---
title: "Automated AI research and recursive self-improvement"
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
source: https://iinventedlight.com/library/notes/auto-research-and-recursive-self-improvement.md
author: "iinventedlight (synthesis of cited sources); text drafted with AI assistance"
date: 2026-10-14
status: reviewed
reviewed_on: 2026-10-14
checked_on: 2026-10-14
sources:
  - id: S1
    title: "RSIGym: A Flexible Environment for Recursive Self-Improvement (arXiv 2610.10310, v1)"
    publisher: arXiv (Meng et al.)
    url: https://arxiv.org/abs/2610.10310
    accessed: 2026-10-14
    license: CC BY 4.0
  - id: S2
    title: "Self-Adapting Language Models (SEAL) (arXiv 2506.10943, v2)"
    publisher: arXiv (Zweiger et al.)
    url: https://arxiv.org/abs/2506.10943
    accessed: 2026-10-14
    license: CC BY 4.0
  - id: S3
    title: "IdeaScientist: Orchestrating Agents for Grounded Scientific Ideation (arXiv 2610.04074)"
    publisher: arXiv (Liu et al.)
    url: https://arxiv.org/abs/2610.04074
    accessed: 2026-10-14
    license: CC BY 4.0
  - id: S4
    title: "Self-Revising Discovery Systems for Science (arXiv 2606.01444, v2)"
    publisher: arXiv (Wang & Buehler)
    url: https://arxiv.org/abs/2606.01444
    accessed: 2026-10-14
    license: "restricted: CC BY-NC-ND 4.0, restated, not copied"
  - id: S5
    title: "Quantifying the Effect of Test Set Contamination on Generative Evaluations (arXiv 2601.04301, v3)"
    publisher: arXiv (Schaeffer et al.)
    url: https://arxiv.org/abs/2601.04301
    accessed: 2026-10-14
    license: CC BY 4.0
  - id: S6
    title: "Sam Altman on Building the Future of AI (OpenAI Forum)"
    publisher: OpenAI (YouTube)
    url: https://www.youtube.com/watch?v=ZpUKNYcgM-E
    accessed: 2026-10-14
    license: "restricted: restated, not copied"
  - id: S7
    title: "Stanford CS153 guest session with Sam Altman (third-party reupload; official URL not found)"
    publisher: prepublish.ai transcript of YouTube knvviM-LAZU
    url: https://prepublish.ai/youtube-transcript/knvviM-LAZU
    accessed: 2026-10-14
    license: "restricted: restated, not copied"
  - id: S8
    title: "Jan Leike on OpenAI's superalignment project (80,000 Hours, 2023)"
    publisher: 80,000 Hours (YouTube)
    url: https://www.youtube.com/watch?v=ZP_N4q5U3eE
    accessed: 2026-10-14
    license: "restricted: restated, not copied"
  - id: S9
    title: "Supervising AI on hard tasks (Jan Leike, FAR.AI, 2024)"
    publisher: FAR.AI (YouTube)
    url: https://www.youtube.com/watch?v=PuESFNSh_Qo
    accessed: 2026-10-14
    license: "restricted: restated, not copied"
  - id: S10
    title: "Ilya Sutskever interview (Dwarkesh Patel, 2025)"
    publisher: Dwarkesh Patel (YouTube)
    url: https://www.youtube.com/watch?v=aR20FWCCjAs
    accessed: 2026-10-14
    license: "restricted: restated, not copied"
---
# Automated AI research and recursive self-improvement

*Note, 2026-10-14. Status: reviewed. Tags: ai, agents, research.*
*Authorship: synthesis by iinventedlight from the cited sources; text drafted with AI assistance (Grok Bot), reviewed by Cole Withington on 2026-10-14.*

## Summary
Automated AI research means AI agents doing parts of the research loop: coming up with ideas, changing training data or code, running experiments and evaluating results. Recursive self-improvement (RSI) is the stronger idea that each round of improvement makes the next round better, so gains compound. As of October 2026, published experiments show agents can improve a smaller model in one budgeted cycle [F1–F4]. Compounding across cycles has not been shown [F6].

## Sections
### Definitions
- **Auto-research:** an agent runs a research workflow: diagnose, propose, train or revise, evaluate, keep or discard [F1]. It can also mean just the ideation step [F9].
- **Self-improvement:** a model writes its own fine-tuning data and is updated on it [F7], or an agent improves another model's data, settings or harness [F1].
- **Recursive self-improvement:** improvements that compound over cycles. Our definition; the papers below test at most one cycle or a few RL iterations [F6, F7].

### Current evidence
- RSIGym: frontier coding agents with a $500 budget closed part of the gap to perfect scores on five benchmarks [F2], with improvement in 29 of 30 agent–benchmark settings [F3]. Changing only the harness also helped [F4].
- SEAL: a 7B model trained to write its own fine-tuning data slightly beat GPT-4.1-generated data in one setting [F7], but not in another [F8].
- IdeaScientist: research proposals were judged more novel than baselines' [F9]. No proposal was carried out [F11].
- Lab goals (stated, not evidence): OpenAI says it aims for an automated AI researcher by March 2028 [F13, F14].

### Limits and caveats
- RSIGym scores were not on held-out tasks, and each configuration was run only once [F5].
- Contamination research shows that gains on test sets the model has seen can be shallow memorization [F16].
- SEAL suffers forgetting and costs a fine-tune for every reward [F8].
- Judging research quality is noisy, even between human experts [F10].
- Supervising AI on tasks without ground truth is an open problem [F15].

### Speculation and hype (labeled)
- **Speculation:** claims that RSI leads to an "intelligence explosion" go beyond anything measured in these sources.
- **Hype check:** social posts called SEAL evidence that a model "might be alive" and said it works "without retraining". SEAL works *by* retraining [F7]. A post called the categorical framework (S4) a "breakthrough", but its authors describe case studies, not validated laws [F12].
- **Stated plans, not results:** lab timelines [F13, F14] and bootstrapped alignment plans [F15] are goals or forecasts.

## Facts
- id: F1
  claim: RSIGym gives a research agent in a CPU-only container metered training, model, rollout, benchmark and sandbox services under a shared budget, and asks it to improve a target model's data, settings and/or harness.
  confidence: medium
  date: 2026-10-07
  sources: [S1]
  url: https://arxiv.org/abs/2610.10310
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F2
  claim: In RSIGym's joint track (base Qwen3.5-35B-A3B-Base, $500 per benchmark run), the best agent closed on average about 48% of the remaining gap to a perfect score across five benchmarks (RSI-Index 0.48).
  confidence: medium
  date: 2026-10-07
  sources: [S1]
  url: https://arxiv.org/abs/2610.10310
  checked_on: 2026-10-14
  license: CC BY 4.0
  note: Model names in the paper (e.g. "Opus 5", "Sol 6.1") are reported as the authors give them.
- id: F3
  claim: In RSIGym, 29 of 30 agent–benchmark settings ended above the starting system.
  confidence: medium
  date: 2026-10-07
  sources: [S1]
  url: https://arxiv.org/abs/2610.10310
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F4
  claim: With the model weights fixed, harness-only revisions raised Terminal-Bench 2.0 from about 30% to about 41% in RSIGym.
  confidence: medium
  date: 2026-10-07
  sources: [S1]
  url: https://arxiv.org/abs/2610.10310
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F5
  claim: RSIGym's reported scores are not held-out (the agent could see the eval tasks during development), and each configuration was run only once.
  confidence: medium
  date: 2026-10-07
  sources: [S1]
  url: https://arxiv.org/abs/2610.10310
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F6
  claim: RSIGym's authors say whether gains compound across later improvement cycles has not yet been evaluated. Doubling the budget did not consistently help, and a run without network access did not beat the starting system.
  confidence: medium
  date: 2026-10-07
  sources: [S1]
  url: https://arxiv.org/abs/2610.10310
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F7
  claim: SEAL uses reinforcement learning to teach a model to write its own fine-tuning data ("self-edits"), applied as LoRA weight updates. On a SQuAD knowledge task with Qwen2.5-7B it reached 47.0% vs 46.3% for GPT-4.1-written data and 32.7% for the base model.
  confidence: medium
  date: 2025-06-12
  sources: [S2]
  url: https://arxiv.org/abs/2506.10943
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F8
  claim: SEAL loses accuracy on earlier passages over sequential edits (forgetting), needs a fine-tune plus evaluation for each reward, and was slightly beaten by GPT-4.1 data in the 200-passage setting.
  confidence: medium
  date: 2025-06-12
  sources: [S2]
  url: https://arxiv.org/abs/2506.10943
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F9
  claim: IdeaScientist splits research ideation into gap-finder, cross-domain innovator and report-writer agents. It scored 14% above the strongest open auto-research baseline overall, mostly on novelty, using LLM-judge metrics on a held-out 277-paper split.
  confidence: medium
  date: 2026-10-02
  sources: [S3]
  url: https://arxiv.org/abs/2610.04074
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F10
  claim: In IdeaScientist, two human experts agreed only moderately on proposal quality (r = 0.56), and RL training added little over the untrained pipeline (+1.9% novelty, +0.4% quality).
  confidence: medium
  date: 2026-10-02
  sources: [S3]
  url: https://arxiv.org/abs/2610.04074
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F11
  claim: IdeaScientist measures proposal quality only. No proposal was executed, so real-world feasibility is untested.
  confidence: medium
  date: 2026-10-02
  sources: [S3]
  url: https://arxiv.org/abs/2610.04074
  checked_on: 2026-10-14
  license: CC BY 4.0
- id: F12
  claim: Wang and Buehler propose separating "search" within a fixed vocabulary from "discovery" that verifiably extends the vocabulary. Their evidence is case studies, which they say do not establish new experimentally validated laws.
  confidence: low
  date: 2026-05-31
  sources: [S4]
  url: https://arxiv.org/abs/2606.01444
  checked_on: 2026-10-14
  license: CC BY 4.0 (our restatement; source is CC BY-NC-ND, not quoted)
- id: F13
  claim: In an April 2026 OpenAI Forum talk, OpenAI said its official goal is an automated AI researcher by March 2028 (stated goal, not a result).
  confidence: low
  date: 2026-04-07
  sources: [S6]
  url: https://www.youtube.com/watch?v=ZpUKNYcgM-E
  checked_on: 2026-10-14
  license: CC BY 4.0 (our restatement)
- id: F14
  claim: At Stanford CS153, Altman described internal goals of an AI "research intern" by September 2026 and a full end-to-end researcher by March 2028 (stated goals; we have not verified whether the 2026 milestone was met).
  confidence: low
  date: 2026-06-15
  sources: [S7]
  url: https://prepublish.ai/youtube-transcript/knvviM-LAZU
  checked_on: 2026-10-14
  license: CC BY 4.0 (our restatement)
  note: Third-party reupload; no official Stanford video found. Speaker is CEO of an AI lab (conflict of interest).
- id: F15
  claim: Jan Leike has argued for aligning an automated alignment researcher only slightly smarter than humans, then bootstrapping. He notes that alignment research lacks ground-truth labels, so evaluating it needs methods such as deliberately planted flaws (tampering) and adversarial evals.
  confidence: low
  date: 2024-09-10
  sources: [S8, S9]
  url: https://www.youtube.com/watch?v=PuESFNSh_Qo
  checked_on: 2026-10-14
  license: CC BY 4.0 (our restatement)
  note: Two talks by the same speaker, not independent sources.
- id: F16
  claim: In small models (≤344M) trained on copies of the MATH test set, contamination gains disappeared when problems were reworded or their numbers changed.
  confidence: medium
  date: 2026-01-07
  sources: [S5]
  url: https://arxiv.org/abs/2601.04301
  checked_on: 2026-10-14
  license: CC BY 4.0
  note: Tested at small scale only; relevance to RSIGym's non-held-out scores is our inference.
- id: F17
  claim: Ilya Sutskever has said a better target than self-improving AI is AI robustly aligned to care for sentient life, and that capping the power of the strongest systems would help (opinion).
  confidence: low
  date: 2025-11-25
  sources: [S10]
  url: https://www.youtube.com/watch?v=aR20FWCCjAs
  checked_on: 2026-10-14
  license: CC BY 4.0 (our restatement)

## Pointers
- answers: Full RSIGym per-agent, per-benchmark tables and the audit protocol
  where: https://arxiv.org/abs/2610.10310 (§4, Table 4); code https://github.com/evolvent-ai/RSIGym
  access: open
- answers: Category-theory formalism for "discovery" vs "search"
  where: https://arxiv.org/abs/2606.01444
  access: open (NC-ND; read the source, do not copy)

## Disagreements
- What "RSI" means: RSIGym uses the name for one budgeted improvement cycle [F1, F6]. The classic meaning is compounding improvement, which none of these sources demonstrate.
- Lab leaders' timelines [F13, F14] are much more confident than the measured evidence [F5, F6, F11]. Sutskever questions self-improvement as the goal at all [F17].