-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
369 lines (355 loc) · 19.3 KB
/
Copy pathindex.html
File metadata and controls
369 lines (355 loc) · 19.3 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
<!doctype html>
<html lang="en" data-lang="en">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<meta name="theme-color" content="#ffffff" />
<meta
name="google-site-verification"
content="D94IoHgseJNY5YrF_3tcZM9igQlPpZqWRlPSXJP3W_g"
/>
<title>Simple Agent Lab | RSIHub, Self-Evolving Agents, AI4AI & RSI</title>
<meta
name="description"
content="Simple Agent Lab builds RSIHub, an open-source framework for self-evolving agents, and researches self-improving systems and skills, AI4AI, and recursive self-improvement (RSI)."
/>
<meta
name="keywords"
content="RSIHub, RSIHub framework, RSIHub agent framework, self-evolving agents, self-improving agents, self-improving systems, self-improving skills, agent self-improvement, AI self-improvement, recursive self-improvement, recursive improvement, RSI, AI agents, autonomous agents, adaptive AI agents, agent skills, agentic AI, AI4AI, autonomous AI research, 自进化智能体, 自改进智能体, 自改进系统, 自改进技能, 递归自我改进"
/>
<meta name="robots" content="index, follow, max-image-preview:large" />
<link rel="canonical" href="https://simpleagentlab.com/" />
<meta property="og:title" content="Simple Agent Lab | RSIHub, Self-Evolving Agents, AI4AI & RSI" />
<meta
property="og:description"
content="Simple yet effective methods for self-evolving agents, self-improving systems and skills, AI4AI, and recursive self-improvement (RSI)."
/>
<meta property="og:type" content="website" />
<meta property="og:url" content="https://simpleagentlab.com/" />
<meta property="og:site_name" content="Simple Agent Lab" />
<meta property="og:image" content="https://simpleagentlab.com/assets/og.png" />
<meta property="og:image:alt" content="Simple Agent Lab — Simple yet effective methods for AI4AI" />
<meta property="og:image:width" content="1200" />
<meta property="og:image:height" content="630" />
<meta name="twitter:card" content="summary_large_image" />
<meta name="twitter:title" content="Simple Agent Lab | RSIHub, Self-Evolving Agents, AI4AI & RSI" />
<meta
name="twitter:description"
content="Researching self-evolving agents, self-improving systems and skills, AI4AI, and RSI."
/>
<meta name="twitter:image" content="https://simpleagentlab.com/assets/og.png" />
<meta name="twitter:image:alt" content="Simple Agent Lab — Simple yet effective methods for AI4AI" />
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@graph": [
{
"@type": "WebSite",
"@id": "https://simpleagentlab.com/#website",
"url": "https://simpleagentlab.com/",
"name": "Simple Agent Lab",
"description": "Simple yet effective methods for self-evolving agents, self-improving systems and skills, AI4AI, and recursive self-improvement.",
"inLanguage": ["en", "zh-CN"],
"keywords": [
"RSIHub",
"Self-evolving agents",
"Self-improving agents",
"Self-improving systems",
"Self-improving skills",
"Agent self-improvement",
"AI self-improvement",
"AI agents",
"AI4AI",
"Recursive self-improvement",
"RSI"
],
"publisher": {
"@id": "https://simpleagentlab.com/#organization"
}
},
{
"@type": "Organization",
"@id": "https://simpleagentlab.com/#organization",
"url": "https://simpleagentlab.com/",
"name": "Simple Agent Lab",
"description": "An open AI research collective studying self-evolving agents, self-improving systems and skills, AI4AI, and recursive self-improvement.",
"logo": "https://simpleagentlab.com/assets/logo.png",
"knowsAbout": [
"RSIHub",
"Self-evolving agents",
"Self-improving agents",
"Self-improving systems",
"Self-improving skills",
"Agent self-improvement",
"AI agents",
"AI4AI",
"Recursive self-improvement",
"RSI"
],
"sameAs": [
"https://github.com/simple-agent-lab"
]
},
{
"@type": "SoftwareSourceCode",
"@id": "https://simpleagentlab.com/rsihub/#software",
"name": "RSIHub",
"url": "https://simpleagentlab.com/rsihub/",
"codeRepository": "https://github.com/simple-agent-lab/RSIHub",
"description": "RSIHub is an open-source framework for self-evolving agents: build agents that improve and keep the evidence.",
"programmingLanguage": "Python",
"license": "https://www.apache.org/licenses/LICENSE-2.0",
"author": {
"@id": "https://simpleagentlab.com/#organization"
},
"sameAs": [
"https://github.com/simple-agent-lab/RSIHub",
"https://simpleagentlab.com/RSIHub/"
]
},
{
"@type": "ScholarlyArticle",
"@id": "https://simpleagentlab.com/ai4ai/#article",
"name": "AI4AI Survey: From Long-Horizon Agents to Recursive Self-Improvement—Definitions, Reliable Horizons, and Open Problems",
"url": "https://simpleagentlab.com/ai4ai/",
"description": "A survey of AI4AI (AI for AI): how far an AI system can reliably carry an improvement process from idea to a validated result, a taxonomy for locating where the loop closes, a closure audit of 35 systems, and the composition gap that bounds recursive self-improvement.",
"genre": "survey",
"datePublished": "2026-08-28",
"inLanguage": ["en", "zh-CN"],
"publisher": {
"@id": "https://simpleagentlab.com/#organization"
},
"about": [
"AI4AI",
"Long-horizon agents",
"Recursive self-improvement",
"Self-improving systems",
"Agent evaluation"
]
}
]
}
</script>
<link rel="icon" href="./assets/favicon.svg?v=4" type="image/svg+xml" />
<link rel="stylesheet" href="./assets/styles.css" />
<script>
try {
var root = document.documentElement;
var savedLanguage = localStorage.getItem("simple-agent-lab-language-choice");
var browserLanguages = navigator.languages && navigator.languages.length
? navigator.languages
: [navigator.language || "en"];
var browserLanguage = browserLanguages.some(function (language) {
return language.toLowerCase().startsWith("zh");
}) ? "zh" : "en";
var language = savedLanguage === "en" || savedLanguage === "zh"
? savedLanguage
: browserLanguage;
root.dataset.lang = language;
root.lang = language === "zh" ? "zh-CN" : "en";
var savedTheme = localStorage.getItem("simple-agent-lab-theme");
var theme = savedTheme === "light" || savedTheme === "dark"
? savedTheme
: window.matchMedia("(prefers-color-scheme: dark)").matches
? "dark"
: "light";
root.dataset.theme = theme;
root.style.colorScheme = theme;
document.querySelector('meta[name="theme-color"]').setAttribute(
"content",
theme === "dark" ? "#111513" : "#ffffff"
);
} catch (error) {}
</script>
<script src="./assets/site.js" defer></script>
</head>
<body>
<a class="skip-link" href="#main">
<span class="lang-en" lang="en">Skip to content</span>
<span class="lang-zh" lang="zh-CN">跳至正文</span>
</a>
<header class="site-header">
<a class="small-brand" href="#top" aria-label="Simple Agent Lab home">
<img src="./assets/logo.svg" width="26" height="26" alt="" />
<span>Simple Agent Lab</span>
</a>
<nav aria-label="Primary navigation">
<a href="#work">
<span class="lang-en" lang="en">Work</span>
<span class="lang-zh" lang="zh-CN">研究工作</span>
</a>
<a href="#about">
<span class="lang-en" lang="en">About</span>
<span class="lang-zh" lang="zh-CN">关于我们</span>
</a>
<a href="https://github.com/simple-agent-lab" target="_blank" rel="noreferrer">GitHub</a>
<button class="theme-toggle" type="button" aria-label="Switch to dark theme">
<span class="theme-action-en" lang="en">Dark</span>
<span class="theme-action-zh" lang="zh-CN">深色</span>
</button>
<button class="language-toggle" type="button" aria-label="切换为中文">中文</button>
</nav>
</header>
<div class="masthead" id="top">
<img class="masthead-mark" src="./assets/logo.svg" width="78" height="78" alt="Simple Agent Lab loop mark" />
<h1>Simple Agent Lab</h1>
<p>
<span class="lang-en" lang="en">Simple yet effective methods for AI4AI.</span>
<span class="lang-zh" lang="zh-CN">探索 AI4AI 中 simple yet effective 的方法。</span>
</p>
</div>
<main id="main">
<article class="manifesto">
<p class="opening">
<span class="lang-en" lang="en">Simple Agent Lab is an open AI research collective. We’re building a future where AI can help build AI, while the methods behind it remain understandable to humans.</span>
<span class="lang-zh" lang="zh-CN">Simple Agent Lab 是一个开放的 AI 研究共同体。我们希望构建一个由 AI 帮助创造、又能被人类理解的未来。</span>
</p>
<p>
<span class="lang-en" lang="en">AI will increasingly write code, generate data, run experiments, evaluate results, and improve models and systems. Improved AI can then return to build the next generation—forming a recursive loop of improvement.</span>
<span class="lang-zh" lang="zh-CN">AI 将越来越多地编写代码、生成数据、运行实验、评估结果，并改进模型与系统。改进后的 AI 可以继续参与下一轮构建，形成递归的自我改进循环。</span>
</p>
<p>
<span class="lang-en" lang="en">But a future built by AI should not become too complex for humans to understand. We want every improvement to remain simple, transparent, and verifiable—so people can understand how it happened and decide where it goes next.</span>
<span class="lang-zh" lang="zh-CN">但由 AI 构建的未来，不应复杂到无法被人类理解。我们希望每一次改进都保持简单、透明、可验证，让人们能够理解它如何发生，并决定它走向何方。</span>
</p>
<h2>
<span class="lang-en" lang="en">From AI4AI to recursive improvement</span>
<span class="lang-zh" lang="zh-CN">从 AI4AI 走向递归改进</span>
</h2>
<p>
<span class="lang-en" lang="en">We research <strong>AI4AI</strong> and <strong>Self-Evolving Agents</strong>, build <strong>Self-Improving Systems</strong> and <strong>Self-Improving Skills</strong>, and explore the path toward <strong>Recursive Self-Improvement (RSI)</strong>.</span>
<span class="lang-zh" lang="zh-CN">我们研究 <strong>AI4AI</strong> 与 <strong>Self-Evolving Agents（自进化智能体）</strong>，构建 <strong>Self-Improving Systems（自改进系统）</strong>与 <strong>Self-Improving Skills（自改进技能）</strong>，并探索通往 <strong>Recursive Self-Improvement（RSI，递归自我改进）</strong> 的路径。</span>
</p>
<h2>
<span class="lang-en" lang="en">Simple yet effective</span>
<span class="lang-zh" lang="zh-CN">Simple yet effective</span>
</h2>
<p>
<span class="lang-en" lang="en">Simple does not mean unsophisticated. Simplicity is what allows humans to continue understanding and participating in a future built by AI. We want people to see how a method works, verify why an improvement happened, and help decide what comes next.</span>
<span class="lang-zh" lang="zh-CN">简单不代表简陋，简单是让人类始终能够理解并参与 AI 构建未来的前提。我们希望人们能够看懂方法如何工作，验证改进为何发生，并参与决定下一步。</span>
</p>
<h2>
<span class="lang-en" lang="en">Research is better when shared</span>
<span class="lang-zh" lang="zh-CN">研究因分享而进步</span>
</h2>
<p>
<span class="lang-en" lang="en">We share code, experiments, evaluations, and failures so others can inspect, reproduce, and build on what we learn. Our members have studied or worked at ByteDance, UC Berkeley, Tsinghua University, Shanghai Jiao Tong University, and Tongji University. All members participate in a personal capacity.</span>
<span class="lang-zh" lang="zh-CN">我们分享代码、实验、评测与失败，让更多人能够检查、复现并继续推进这些探索。我们的成员曾在 ByteDance、UC Berkeley、清华大学、上海交通大学和同济大学学习或工作。所有成员均以个人身份参与。</span>
</p>
<section id="work" class="work-section">
<h2>
<span class="lang-en" lang="en">Selected work</span>
<span class="lang-zh" lang="zh-CN">代表工作</span>
</h2>
<div class="work-list">
<div class="work-item">
<a
class="work-summary"
href="./rsihub/"
aria-label="RSIHub — self-evolving agent framework"
>
<span class="work-tag">
<span class="lang-en" lang="en">New</span>
<span class="lang-zh" lang="zh-CN">最新</span>
</span>
<strong>RSIHub</strong>
<small>
<span class="lang-en" lang="en">Build agents that improve and keep the evidence. A simple framework for agent evolution.</span>
<span class="lang-zh" lang="zh-CN">让智能体持续改进，并保留证据。一个简单的智能体进化框架。</span>
</small>
</a>
<span class="work-actions">
<a href="https://github.com/simple-agent-lab/RSIHub" target="_blank" rel="noreferrer">
<span>Code</span>
<b aria-hidden="true">↗</b>
</a>
<a href="https://simpleagentlab.com/RSIHub/" target="_blank" rel="noreferrer">
<span class="lang-en" lang="en">Docs</span>
<span class="lang-zh" lang="zh-CN">文档</span>
<b aria-hidden="true">↗</b>
</a>
</span>
</div>
<div class="work-item">
<a
class="work-summary"
href="./ai4ai/"
aria-label="AI4AI Survey — from long-horizon agents to recursive self-improvement"
>
<span class="work-tag">
<span class="lang-en" lang="en">Survey</span>
<span class="lang-zh" lang="zh-CN">综述</span>
</span>
<strong>AI4AI Survey</strong>
<small>
<span class="lang-en" lang="en">A survey of AI4AI, from long-horizon agents to recursive self-improvement: definitions, reliable horizons, and a closure audit of 35 systems. Now published on Preprints.org.</span>
<span class="lang-zh" lang="zh-CN">AI4AI 综述：从长程智能体到递归自进化。梳理核心定义、长程可靠性边界，并对 35 个前沿自改进系统展开闭合度审计。已在 Preprints.org 正式发布。</span>
</small>
</a>
<span class="work-actions">
<a href="./ai4ai/">
<span class="lang-en" lang="en">Read the AI4AI survey</span>
<span class="lang-zh" lang="zh-CN">阅读 AI4AI 综述</span>
</a>
<a href="https://doi.org/10.20944/preprints202608.2108.v1" target="_blank" rel="noreferrer">
<span class="lang-en" lang="en">Paper</span>
<span class="lang-zh" lang="zh-CN">论文</span>
<b aria-hidden="true">↗</b>
</a>
</span>
</div>
<div class="work-item">
<a
class="work-summary"
href="https://arxiv.org/abs/2606.31551"
target="_blank"
rel="noreferrer"
aria-label="AutoTrainess paper"
>
<strong>AutoTrainess</strong>
<small>
<span class="lang-en" lang="en">Teaching language models to improve language models autonomously through a training-specialized Agent–Computer Interface.</span>
<span class="lang-zh" lang="zh-CN">通过面向训练的专用 Agent–Computer Interface，让语言模型自主改进语言模型。</span>
</small>
</a>
<span class="work-actions">
<a href="https://arxiv.org/abs/2606.31551" target="_blank" rel="noreferrer">
<span class="lang-en" lang="en">Paper</span>
<span class="lang-zh" lang="zh-CN">论文</span>
<b aria-hidden="true">↗</b>
</a>
<a href="https://github.com/simple-agent-lab/AutoTrainess" target="_blank" rel="noreferrer">
<span>Code</span>
<b aria-hidden="true">↗</b>
</a>
</span>
</div>
<a href="https://github.com/simple-agent-lab/simple-long-horizon-agent" target="_blank" rel="noreferrer">
<span>
<strong>Simple Long Horizon Agent</strong>
<small>
<span class="lang-en" lang="en">A minimal, understandable agent loop for real long-horizon work.</span>
<span class="lang-zh" lang="zh-CN">一个保持简单、可理解，同时面向真实长程任务的 Agent Loop。</span>
</small>
</span>
<b aria-hidden="true">↗</b>
</a>
</div>
</section>
<section id="about" class="closing">
<p><strong>AI4AI</strong> <span class="lang-en" lang="en">is our direction.</span><span class="lang-zh" lang="zh-CN">是我们的方向。</span></p>
<p><strong>Recursive Improvement</strong> <span class="lang-en" lang="en">is our path.</span><span class="lang-zh" lang="zh-CN">是我们的路径。</span></p>
<p><strong>Simple</strong> <span class="lang-en" lang="en">is our principle.</span><span class="lang-zh" lang="zh-CN">是我们的原则。</span></p>
</section>
</article>
</main>
<footer>
<span>Simple Agent Lab</span>
<a href="./disclaimer.html">
<span class="lang-en" lang="en">Disclaimer</span>
<span class="lang-zh" lang="zh-CN">免责声明</span>
</a>
<span>© 2026</span>
</footer>
</body>
</html>