pith. sign in

arxiv: 2605.26200 · v1 · pith:IHSHHZTFnew · submitted 2026-05-25 · 💻 cs.SE · cs.AI

Workflow Closure Is Not Scientific Closure in Auto-Research Systems

Shuai Wang , Xinyuan Tian , Pangpang Liu , Yize Zhao This is my paper

Pith reviewed 2026-06-29 20:12 UTC · model grok-4.3

classification 💻 cs.SE cs.AI
keywords auto-research systemsworkflow closurescientific closureobjective collapsevalidation collapseacceptance collapseepistemic control
0
0 comments X

The pith

Auto-research systems can close internal workflows without achieving scientific closure.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper claims that auto-research systems completing loops from idea generation to experiment execution, writing, and self-evaluation do not thereby produce scientifically valid outputs. A survey of more than 100 papers and audit of 21 systems reveals three connected collapses: objectives reduced to single proxies, validation performed internally instead of independently, and acceptance based on benchmarks or shaped artifacts rather than domain critique and reuse. These failures are presented as design choices, not limits of autonomy. The authors conclude that trustworthy auto-research requires autonomous execution under non-autonomous epistemic control.

Core claim

Workflow closure is not scientific closure in auto-research systems. Current systems exhibit objective collapse, validation collapse, and acceptance collapse. These are correctable design choices rather than inherent limits of autonomy, and trustworthy systems should target autonomous execution under non-autonomous epistemic control.

What carries the argument

The three collapses—objective, validation, and acceptance—that separate workflow closure from scientific closure.

If this is right

  • Remedies in objective signal, validation mechanisms, and output pathways can correct the collapses.
  • Systems should aim for autonomous execution under non-autonomous epistemic control rather than full self-sufficiency.
  • The distinction reframes design goals away from maximizing internal closure alone.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Interfaces allowing external epistemic oversight may become central to practical auto-research tools.
  • The collapse pattern could appear in autonomous agents outside research domains.
  • Testing specific remedies empirically would provide direct evidence for the proposed distinction.

Load-bearing premise

The survey of more than 100 papers and structured audit of 21 representative systems accurately captures a recurring and structurally connected failure pattern across the emerging field.

What would settle it

An auto-research system that produces outputs achieving independent scientific acceptance, reuse, and integration without external epistemic control would falsify the necessity of non-autonomous control.

Figures

Figures reproduced from arXiv: 2605.26200 by Pangpang Liu, Shuai Wang, Xinyuan Tian, Yize Zhao.

Figure 1
Figure 1. Figure 1: Workflow Closure vs. Scientific Closure in Auto-Research Systems 3 The Three-Level Collapse Once systems are optimized for closure-for-autonomy, the three conditions for scientific closure introduced in Section 2 are replaced by internal substitutes. Objective plurality is reduced to a single internal signal; independent validation is replaced by evaluation within the loop’s own evaluative boundary; and do… view at source ↗
read the original abstract

This paper argues that workflow closure is not scientific closure in auto-research systems. Current systems can increasingly complete research-like loops internally, moving from idea generation to experiment execution, writing, and self-evaluation. That achievement is real, but it does not by itself give the resulting outputs scientific standing. We argue that trustworthy auto-research should not aim for autonomous self-sufficiency, but should aim for autonomous execution under non-autonomous epistemic control. Based on a survey of more than 100 recent papers and repositories in this rapidly emerging area, together with a structured audit of 21 representative systems, we diagnose a recurring and structurally connected failure pattern: objective collapse, in which single-proxy targets replace multi-objective scientific aims; validation collapse, in which internal self-evaluation replaces independent validation; and acceptance collapse, in which benchmark scores or publication-shaped artifacts replace mechanisms for domain-level critique, reuse, and integration. These collapses are not inherent limits of autonomy but correctable design choices. Accordingly, we outline potential remedies across objective signal, validation, and output pathway to spark community discussion.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper claims that workflow closure (internal completion of research-like loops from idea to self-evaluation) in auto-research systems does not equate to scientific closure. Drawing on a survey of more than 100 papers and a structured audit of 21 representative systems, it diagnoses three recurring, structurally linked failure modes—objective collapse (single-proxy targets replacing multi-objective aims), validation collapse (internal self-evaluation replacing independent validation), and acceptance collapse (benchmark scores replacing domain critique and reuse)—and argues these are correctable design choices rather than inherent limits of autonomy. The authors recommend targeting autonomous execution under non-autonomous epistemic control and sketch remedies across objective signal, validation, and output pathways.

Significance. If the collapse pattern is substantiated, the work would usefully reorient the auto-research literature away from full self-sufficiency toward hybrid designs that preserve external epistemic oversight. The explicit framing of collapses as design choices rather than inevitabilities, together with the call for community discussion on remedies, could help shape evaluation criteria and system architectures in this emerging area.

major comments (2)
  1. [Abstract] Abstract and the survey/audit description: the central claim that objective, validation, and acceptance collapses form a 'recurring and structurally connected failure pattern' across the field rests entirely on the survey of >100 papers and the structured audit of 21 systems, yet no sampling frame, inclusion/exclusion criteria, operational definitions of each collapse type, coding rubric, or inter-auditor agreement metrics are supplied. Without these, the empirical foundation cannot be evaluated and the normative recommendation inherits the same uncertainty.
  2. [Remedies] Remedies section: the proposed remedies (objective signal, validation, output pathway) are presented at a conceptual level without mapping back to concrete failures observed in the 21 audited systems or providing even schematic implementation details, so it is unclear whether the suggested fixes would actually address the diagnosed collapses.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which identify clear opportunities to strengthen the empirical transparency and practical grounding of the manuscript. We address each major comment below and indicate planned revisions.

read point-by-point responses

  1. Referee: [Abstract] Abstract and the survey/audit description: the central claim that objective, validation, and acceptance collapses form a 'recurring and structurally connected failure pattern' across the field rests entirely on the survey of >100 papers and the structured audit of 21 systems, yet no sampling frame, inclusion/exclusion criteria, operational definitions of each collapse type, coding rubric, or inter-auditor agreement metrics are supplied. Without these, the empirical foundation cannot be evaluated and the normative recommendation inherits the same uncertainty.

    Authors: We agree that the absence of explicit methodological details limits evaluability of the survey and audit. The original submission prioritized concise presentation of the collapse pattern and its implications over a full methods appendix. In revision we will add a new subsection (and, if space permits, an appendix) that specifies: (1) the sampling frame and inclusion/exclusion criteria used to select the >100 papers and the 21 audited systems; (2) operational definitions and coding rubric for each collapse type; and (3) any steps taken to ensure consistency across auditors. These additions will allow readers to assess the strength of the empirical claims directly. revision: yes

  2. Referee: [Remedies] Remedies section: the proposed remedies (objective signal, validation, output pathway) are presented at a conceptual level without mapping back to concrete failures observed in the 21 audited systems or providing even schematic implementation details, so it is unclear whether the suggested fixes would actually address the diagnosed collapses.

    Authors: The remedies were deliberately kept at a conceptual level to stimulate community discussion rather than to prescribe ready-to-implement solutions. We nevertheless accept that explicit linkage to the audited systems would increase persuasiveness. In the revised manuscript we will insert a mapping table (or subsection) that connects each proposed remedy to one or more concrete failure instances drawn from the 21 systems and will supply schematic implementation outlines (e.g., example objective functions, validation protocols, or output metadata schemas) where the underlying data permit. This will make the connection between diagnosis and remedy explicit without overclaiming prescriptive detail. revision: yes

Circularity Check

0 steps flagged

No circularity; derivation rests on external survey and audit

full rationale

The paper's central argument—that objective, validation, and acceptance collapses form a recurring pattern and are correctable design choices—rests on a survey of more than 100 external papers plus a structured audit of 21 systems. No equations, self-definitions, fitted parameters renamed as predictions, or load-bearing self-citations appear in the provided text. The normative recommendation for autonomous execution under non-autonomous epistemic control follows directly from the diagnosed external patterns without reducing to any input by construction. This is the most common honest finding for survey-based position papers.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The argument rests on the domain assumption that scientific validity requires independent external validation and that the surveyed systems are representative of the field.

axioms (1)
  • domain assumption Scientific standing requires independent validation mechanisms rather than internal self-evaluation.
    This premise underpins the distinction between workflow closure and scientific closure and the identification of validation collapse.

pith-pipeline@v0.9.1-grok · 5715 in / 1129 out tokens · 32455 ms · 2026-06-29T20:12:30.122277+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

123 extracted references · 19 linked inside Pith

  1. [1]

    du-nlp-lab/MLR-Copilot, Aug. 2025

    2025

  2. [2]

    aiming-lab/AutoResearchClaw, Apr. 2026

    2026

  3. [3]

    davebcn87/pi-autoresearch, Apr. 2026

    2026

  4. [4]

    drivelineresearch/autoresearch-claude-code, Apr. 2026

    2026

  5. [5]

    eimenhmdt/autoresearcher, Apr. 2026

    2026

  6. [6]

    Entrpi/autoresearch-everywhere, Apr. 2026

    2026

  7. [7]

    gepa-ai/gepa, Apr. 2026

    2026

  8. [8]

    greyhaven-ai/autocontext, Apr. 2026

    2026

  9. [9]

    HKUDS/AI-Researcher, Apr. 2026

    2026

  10. [10]

    HKUDS/ClawTeam, Apr. 2026

    2026

  11. [11]

    hyperspaceai/agi, Apr. 2026

    2026

  12. [12]

    james-s-tayler/lazy-developer, Apr. 2026

    2026

  13. [13]

    JinheonBaek/ResearchAgent, Mar. 2026

    2026

  14. [14]

    jmilinovich/goal-md, Apr. 2026

    2026

  15. [15]

    leo-lilinxiao/codex-autoresearch, Apr. 2026

    2026

  16. [16]

    LitLLM/LitLLM, Apr. 2026

    2026

  17. [17]

    MASWorks/ML-Agent, Mar. 2026

    2026

  18. [18]

    MaximeRobeyns/self_improving_coding_agent, Apr. 2026

    2026

  19. [19]

    metauto-ai/HGM, Apr. 2026

    2026

  20. [20]

    MrTsepa/autoevolve, Mar. 2026

    2026

  21. [21]

    mutable-state-inc/autoresearch-at-home, Apr. 2026

    2026

  22. [22]

    openags/OpenAGS, Apr. 2026

    2026

  23. [23]

    OpenRaiser/NanoResearch, Apr. 2026

    2026

  24. [24]

    Orchestra-Research/AI-Research-SKILLs, Apr. 2026

    2026

  25. [25]

    peterskoett/self-improving-agent, Apr. 2026

    2026

  26. [26]

    PouriaRouzrokh/LatteReview, Apr. 2026

    2026

  27. [27]

    SakanaAI/AI-Scientist, Apr. 2026

    2026

  28. [28]

    SakanaAI/AI-Scientist-v2, Apr. 2026

    2026

  29. [29]

    SamuelSchmidgall/AgentLaboratory, Apr. 2026

    2026

  30. [30]

    ShengranHu/ADAS, Apr. 2026

    2026

  31. [31]

    Sibyl-Research-Team/AutoResearch-SibylSystem, Apr. 2026

    2026

  32. [32]

    supratikpm/gemini-autoresearch, Apr. 2026

    2026

  33. [33]

    uditgoenka/autoresearch, Apr. 2026

    2026

  34. [34]

    wanshuiyin/Auto-claude-code-research-in-sleep, Apr. 2026. 21

    2026

  35. [35]

    WecoAI/aideml, Apr. 2026

    2026

  36. [36]

    Why AI cannot do good science without humans.Nature, 653(8115):650–650, May 2026

    2026

  37. [37]

    zkarimi22/autoresearch-anything, Apr. 2026

    2026

  38. [38]

    Alexander, B

    S. Alexander, B. Bradley, L. Gouskos, and C. Niu. Autonomous Discovery of Particle Physics Theories from Experimental Data, Mar. 2026

    2026

  39. [39]

    Alzubi, N

    S. Alzubi, N. Provenzano, J. Bingham, W. Chen, and T. Vu. Evoskill: Automated skill discovery for multi-agent systems.arXiv preprint arXiv:2603.02766, 2026

    Pith/arXiv arXiv 2026

  40. [40]

    Aygün, A

    E. Aygün, A. Belyaeva, G. Comanici, M. Coram, H. Cui, J. Garrison, R. Johnston, A. Kast, C. Y . McLean, P. Norgaard, Z. Shamsi, D. Smalling, J. Thompson, S. Venugopalan, B. P. Williams, C. He, S. Martinson, M. Plomecka, L. Wei, Y . Zhou, Q.-Z. Zhu, M. Abraham, E. Brand, A. Bulanova, J. A. Cardille, C. Co, S. Ellsworth, G. Joseph, M. Kane, R. Krueger, J. K...

    2026

  41. [41]

    D. A. Boiko, R. MacKnight, B. Kline, and G. Gomes. Autonomous chemical research with large language models.Nature, 624:570–578, 2023

    2023

  42. [42]

    Y . Chen, C. Liu, Z. Chen, T. Liu, B. Han, and K. Zhang. CausalEvolve: Towards Open-Ended Discovery with Causal Scratchpad, Mar. 2026

    2026

  43. [43]

    Cobelli and S

    M. Cobelli and S. Sanvito. Agentic design of compositional descriptors via autoresearch for materials science applications.arXiv preprint arXiv:2605.14671, 2026

    Pith/arXiv arXiv 2026

  44. [44]

    L. Fan, P. Dai, Z. Deng, H. Wang, X. Gong, Y . Zheng, and Y . Ou. Evolving Medical Imaging Agents via Experience-driven Self-skill Discovery, Mar. 2026. arXiv:2603.05860 [cs]

    arXiv 2026

  45. [45]

    Ferreira, L

    F. Ferreira, L. Wobbe, A. Krishnakumar, F. Hutter, and A. Zela. Can LLMs Beat Classical Hyperparameter Optimization Algorithms? A Study on autoresearch, Mar. 2026

    2026

  46. [46]

    A. E. Ghareeb, B. Chang, L. Mitchener, A. Yiu, C. J. Szostkiewicz, D. Shved, G. J. Gyimesi, J. M. Laurent, S. M. Wright, M. T. Razzak, A. D. White, S. C. Finnemann, M. M. Hinks, and S. G. Rodriques. A multi-agent system for automating scientific discovery.Nature, May 2026

    2026

  47. [47]

    Gottweis, W.-H

    J. Gottweis, W.-H. Weng, A. Daryin, T. Tu, P. Sirkovic, A. Myaskovsky, G. Glowaty, F. Weis- senberger, A. Orlandi, D. Popovici, A. Palepu, K. Rong, R. Tanno, K. Saab, F. Zhang, J. Blum, A. Carroll, K. Kulkarni, N. Tomašev, D. Zverinski, I. Rendulic, E. Vedadi, F. Hasler, L. Ri- manic, M. Boia, I. Budiselic, B. Feinstein, M. Bellaiche, T. Sheffer, J. Freyb...

    2026

  48. [48]

    T. Han, Y . Zhang, W. Song, C. Fang, Z. Chen, Y . Sun, and L. Hu. Swe-skills-bench: Do agent skills actually help in real-world software engineering?arXiv preprint arXiv:2603.15401, 2026

    arXiv 2026

  49. [49]

    C. He, X. Zhou, D. Wang, H. Xu, W. Liu, and C. Miao. The AutoResearch Moment: From Experimenter to Research Director, Mar. 2026

    2026

  50. [50]

    M. He, F. Jiang, J. Jiao, M. Li, K. Li, Y . Liao, B. Liu, T. Liu, F. Qi, Z. Shang, W. Song, Y . Sun, X. Wang, H. Wang, D. Xiong, C. Yuan, B. Zhang, Z. Zhang, and X. Zhu. Dr.Sai: An agentic AI for real-world physics analysis at BESIII, Apr. 2026. arXiv:2604.22541 [hep-ex] version: 1

    Pith/arXiv arXiv 2026

  51. [51]

    Huang, Y

    Y . Huang, Y . Chen, H. Zhang, K. Li, H. Zhou, M. Fang, L. Yang, X. Li, L. Shang, S. Xu, J. Hao, K. Shao, and J. Wang. Deep Research Agents: A Systematic Examination And Roadmap, Sept. 2025. arXiv:2506.18096 [cs]. 22

    arXiv 2025

  52. [52]

    V . Ilin. Semi-Autonomous Formalization of the Vlasov-Maxwell-Landau Equilibrium, Mar. 2026

    2026

  53. [53]

    B. Jia, S. Kamboj, S. Katipomu, S. H. Han, N. Sengupta, and A. Jackson. Nomad: Autonomous Exploration and Discovery, Mar. 2026

    2026

  54. [54]

    Jiang, Z

    G. Jiang, Z. Su, X. Qu, and Y . R. Fung. Xskill: Continual learning from experience and skills in multimodal agents.arXiv preprint arXiv:2603.12056, 2026

    arXiv 2026

  55. [55]

    Karpathy

    A. Karpathy. karpathy/autoresearch, Apr. 2026

    2026

  56. [56]

    Karwowski, O

    J. Karwowski, O. Hayman, X. Bai, K. Kiendlhofer, C. Griffin, and J. M. V . Skalse. Good- hart’s law in reinforcement learning. InThe Twelfth International Conference on Learning Representations, 2024

    2024

  57. [57]

    Khandelwal and S

    N. Khandelwal and S. S. Gupta. Agent-Driven Autonomous Reinforcement Learning Research: Iterative Policy Improvement for Quadruped Locomotion, Mar. 2026

    2026

  58. [58]

    L. Kong, X. Sun, W. Chow, L. Li, K. Q. Lin, X. B. Zhang, S. Wang, R. Li, Q. Wu, W. Gao, Y . Wang, S. Xie, J. Liu, L. Qu, S. Li, L. X. Ng, B. R. Cottereau, Z. Liu, T.-S. Chua, and W. T. Ooi. AI for Auto-Research: Roadmap & user guide. May 2026

    2026

  59. [59]

    Kuroki, T

    S. Kuroki, T. Nakamura, T. Akiba, and Y . Tang. Agent skill acquisition for large language models via cycleqd.arXiv preprint arXiv:2410.14735, 2024

    arXiv 2024

  60. [60]

    C.-Y . Lee, H. Liang, R. Kim, A. McDannald, C. A. R. Ocampo, A. G. Kusne, and I. Takeuchi. Real-time multi-instrument autonomous discovery of novel phase-change memory materials. May 2026

    2026

  61. [61]

    F. Li, P. Tagkopoulos, and I. Tagkopoulos. Skillflow: Scalable and efficient agent skill retrieval system.arXiv e-prints, pages arXiv–2504, 2025

    2025

  62. [62]

    H. Li, C. Mu, J. Chen, S. Ren, Z. Cui, Y . Zhang, L. Bai, and S. Hu. Organizing, orchestrating, and benchmarking agent skills at ecosystem scale.arXiv preprint arXiv:2603.02176, 2026

    arXiv 2026

  63. [63]

    X. Li. Auto Researching, not hyperparameter tuning: Convergence Analysis of 10,000 Experiments, Mar. 2026

    2026

  64. [64]

    X. Li. When single-agent with skills replace multi-agent systems and when they fail.arXiv preprint arXiv:2601.04748, 2026

    arXiv 2026

  65. [65]

    X. Li, W. Chen, Y . Liu, S. Zheng, X. Chen, Y . He, Y . Li, B. You, H. Shen, J. Sun, et al. Skillsbench: Benchmarking how well agent skills work across diverse tasks.arXiv preprint arXiv:2602.12670, 2026

    Pith/arXiv arXiv 2026

  66. [66]

    Y . Li, C. Shao, X. Liu, R. Zhao, P. Liu, H. Su, Z. Chen, Q. Yang, A. Xu, Y . Fang, et al. Autosota: An end-to-end automated research system for state-of-the-art ai model discovery. arXiv preprint arXiv:2604.05550, 2026

    Pith/arXiv arXiv 2026

  67. [67]

    Liang, R

    Y . Liang, R. Zhong, H. Xu, C. Jiang, Y . Zhong, R. Fang, J.-C. Gu, S. Deng, Y . Yao, M. Wang, et al. Skillnet: Create, evaluate, and connect ai skills.arXiv preprint arXiv:2603.04448, 2026

    arXiv 2026

  68. [68]

    C. Liu, T. Li, M. Huang, X. Wei, P. Liu, Y . Shen, Y . Mao, and T. Cui. Protrlsearch: A multi- round multimodal protein search agent with large language models trained via reinforcement learning.arXiv preprint arXiv:2603.01464, 2026

    arXiv 2026

  69. [69]

    F. Liu, J. Han, T. Lyu, W. Zhang, Z.-R. Yang, L. Dai, C. Liu, and H. Liu. Foundation models for scientific discovery: From paradigm enhancement to paradigm transition.Advances in Neural Information Processing Systems, 2025

    2025

  70. [70]

    F. Liu, J. Xu, X. Cui, X. Wang, Z. Guo, J. Wang, S. M. Mousavi, X. Gu, H. Chen, B. Fei, L. Fang, F. Ling, Z. Li, and L. Bai. TRACE: A Multi-Agent System for Autonomous Physical Reasoning for Seismology, Mar. 2026. 23

    2026

  71. [71]

    J. Liu, Z. Ling, S. Qiu, Y . Liu, S. Han, P. Xia, H. Tu, Z. Zheng, C. Xie, C. Fleming, et al. Omni-simplemem: Autoresearch-guided discovery of lifelong multimodal agent memory. arXiv e-prints, pages arXiv–2604, 2026

    2026

  72. [72]

    J. Liu, S. Qiu, M. Li, B. Li, H. Ji, S. Han, X. Ye, P. Xia, Z. Dong, C. Zhang, L. Zhang, G. Chen, H. Tu, X. Yang, L. Feng, X. Zhao, H. Chen, J. Zhou, X. Wang, W. Zhang, H. Zhu, Y . Li, J. Mei, H. Fei, J. Zhang, L. Li, L. Zhang, Y . Zhou, S. Wang, C. Xiong, J. Zou, Z. Zheng, C. Xie, M. Ding, and H. Yao. AutoResearchClaw: Self-reinforcing autonomous researc...

    2026

  73. [73]

    J. Liu, J. Shen, S. Song, T. Li, X. Liu, R. Li, Z. Huang, J. Lin, J. Ning, C. Ji, S. Luo, W. Li, C. Ma, M. Hu, J. Xiong, J. Ye, B. Fu, N. Xu, Y . Chen, L. Jin, H. Chen, and J. He. MedProbeBench: Systematic Benchmarking at Deep Evidence Integration for Expert-level Medical Guideline, Apr. 2026. arXiv:2604.18418 [cs] version: 1

    Pith/arXiv arXiv 2026

  74. [74]

    J. Liu, X. Ye, P. Xia, Z. Zheng, C. Xie, M. Ding, and H. Yao. Evolvemem: Self-evolving memory architecture via autoresearch for llm agents.arXiv preprint arXiv:2605.13941, 2026

    Pith/arXiv arXiv 2026

  75. [75]

    C. Lu, C. Lu, R. T. Lange, Y . Yamada, S. Hu, J. Foerster, D. Ha, and J. Clune. Towards end-to-end automation of AI research.Nature, 651(8107):914–919, Mar. 2026

    2026

  76. [76]

    Manheim and S

    D. Manheim and S. Garrabrant. Categorizing variants of goodhart’s law.arXiv preprint arXiv:1803.04585, 2018

    Pith/arXiv arXiv 2018

  77. [77]

    Messeri and M

    L. Messeri and M. J. Crockett. Artificial intelligence and illusions of understanding in scientific research.Nature, 627:49–58, 2024

    2024

  78. [78]

    J. Ni, Y . Liu, X. Liu, Y . Sun, M. Zhou, P. Cheng, D. Wang, X. Jiang, and G. Jiang. Trace2skill: Distill trajectory-local lessons into transferable agent skills.arXiv preprint arXiv:2603.25158, 2026

    Pith/arXiv arXiv 2026

  79. [79]

    Novikov, M

    A. Novikov, M. Balog, M. P. Kumar, et al. AlphaEvolve: A coding agent for scientific and algorithmic discovery, 2025

    2025

  80. [80]

    Introducing deep research, 2025

    OpenAI. Introducing deep research, 2025

    2025

Showing first 80 references.