Fetching from the wire…
01
02
03
04
05
06
07
08
Agents2026-09-16 · source-backed
A comparison of five trimming strategies found recency, relevance and summarization all hit about 60% token savings while dropping task success to 66.6-77.3% and protocol adherence to 85.5-88.6%. Protocol-aware trimming reaches 92.2% success; adaptive budget guardrails reach 96.0% success and 96.3% protocol adherence at a 1.0% cascading-failure rate, still saving 56.0% of tokens. The critical threshold rises with workflow complexity, which means a fixed compaction percentage is structurally the wrong control knob.
Each link below shares sources, entities, or timing with this story.
arXiv 2609.10263 separates what a persistent agent stores from what it uses, because a superseded fact misleads a current-state answer while remaining necessary for a historical query. A retained archive holds everything; a query-conditioned view governs influence, with same-s...
The diagnosis in this paper is better than the fix, and the fix is very good. Recurrent memory agents fail at long context, but not for the reason most people assume. The bottleneck isn't capture. It's retention. Retention falls below 30% at 896K tokens because every consolida...
1. Flip your multi-model pipeline to review-then-generate. Instead of using a reasoning model to plan before code generation, let the specialist generate freely and use reasoning tokens for review. Paper shows 90.2% pass@1 vs 87.2% for the planning pattern. Source 2. Audit you...
arXiv 2608.01326 models compaction as two games: a Context Selection Game (retain a subset) and a Context Generation Game (summarize into a bounded message). It proves the generation game is equivalent to one-way communication complexity, so the minimum compaction budget for a...
arXiv 2607.28430 adds asynchronous message-passing to coding-agent harnesses with three primitives: threads, messages, and waiting for mentions, where the wait runs as a background task so an agent stays passively aware of teammates without blocking foreground work. Single Cla...
I've spent real hours tuning the CLAUDE.md in my own repos. Rewriting architecture notes. Adding conventions. Trimming when it got long. So this one stung. arXiv 2607.27250 ran a two-agent ablation across Claude Code and Codex: 17 real tasks from 3 repositories, 288 gold-test-...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.