Apple announced new restrictions on Full Disk Access in macOS. Apps will need 'very explicit user action' to obtain the permission. Apple says some developers' use of Full Disk Access could expose users' files, email, messages, and browsing history, and that the risk from such access will rise sharply as AI agents become more capable and autonomous.
Meta open-sourced code that lets people build their own hardware for Muse, its new AI agent. Suggested projects include putting Muse on a color e-ink screen for reminders, on an HDMI stick for a big-screen display, or on a small touchscreen as a DIY Muse Charm. Users can connect displays, buttons, sensors, and actuators using off-the-shelf ESP32 development boards or Raspberry Pi with Meta's SDK.
Apple announced additional controls for Full Disk Access in macOS, citing risks increased by AI agents. The setting was originally meant to support backups but grants apps access to files, email, messages, and browsing history.
NVIDIA will release a DGX Spark with 64GB of unified memory this month through Acer, ASUS, Dell, Gigabyte, HP, and MSI. It goes on sale October 23, starting at $4,999, and retains the GB10 Grace Blackwell superchip, DGX OS, and full NVIDIA AI software stack. It supports running models with up to 100 billion parameters locally.
Why it matters: The source gives the 64GB configuration's price, launch date, and two-device clustering option, helping readers assess the hardware threshold for running large models locally.
Hearing-tech startup Legato announced that its AI hearing-assistance glasses, Legato Frames, are now on sale starting at $999 for adults with mild to moderate hearing loss. Hearing technology is built into the temples, and an AI system distinguishes speech from background noise and amplifies voices. A dual-speaker system reduces sound a few inches outside the ear by 99%. All processing happens locally on the glasses; they have no camera or audio recording.
Simon Willison built experimental tool Photo Scrubber with GPT-6 Astra. It detects and blurs faces automatically and removes photo metadata. It uses Google's MediaPipe C++ library compiled to WebAssembly through @mediapipe/tasks-vision and the BlazeFace face-detection model. All processing happens locally.
Huawei WATCH D3 went on sale on September 29 starting at RMB 2,988, or RMB 2,540 after government subsidies. It introduces what Huawei calls the industry's first 'low-interference ambulatory blood-pressure study', combining oscillometry and PPG with four micropump measurements during the day and three at night. It is registered as a Class II medical device with China's NMPA and has European CE-MDR certification. The device is 11.3 mm thick, with a 1.92-inch AMOLED display and 17 health indicators in its upgraded quick health check.
Tencent's operating-system-level personal AI butler Marvis announced a repositioning four months after launch, shifting from “find an app” to “find a service” and explicitly taking on the role of managing users' digital assets over time. A new “Butler Team” appears in the sidebar; its PC Butler, co-developed with Tencent's Enterprise IT “8000” team, offers a Skill catalog covering 27 real-world operations scenarios across four categories.
The new Surface PCs Microsoft announced this week no longer carry the Copilot+ PC branding. Surface vice president Brett Ostrum said the new devices “aren't called Copilot+ PCs,” though they still meet every requirement previously associated with the label.
Google's Project Suncatcher is taking its first step: the initial minimum viable product satellite is scheduled to launch on October 1 to test the idea of an orbital AI data center. About the size of a refrigerator, it carries four Google-designed TPU AI accelerators and solar panels providing around one kilowatt of power. Google supplies the AI hardware, while satellite-imaging company Planet Labs provides the spacecraft.
🚀 Tencent Hy Translation just landed.Powered by Hy-MT2. 33 languages. 5 Chinese minority languages & dialects. Voice. Photo. Full offline — on-device, no network required. Already live in 12 countries and regions. Travel, drive, work or read abroad. Accurate. Natural. Always available. ⏬ ⏬⏬ https://apps.apple.com/us/app/tencent-hy-translate/id6801428074
Jun Kim, creator and maintainer of oMLX, has joined Hugging Face full-time to build the MLX community. oMLX will transition from a side project to a funded, long-term effort, remain under the Apache 2.0 license, and continue under Kim's leadership, with a focus on improving stability and speeding up development.
Hugging Face has added GGUF support to Transformers. Users can select a GGUF checkpoint from the Hub and pass `gguf_file` to `from_pretrained` to generate locally, with no extra setup.
Why it matters: Hugging Face's official post explains how to load GGUF directly in Transformers and shares comparison results, helping readers judge whether to move their local-inference workflow.
Meta says its personal agent Muse is now available on Mac. With the user's explicit authorization, it can perform tasks directly on the computer, such as organizing the Downloads folder, finding missing files, and summarizing messages and notes. More capabilities are planned. Download: http://ai.meta.com/muse/download/.
Google Research has proposed PhotoScan, a deep-learning framework that estimates body-fat percentage, the android-to-gynoid fat ratio (A/G), and the visceral-to-subcutaneous fat area ratio (V/S) from ordinary 2D phone photos. It was pretrained on more than 35,000 UK Biobank records, fine-tuned on a cohort of 677 adults, and validated on an independent cohort of 132 people.
Google DeepMind has released SL2T, a multilingual sign-language-to-text model, bringing sign-language AI to consumer products for the first time. On Pixel 11, Gboard and Live Transcribe support dictation from American Sign Language (ASL) to English; more devices and languages are planned.
Why it matters: The source describes the capabilities and limits of a multilingual sign-to-text model and where it is available, helping readers assess sign-language translation in real products.
Google Research and Google DeepMind have introduced SensorFM, a large sensor foundation model that learns directly from unlabeled wearable data. Its pretraining corpus contains more than one trillion minutes of multimodal sensor signals from five million consenting participants across more than 100 countries, all 50 U.S. states, and over 20 Fitbit and Pixel Watch devices.
Why it matters: The article reports pretraining-data scale, scaling experiments, and downstream-task comparisons, helping readers assess how generalizable wearable foundation models are.
Mistral AI has released Robostral Navigate, its first embodied-navigation model. The 8B model uses a single standard RGB camera, with no LiDAR or depth sensor, and achieves a 76.6% success rate on R2R-CE validation unseen—9.7 percentage points above the best single-camera approach and 4.5 points above the best systems using depth or multiple cameras.
Why it matters: The source reports navigation success using a monocular RGB camera and describes the training approach, giving a view of how sensor requirements for embodied navigation are changing.
Google Research has introduced an architecture that attaches a multi-token-prediction (MTP) head to the frozen Gemini Nano v3 model, speeding up on-device generation without changing the backbone weights.
Why it matters: Google attaches an MTP head to frozen Gemini Nano v3 and reports concrete methods and gains under on-device memory and energy constraints.
Google DeepMind has released Gemma 4 12B, an agentic multimodal model aimed at laptops, positioned between the edge-focused E4B and the 26B MoE. It is the company's first medium-sized model with native audio input.
Why it matters: Gemma 4 12B removes multimodal encoders and feeds visual and audio input directly into the LLM, giving readers a view of hardware requirements for local multimodal agents.
In a paper published in Nature, Google Research describes PHRM, a system that uses the front-facing camera to record a few seconds of video after face unlock and estimate heart rate and resting heart rate in the background using deep learning.
Why it matters: Google passively measures heart rate with a phone's front camera after unlock and has released its largest phone-video dataset to date along with a pretrained model.
Google has introduced zero-trust aggregation for its privacy analytics service, combining a new cryptographic aggregation method with hardware-protected TEEs. It can receive device data in a single message without multi-round online interaction. Google receives only anonymous, aggregate insights about groups; raw data is neither exposed nor reconstructible, even within the hardware-protected environment. TEE attestation provides participants with proof that the protocol ran as intended.
Google DeepMind has released the open Gemma 4 family in four sizes: E2B, E4B, 26B MoE, and 31B Dense, under the Apache 2.0 license.
Why it matters: Gemma 4 spans edge devices to workstations in four sizes and reports Arena rankings under Apache 2.0, helping readers assess the practical range of open models.
Mistral has released the Mistral 3 family: dense small models at 3B, 8B, and 14B, plus Mistral Large 3, a sparse MoE model with 41B active and 675B total parameters. All are open-source under Apache 2.0.
Why it matters: Mistral has released Apache-2.0 weights ranging from 3B to 675B, giving readers a view of open-model options from edge devices to frontier scale.
Mistral AI has released Mistral Small 3.1, improving text performance and multimodal understanding over Mistral Small 3. It extends the context window to 128K tokens, runs at 150 tokens per second, and is open-source under Apache 2.0.
Why it matters: Mistral reports comparisons for a small model across text, multimodal, and long-context tasks, helping readers assess on-device deployment options.
Mistral AI has released Mistral Small 3, a 24B-parameter model optimized for low latency and open-sourced under Apache 2.0. It provides both pretrained and instruction-tuned checkpoints.
Why it matters: A 24B model under Apache 2.0 is positioned against models three times its size, with quantization details for single-GPU local deployment to help readers assess its fit.
