Paper

{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.19461v1</id>\n    <title>Hyperagents</title>\n    <updated>2026-03-19T20:50:08Z</updated>\n    <link href='https://arxiv.org/abs/2603.19461v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.19461v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Self-improving AI systems aim to reduce reliance on human engineering by learning to improve their own learning and problem-solving processes. Existing approaches to self-improvement rely on fixed, handcrafted meta-level mechanisms, fundamentally limiting how fast such systems can improve. The Darwin Gödel Machine (DGM) demonstrates open-ended self-improvement in coding by repeatedly generating and evaluating self-modified variants. Because both evaluation and self-modification are coding tasks, gains in coding ability can translate into gains in self-improvement ability. However, this alignment does not generally hold beyond coding domains. We introduce \\textbf{hyperagents}, self-referential agents that integrate a task agent (which solves the target task) and a meta agent (which modifies itself and the task agent) into a single editable program. Crucially, the meta-level modification procedure is itself editable, enabling metacognitive self-modification, improving not only the task-solving behavior, but also the mechanism that generates future improvements. We instantiate this framework by extending DGM to create DGM-Hyperagents (DGM-H), eliminating the assumption of domain-specific alignment between task performance and self-modification skill to potentially support self-accelerating progress on any computable task. Across diverse domains, the DGM-H improves performance over time and outperforms baselines without self-improvement or open-ended exploration, as well as prior self-improving systems. Furthermore, the DGM-H improves the process by which it generates new agents (e.g., persistent memory, performance tracking), and these meta-level improvements transfer across domains and accumulate across runs. DGM-Hyperagents offer a glimpse of open-ended AI systems that do not merely search for better solutions, but continually improve their search for how to improve.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-19T20:50:08Z</published>\n    <arxiv:comment>Code at https://github.com/facebookresearch/Hyperagents</arxiv:comment>\n    <arxiv:primary_category term='cs.AI'/>\n    <author>\n      <name>Jenny Zhang</name>\n    </author>\n    <author>\n      <name>Bingchen Zhao</name>\n    </author>\n    <author>\n      <name>Wannan Yang</name>\n    </author>\n    <author>\n      <name>Jakob Foerster</name>\n    </author>\n    <author>\n      <name>Jeff Clune</name>\n    </author>\n    <author>\n      <name>Minqi Jiang</name>\n    </author>\n    <author>\n      <name>Sam Devlin</name>\n    </author>\n    <author>\n      <name>Tatiana Shavrina</name>\n    </author>\n  </entry>"
}