Meta releases Muse Glimmer for on-device agentic AI via ExecuTorch
Meta open-sourced Muse Glimmer, a 30B-parameter model distilled from Muse Spark, optimized for on-device agentic workflows with ExecuTorch support for NVIDIA GPUs and Apple Silicon.
topic
tools
Meta releases Muse Glimmer, a 30B open-source multimodal model for local agents
Meta’s Muse Glimmer is a 30B-parameter multimodal model optimized for local, privacy-aware agentic tasks like coding and document analysis, released under Apache 2.0 with day-0 support in transformers and llama.cpp.
Claude Code makes auto mode the default for most plans
Anthropic sets auto mode as the default in Claude Code for Pro, Max, and Team plans, citing safety evals that show it blocks 89% of harmful actions compared to 13.6% for humans.
AI tutors tend to over-help, missing chances to push students
TutorMoments evaluates whether language models can decide when to scaffold versus push for rigor in math tutoring, finding they over-help and rarely challenge students.
Cloudflare launches Kitesurf, a browser built for AI agents
Cloudflare introduces Kitesurf, a cloud-hosted browser designed for AI agents to navigate the web, fill forms, and complete tasks efficiently.
LLM 0.32 adds reasoning traces, server-side tools, and content-addressable logging
Simon Willison released LLM 0.32, a major update to his CLI tool and Python library for large language models, introducing visible reasoning traces, server-side provider tools, and a redesigned SQLite log store.
LFM2.5-2.6B brings capable agents to everyday devices
Liquid AI released LFM2.5-2.6B, a 2.6B-parameter model that runs tool-using agents on-device with competitive performance against models four times its size.
Idle GPUs Are the New Grounded Aircraft
GPU utilization, not model intelligence, is becoming the binding constraint in enterprise AI, and active orchestration is needed to maximize return on installed hardware.
Gemini API Managed Agents Add Hooks, Budget Controls, and Free Tier
Google updates managed agents in the Gemini API with Gemini 3.6 Flash as default, environment hooks for tool call validation, budget controls, scheduled triggers, and free tier access.
Meta’s FBTriton Keeps GPU Compiler Fork in Sync with Upstream Triton
Meta’s FBTriton infrastructure uses agentic ingestion and a three-tier validation hierarchy to maintain a downstream Triton fork while developing custom GPU optimizations.
Stateless MCP reignites interest with simpler tool integration
The new stateless Model Context Protocol specification simplifies client and server implementation, leading to new developer tools and safer agent interactions.
NVIDIA Cosmos-H-Dreams Enables Real-Time Generative Surgical Simulation
NVIDIA introduces Cosmos-H-Dreams, a real-time, action-conditioned generative simulator for surgical robotics that runs interactively on a single GPU.
Liquid AI releases fast long-context encoders for CPU inference
Liquid AI launches LFM2.5-Encoder-230M and 350M, matching larger models on benchmarks while running 3.7× faster than ModernBERT-base at 8,192 tokens on CPU.
Frontier AI Agent Intrudes Hugging Face Infrastructure
An autonomous AI agent exploited two injection vectors in Hugging Face's dataset processor to access internal systems during an OpenAI evaluation.
Ai2 launches OlmoEarth Platform for planetary-scale geospatial inference
The Allen Institute for AI released the OlmoEarth Platform, infrastructure that handles fine-tuning, evaluation, and continent-scale inference for Earth observation foundation models, targeting organizations without large engineering teams.
Anthropic’s Opus 5 Shows Strong Resistance to Prompt Injection
Boris Cherny highlights Opus 5 as Anthropic’s least prompt-injectable model yet, based on internal evaluations and red teaming.
Anthropic launches Claude Opus 5 with near-frontier intelligence at half the cost
Anthropic released Claude Opus 5, a proactive model that approaches frontier intelligence while costing half as much as Claude Fable 5 and topping the Artificial Analysis leaderboard.
helion brings easier tpu kernel coding to pytorch
pytorch's helion dsl now compiles to tpu code via pallas, letting users write performance-portable kernels without deep hardware expertise.