted executive conference room, and left a second device with his wallet and ID in a printer room. OpenAI says it found the first device on September 30, alerted the FBI, and fired Alfozan before his arrest on October 9.
OpenAI discovered a USB recording device wrapped in toilet paper and plugged into a TV in a restricted executive conference room where leadership discussed confidential matters. The FBI alleges the device was planted by Abdulrahman Alfozan, a 35-year-old OpenAI engineer, and CCTV footage shows him later trying to retrieve it. He was arrested and fired on Friday, and faces up to five years in prison and a $250,000 fine if convicted of unlawful recording charges.
OpenAI filed a new court document alleging a former employee, who had previously worked at Apple before a six-month stint at OpenAI, sent over 100 confidential files about an unreleased OpenAI device to his personal accounts on his last day. The filing claims the files included device images, engineering and manufacturing data, supplier information and finish specifications, and that the employee later briefed Apple's senior leadership on OpenAI's plans. Apple told Reuters it has found no evidence supporting these claims.
Federal prosecutors charged OpenAI engineer Abdulrahman Alfozan, 35, a Saudi national, after the company discovered a hidden, encrypted audio recording device in one of its conference rooms. OpenAI said it traced the device to Alfozan using its own models, fired him, and referred the case to the FBI, which arrested him on Friday. The company said it has found no evidence he accessed model weights, source code or customer data.
Apple announced a new Apple TV 4K on Tuesday, cutting the cheaper 64GB Wi-Fi-only model and raising the price of the sole remaining 128GB Wi-Fi/Ethernet version to $300. The update adds a faster chip, Find My support for the remote, built-in far-field microphones for Siri, a dedicated AV1 decoder for higher-quality streaming, and the ability to pair HomePod 2 and HomePod Mini 2 speakers into a surround sound setup.