6Latent Space (swyx)·
Lilian Weng summarizes 35 papers on Harness Engineering for RSI
Latent Space's AINews digest covers a summary by Lilian Weng of 35 papers on Harness Engineering for Recursive Self-Improvement (RSI), a topic at the intersection of agent scaffolding, self-improvement loops, and AI safety. Weng, a prominent researcher at OpenAI, synthesizing this volume of work signals growing institutional attention to RSI as a research area. The digest frames this as a quiet-day read, suggesting it is a curated secondary synthesis rather than a primary research release.
Related guides (4)
Related events (8)
Import AI 456: RSI and Economic Growth, AI Regulation Optionality, and Neural Computer
Import AI issue 456 covers three topics: recursive self-improvement (RSI) and its implications for economic growth, frameworks for 'radical optionality' in AI regulation, and a neural computer architecture. The newsletter synthesizes recent developments in AI capability trajectories and governance approaches. As a tier-2 commentary source, it provides synthesis and analysis rather than primary research.
Latent Space AINews: Meta-Harness Summer roundup
Latent Space's AINews digest covers a period they're calling 'Meta-Harness Summer,' signaling a trend toward higher-order agent harness tooling — frameworks that orchestrate or compose other harnesses. The piece appears to be a community news roundup from a tier-2 commentary source. The framing suggests growing ecosystem maturity in agent orchestration tooling.
The Batch explains recursive self-improvement hype following Anthropic's coding productivity report
The Batch analyzes the surge of interest in recursive self-improvement (RSI) triggered by Anthropic's report that Claude now authors or co-authors 80% of the company's code, up from under 5% before Claude Code launched. The piece documents concrete productivity metrics—engineers contributing 8x more code lines in Q2 2026 versus Q1 2023, and 800 API fixes shipped in April that would have taken humans four years alone—alongside a spectrum of community reactions ranging from skeptical (Brundage, Mollick) to opportunistic (OpenAI, Sakana AI's new RSI Lab). The commentary frames RSI as theoretically distant but notes the marketing dimension of Anthropic's framing and the gap between agentic coding assistance and true self-directed improvement.
Import AI 460: Reward hacking, RSI data from Anthropic, and RL-based quadcopter racing
Import AI issue 460 covers three main topics: reward hacking as a societal-scale concern, repetitive strain injury (RSI) data released by Anthropic related to AI labor/usage patterns, and reinforcement learning applied to quadcopter racing. The newsletter also raises the question of when financial markets will begin pricing in transformative AI scenarios. This is a curated commentary digest from Jack Clark covering recent AI research and industry developments.
Survey of recursive self-improvement in AI: taxonomy, evaluator hierarchy, and governance gaps
A new arXiv survey covers 1,250 papers (2024–2026) on AI self-improvement, proposing a two-axis taxonomy distinguishing what is improved (behavior, policy, evaluator, or research process) from the degree of loop closure (human-in-the-loop to fully closed). The authors construct a verification hierarchy for self-evaluation signals—from formal verifiers (strongest) to intrinsic self-assessment (weakest)—and find that demonstrated self-improvement strength tracks this hierarchy while failure modes (self-confirming loops, model collapse, diversity collapse) arise from its violations. The paper argues that 'research direction-setting' remains the key bottleneck keeping humans in the loop, and identifies governance-grade measurement of self-improvement as the most underpopulated niche in the field. The work connects technical RSI limits to safety and governance concerns raised by frontier labs experimenting with closed-loop AI research.
Awesome Harness Engineering: Curated List for AI Agent Infrastructure
A GitHub repository aggregating resources on AI agent harness engineering, covering tools, patterns, evaluations, memory systems, MCP (Model Context Protocol), permissions, observability, and orchestration. The list has accumulated 1,318 stars with 39 added today, indicating moderate community traction. It serves as a reference index rather than original research or tooling.
SIA: Self-Improving AI via Joint Harness and Weight Updates
SIA proposes a self-improving loop in which a Feedback-Agent simultaneously updates both the scaffold (harness) and model weights of a task-specific agent, unifying two previously disjoint research lines: meta-agent scaffold rewriting and test-time training. The system is evaluated on three diverse benchmarks—Chinese legal charge classification, GPU kernel optimization, and single-cell RNA denoising—achieving gains of 56.6%, 91.9% runtime reduction, and 502% respectively over baselines. The paper argues that harness updates shape agentic behavior while weight updates instill domain intuition that prompting alone cannot provide, and that combining both levers consistently outperforms either alone.
Import AI 455: AI systems are about to start building themselves
Import AI issue 455 covers the emerging trend of AI systems automating AI research, framing it as a first step toward recursive self-improvement. The commentary synthesizes recent developments suggesting AI is beginning to participate meaningfully in its own development pipeline. As a tier-2 newsletter, this represents curated analysis of frontier AI research directions rather than primary reporting.