Desk
Broad artificial intelligence coverage.
Liquid AI’s d1-ultra returns calibrated probabilities in one forward pass — 65 on Artificial Analysis’s Decision Index at $0.042 per million input tokens.
MarkTechPost’s wire on Muse Spark Open: 816B MoE, 42B active, 1M context — Meta’s open multimodal push lands with MCP-Atlas bragging rights.
Reflection’s Beam lands on Hugging Face: 501B MoE, 23B active, 1M context, Apache 2.0 — with vLLM, SGLang, and a Rust harness.
ChatGPT’s Chrome extension can read your open tabs and — in Work — drive the browser like a junior assistant sitting in your sidebar.
WIRED’s Uncanny Valley: Muse will sext, a pro-AI super PAC is moving money through LLCs, and Apple’s HomeView underwhelms.
A Truth Social “before and after” of a White House column carries OpenAI’s SynthID watermark — politics meeting generative edits in plain sight.
Meta’s Muse Spark Open — 816B parameters, 1M context — ships as a Mac-ready open-weight coding and agent model, if you’ve got a terabyte of RAM.
MarkTechPost’s take on Nemotron 3.5 Super VL: 120B open multimodal MoE, 12B active, 1M context, NVFP4 weights aimed at agents.
Anthropic is embedding six METR researchers with employee-level access — and giving up the right to kill their public assessments.
Meta’s Forum app was a testbed for human-on-the-loop agents — success measured as verified production changes, not lines of code.
OpenAI’s Rosalind Academic Plan: up to 30 free Business seats for a year and trusted-access GPT-Rosalind for eligible U.S. university labs.
Zuckerberg and Apple’s John Ternus are publicly cheering Muse Spark Open on dual M5 Ultra Mac Studios — rivals playing nice for on-desk open weights.