Highly recommended paper from NVIDIA.
This is a dev post classified by AGK Intelligence as AI dev tools (a free resource), kept by the Dev Radar because it carries real work, not commentary.
Highly recommended paper from NVIDIA. (bookmark it) There is a lot of debate about whether to use context compaction aggressively. They show that with frequent compaction, frontier language models lose little to no accuracy on a broad range of long-horizon and long-input tasks while keeping their context short. With one action per context, Claude Opus 4.5 scores 70.2 on SWE-bench Verified against 70.0 with the full history. Median peak context drops from 32K to 8K tokens. The agent carries a short summary of its state between steps instead of the whole transcript. Simple compaction at the
Posted by elvis (327.7k followers) 1 h ago · 76 likes · 3.7k views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: You don't need long context.
More dev work like this
- ChatAndBuild is launching Living Brain. — @wallstengine
- I asked @bot to go through my original post’s replies and quotes to see the savings:… — @shuding
- 🫢牛逼！你的AI代理还在“瞎猜列名”？NVIDIA开源神器，直接把AI训练成懂业务的“数据打工人”！并且你能一眼看出AI代理是否说谎！ — @NFT_Chen
- Build videos in TypeScript — without Chrome in your render pipeline. — @DanKornas
- Drama 3 now cleans up after itself! 🔊 — @plotarmordev
- @Kimi_Moonshot open-sourced TCSExam, a 100 Lean-checked theoretical CS research tasks… — @javierbits
- Made an automation so every morning GitHub Copilot prints this out for me. — @madebygps
- 💥码住！DeepSeek V4.1 Flash 1017 推理档位怎么用最省钱？！ — @NFT_Chen
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.3k posts from 5k X accounts over the last 21 days, 9.3k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-04 14:12 UTC. Full method.