The 500ml and 2.9 Wh figures come from real papers that said something narrower. Here is every published measurement of AI energy and water per query, with its scope.
MLPerf Inference v6.1 added an Edge Agentic benchmark. NVFP4, FP8 KV cache, 96% cache reuse and tree MTP explain the 6.4x gap over the llama.cpp reference.
Isaac Lab 3.0 flips quaternions to XYZW, replaces torch tensors with ProxyArray, and runs without Isaac Sim. Every breaking change, and how to migrate.
Postgres 19 brings REPACK into core. How it compares to pg_repack and VACUUM FULL on locking, disk headroom, and failure recovery - plus what got cut from 19.
Gateway API v1.6 graduates TCPRoute and UDPRoute to Standard and v1. What that means for MQTT, Postgres and DNS traffic you used to route with vendor CRDs.