← 回總覽

Karpathy 的 AutoResearch 将象棋引擎提升至前 50 名特级大师水平

📅 2026-03-22 10:56 Deedy 人工智能 2 分鐘 1565 字 評分: 84
Karpathy AutoResearch 象棋引擎 Rust AI 自动化
📌 一句话摘要 Andrej Karpathy 的 AutoResearch 工具通过 70 多次实验,将一个 vibe coding 的 Rust 象棋引擎从"专家"级别提升至全球前 50 名特级大师水平,达到 2718 ELO。 📝 详细摘要 这条推文分享了 Andrej Karpathy 的 AutoResearch 工具的一个令人印象深刻的实际成果。作者 Deedy 展示了 AutoResearch 是如何自主运行超过 70 次实验对他的 vibe coding Rust 象棋引擎进行优化,并使用爬山算法达到了 2718 ELO 评分,使其成为全球排名第 311 位的象棋引擎。这展
![Image 1: Deedy](https://www.bestblogs.dev/en/tweets?sourceId=SOURCE_54ebf3a2)

Karpathy's Autoresearch pushed my vibecoded Rust chess engine AI from "expert" to a top 50 grandmaster, a #311 chess engine. It ran over 70 experiments on its own and tried to hill climb to the top ELO score it could, landing at 2718!

!Image 2: 视频缩略图

01:01

55 Replies

120 Retweets

1,729 Likes

148.1K Views ![Image 3: Deedy](https://www.bestblogs.dev/en/tweets?sourceid=54ebf3a2)

One Sentence Summary

Andrej Karpathy's AutoResearch tool automatically improved a vibecoded Rust chess engine from 'expert' to top 50 grandmaster level, reaching 2718 ELO through 70+ experiments.

Summary

This tweet shares an impressive real-world result of Andrej Karpathy's AutoResearch tool. The author, Deedy, demonstrates how AutoResearch autonomously ran over 70 experiments on his vibecoded Rust chess engine and used hill climbing optimization to achieve a 2718 ELO rating, ranking it as the #311 chess engine globally. This showcases the potential of AI-driven automated research and optimization in software development. The tweet includes a video demonstrating the engine's gameplay.

AI Score

84

Influence Score 406

Published At Today

Language

English

Tags

Karpathy

AutoResearch

Chess Engine

Rust

AI Automation

查看原文 → 發佈: 2026-03-22 10:56:13 收錄: 2026-03-22 16:00:34

🤖 問 AI

針對這篇文章提問,AI 會根據文章內容回答。按 Ctrl+Enter 送出。