Explore BIG CHANGE reporting on AI Models & Tools: what changed, how it works and why it matters.
Xiaomi researchers released an audio-language model for spatial questions. Its preprint reports stronger directional benchmark scores with two-channel input, while complex scenes remain difficult.
Meta released Muse Spark Open weights on October 15. Its BF16 and NVFP4 checkpoints, hosted APIs and Muse Model License give developers several routes with distinct limits.
Mistral's Timestral package produces 21 quantiles per forecast step. This documentation-only guide covers local setup, output axes, covariates and a held-out comparison.
Google traced one peptide candidate through sequence, structure and genomic checks with Antigravity. Its predicted activity still awaits lab results.
The hosted Voice Agent API now uses Grok Voice 3.0. An independent simulated support score rises, while first-audio time increases and the listed voice price has a separate text charge.
A new Science Robotics paper details the camera, IMU and onboard control behind a 2025 drone racing win. The result is bounded to a known competition course.
Intel's new Inference Microservices select a Xeon serving profile for one model per container. Developers still need to secure the HTTP endpoint and test their workload.
DeepMind and Singapore's GovTech describe a two-stage phone-based scam detector: quick message screening followed by Gemma conversation analysis. Its rollout and results are unknown.
Meta is rolling out topic controls and more detailed post feedback for Facebook Feed and Reels. This documentation-only guide explains the available steps and what remains untested.
Physical Intelligence’s Pi API beta lets selected robot teams upload task data, fine-tune a model on managed infrastructure and use hosted inference. Teams still need their own robots, demonstrations and repeated evaluation.
Liquid AI's new LFM2.5-24B-A2B has downloadable checkpoints for local and GPU serving. Its license and separate edge and server benchmarks matter when evaluating it for tool-using agents.
Anthropic's internal Scion system moves source-control history, commits and merges to a service while agent and CI clients fetch only the code they need.
Shopify's new Shop MCP connects catalog search, merchant carts and checkout. This documentation-based guide shows the setup, shopper approval step and order verification boundary.
REA maps modules and Electron boundaries in a local app. This guide shows how to trace one feature through its evidence and check the behavior separately.
DOJ charged an OpenAI engineer after a recorder was found in a conference room. OpenAI says its models helped identify an employee, but has not explained how.
A practical sequence for testing one Spark UDF, choosing SQL, cuDF or CUDA, checking for CPU fallback, and measuring the result on a GPU-capable pipeline.
Adults 18 and over in select markets can privately launch AI Studio Android apps through Play with a standard Google account. Sharing is planned; public release requires a developer account and review.
TESS has an active target for TIC 4206066. New observations will test a researcher's preregistered transit predictions, while the planet interpretation remains unconfirmed.
