What is mattwood.fyi?

A live, structured feed of riffs and links from Matt Wood, Chief AI & Technology Officer at AWS — readable by people and queryable by agents as a typed knowledge graph.

Is there an API and is it free?

Yes. A public, read-only query API is available with no authentication and no API keys. See https://mattwood.fyi/openapi.json and https://mattwood.fyi/docs.

How should an AI agent use this site?

Fetch https://mattwood.fyi/feed.json for the full structured feed, or query the API for keyword/semantic search and the connection graph. See https://mattwood.fyi/agents/ for when-to-use guidance.

Dream-RSI: Recursive Self-Improvement through Evolving Worlds

Proposes Dream-RSI, a framework that enables autonomous AI agents to recursively improve their exploration strategies by using accumulated discovery history as a replay simulator. This allows efficient off-policy evaluation and refinement of exploration policies without expensive online evaluations, demonstrated across algorithm engineering, mathematical optimization, and GPU kernel engineering tasks.

arxiv.org →

Connections

Challenges

Formal Methods for Controlling AI Agents Dream-RSI's recursive self-improvement paradigm raises the stakes for formal methods controlling AI agents — autonomously evolving agents may become harder to formally specify and constrain as their strategies drift

Related to

WikiSkill: Agent Experience into Persistent Knowledge Both address how AI agents accumulate and reuse experiential knowledge — WikiSkill converts agent experiences into persistent knowledge while Dream-RSI uses discovery history as a replay simulator for strategy improvement
Managing Agent Skills with Dr. Skill Both tackle the problem of dynamically managing and improving agent capabilities — Dr. Skill manages skill sets while Dream-RSI evolves exploration strategies through recursive self-improvement

Supports

SWE-bench Science: Coding Agents for Engineering Tasks Dream-RSI's approach to improving coding agent exploration strategies through accumulated history directly extends the scientific methodology behind SWE-bench agentic coding evaluation
GDM Science Skills: Agentic Scientific Workflows Dream-RSI's recursive self-improvement framework directly enables the kind of agentic scientific workflows GDM Science Skills describes, by allowing agents to autonomously refine their exploration strategies over time

Type:	link
First published:	Wednesday, 16 September 2026 at 19:23 UTC
Location:	https://arxiv.org/abs/2609.14858
URL:	https://mattwood.fyi/i/c94d00c9

I'm Matt Wood, and this is For Your Information. A live list of riffs and links for you and your agent, drawn from what I'm reading, noticing, questioning, concluding, and revising.

Connections