Claude Code 2.1.323 Ships Code Mode: Batch Tool Calls in JavaScript
Claude Code 2.1.323 adds experimental Code Mode — one JavaScript program calls multiple tools in a single pass. Here is what changed, how ...
AMD EPYC 9006: 256 Cores, 1.8x LLM Throughput for Agents
AMD's 6th Gen EPYC Venice ships with Zen 6 and Zen 6c cores, 1.6 TB/s ...
Plannotator: Review AI Agent Plans Before They Run (11K Stars)
Plannotator is a free open-source browser workspace for reviewing AI coding agent plans and code ...
Claude Fable 5.2: 20% Cheaper and the New #1 Coding Model
Anthropic's Claude Fable 5.2 scores 49.1% on FrontierCode 2.0, beats Opus 5.5 on coding at ...
Neo4j Relate: Embed a Graph Database in Your AI Agent
Neo4j shipped Relate on November 11 — an embedded, agent-native graph database written in Rust ...
RTK: Cut AI Coding Agent Token Usage by 60–90%
RTK is a single Rust binary that intercepts CLI output before your AI coding agent reads it, compressing it by 60–90% on verbose ...
Google Antigravity Orchestrator: Two Models, 24% Less Quota
Google Antigravity's new Orchestrator feature automatically swaps between Gemini 4 Carbon and Hydrogen mid-task — ...
90% of AI-Written Unit Tests Are Useless — Here’s the Fix
Momentic audited 1,000 AI-written unit tests. Only 9.6% were useful. Here’s why AI coding agents ...
Gemini Robotics 2.5 Runs Humanoids on a Mercedes Line
Google DeepMind deployed Apollo 2 humanoids on a live Mercedes-Benz factory floor with 96% success. ...
Runway Praxis-1 Open Weights: 27B Robot Brain, Apache 2.0
Runway released Praxis-1, a 27B open-weight world action model for robot control, under Apache 2.0. ...