#llms

The latest AI and developer news about #llms, with the original source and a short summary.

Analysis: LLMs are not superhuman at reading and writing neural network weights

A recent paper's claim that GPT-6 Astra can improve a 0.8B Qwen model by editing weights is dismissed as a task any competent human researcher could perform using steering vectors.

·1·2026.11.03 09:00

Project Zero outlines emergency patching systems for urgent vulnerabilities

The guide highlights feature flags, filtering, and hotpatching as key methods to bypass slow standard update cycles.

·2026.10.07 09:00

300 Agents

Moonshot has released the open-source model Kimi K2.6.

·2026.04.22 01:00

System prompt changes between Claude Opus 4.6 and 4.7

Claude Opus 4.7's system prompt tightens safety, tool-use, and conciseness instructions.

·2026.04.19 12:00

Pelican showdown

Qwen3.6-35B-A3B produced a better pelican SVG than Claude Opus 4.7 on a local laptop.

·2·2026.04.17 02:00