# Radixark and Zhuofeng Li share lessons from autonomous infrastructure optimization experiments

> Original title: Turn infrastructure optimization into verifiable tasks.

- Source: LMSYS Org
- Published: 2026-11-09T01:58:49.000Z
- Bad Influence: https://badinfluence.news/items/ao3f10lfdcb1yqq751n79qg9b
- Original: https://x.com/lmsysorg/status/2119610628944138602

## Summary

Chenyang Zhao of RadixArk and Zhuofeng Li of OpenRSI and Google DeepMind share lessons from long-running autonomous experiments on infrastructure optimization. The post says improvements from these experiments have been merged upstream. The post frames infrastructure optimization as a set of verifiable tasks.

## Post

Turn infrastructure optimization into verifiable tasks.

Chenyang Zhao (@radixark) and Zhuofeng Li (@OpenRSI  / @GoogleDeepMind) share lessons from long-running autonomous experiments and improvements merged upstream.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWpedr9bUAAqnyG.jpg&mode=thumb&exp=1794528000&sig=e8f08b7a438207d1)