Fetching from the wire…
Skills2026-06-24 · source-backed
Compaction trims what's already in the window (Anthropic productized a server-side compaction API on Opus 4.6 that auto-summarizes older turns mid-session). Agentic memory moves information entirely outside context as structured notes the agent writes and pulls back on demand. Conflate them and your long-horizon agent either loses critical state to summarization or drowns in context it should have offloaded (source).
Each link below shares sources, entities, or timing with this story.
Context engineering just got its cookbook. Anthropic's Claude Cookbook published a complete guide to three API-level primitives that, used together, reduced a research agent's peak context from 335K tokens to a sustained 50-80K range. Server-side compaction. Tool-result cleari...
The Compaction API enables server-side context summarization for effectively infinite conversations without context limits, available on Opus 4.6. MCP servers can now request structured input mid-task — interactive dialogs with text fields, dropdowns, and checkboxes — via new...
Fable 5.1 writes short progress updates between tool calls, each arriving as its own thinking block, but under the default omitted those come back empty and a long turn looks silent to users (Anthropic). The new option (beta header thinking-display-updates-2026-08-18) returns...
The announcement describes the same underlying model at two safeguard levels: Fable generally available, Mythos restricted to vetted cybersecurity and life-sciences organizations, currently US-only. Fable 5.1 scores 52.6% on Terminal-Bench-Science 0.1 (against 24.7% for Fable...
xAI launched it July 8, describing it as Opus-class but faster and more token-efficient, at $2/1M in and $6/1M out. Trained across tens of thousands of NVIDIA GB300 GPUs with RL over hundreds of thousands of multi-step software engineering tasks, and trained *alongside Cursor*...
A solo Claude Opus 4.5 agent spent $9 and 20 minutes building a retro game. It was broken. The same model, wrapped in Anthropic's multi-agent harness, spent $200 over 6 hours and produced a fully playable game with physics, sprite editors, and AI integration. Anthropic's engin...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.