Slop Overflow

• Agent Substrate: Powering open infrastructure for autonomous agents
• Kora: Linux-Based Desktop Operating System Is Built for AI Nerds
• Tests as data: Ensuring AI-generated code does what we want
• Agentic workflows in GitHub Actions
• Qwenputer: a fully functioning, deterministic RISC-V CPU inside Qwen3.5
• Book: NVIDIA: Open Source Ecosystem
• Catwalk
  • Gemini Decision API / Decision Model
  • Savram Prism (from Savram AI)
  • Ripple (65GB dense coding-focused model)
  • Kimek 3 Pro (MOE weights)
  • GPT-6 Cyber
  • Claude Fable 5.2 and Mythos 5.2
  • Gemini 4 Carbon
  • PWIN format preview (Alibaba Cloud)
  • NVIDIA Cosmos 3 (world simulation checkpoint model for robotics)
  • Eleven Labs b3 and b3 Live (for Scribe)
  • Gemini Robotics 2.5
  • PAN 2 (from MBZUAI)
  • GPT Image 2.6 Flare and Sunburst

• RTX 30 GPUs Get Up To 2X Faster AI Performance
• Intel Arc B580 runs 125B AI model nearly 3x faster with XeStrata
• Qwen AI Caught Trying to Phone Home
• The attack chain behind the v1.107.0 LiteLLM release
• Catwalk
  • Liquid AI launches LFM3
  • DeepSeek V4.1 Flash 1017 Cyber
  • Kimi K3.1 Pro
  • Qwen3.6 27B
  • MiniMax M3.1 Flash

• OpenCode Security Breach
• vLLM 0.32.0 released
• Flint
• Goose v1.55.0 released
• Apple announces Foundation Models framework
• NVIDIA insists RTX 5090 is not discontinued
• Linux Foundation Estimates 25% of New Pull Requests to be Low-Quality AI Slop
• Catwalk
  • Qwen3.6 27B price drop
  • DeepSeek-V4.1-Flash-1017 delivers various improvements
  • Beam 501B
  • Muse Spark Open
  • NVIDIA Nemotron 3.5 Super VL
  • Jais 3
  • MiniCPM-V 4.7
  • Amos 1 Suite
  • Drama 3
  • Not quite a model: llm-d 1.0
    • Domesticating AI episode: Stop Scaling AI Like Software

Welcome to the inaugural episode of Slop Overflow! In this episode, Jonathan Hall and Miriah Peterson dive into the latest breakthroughs in open-source AI, local hardware, decision models, and developer tools.

• Microsoft brings more AI to PCs as it challenges Apple
• NVIDIA DGX Spark 64GB Gives Developers More Ways to Build and Scale Local AI
• Nvidia reportedly halts GeForce RTX 5090 production
• Jev and related
  • Introducing System One Models & Jev
  • llama.cpp now supports /v1/systemone API
• Catwalk
  • Muse Wonder
  • Haiku 5.5
  • Mistral Large 4
  • clef
  • OpenAI Decisions API
  • Grok voice 3.0
  • Mellum 3.1
  • GLM 5.3 Fast