{"slug": "the-red-queen-hypothesis-a-new-way-forward-for-self-improving-ai", "title": "The Red Queen hypothesis – a new way forward for self-improving AI", "summary": "A research team including collaborators from NVIDIA and Flower Labs has published the arXiv pre-print \"The Red Queen Gödel Machine: Co-Evolving Agents and Their Evaluators,\" a method that lets self-improving AI agents and their evaluators evolve together so agents avoid the ceiling imposed by a fixed test suite. First author Alex Iacob, a PhD student in Cambridge's Machine Learning Systems Lab supervised by Prof Nic Lane, reported that co-evolved scientific paper writers reach 1.78×–1.86× higher acceptance rates under a diverse agent-as-a-judge panel, while co-evolved graders reach 9% higher ground-truth accuracy. The method also suggests a way to cut the computational cost of developing such agents, addressing the limit Iacob described: \"A self-improving agent can only get as good as the test that scores it.", "body_md": "# The Red Queen hypothesis - a new way forward for self-improving AI\n\n## At a time when there's keen public interest in AI that can make itself better, researchers here have tackled one of the central challenges affecting its development.\n\nThe research team, which includes collaborators from NVIDIA and Flower Labs, have come up with a new method for recursive self-improving AI agents to continue improving themselves (by repeatedly testing and enhancing their own code) without hitting the evaluation ceiling that they frequently encounter.\n\nTheir method also suggests a way of cutting the costs of the computational resource needed for the development of such AI agents.\n\nWhile agents can already improve themselves by editing their own code, testing variants, and keeping changes that perform better, this process is usually limited by a fixed evaluator, benchmark, or test suite. Once the agent has learned everything that fixed signal can distinguish, improvement slows or stalls.\n\n\"A self-improving agent can only get as good as the test that\nscores it,\" explains team member **[Alex\nIacob](https://www.linkedin.com/in/alex-iacob-898775194/)**, a PhD student in the **[Machine Learning Systems\nLab](https://mlsys.cst.cam.ac.uk/)** under the supervision of\nProf **[Nic\nLane](https://www.cl.cam.ac.uk/people/ndl32)**. \"The test does not merely measure progress, it\ndefines it, so the efficacy of the test becomes a ceiling the agent\ncannot climb past.\"\n\nNow the researchers have addressed this issue by having both the\nself-improving agent and the evaluator evolve together. \"Instead\nof improving an agent against a fixed test, we let\nthe **evaluation evolve alongside the agent**,\" Alex\nadds. \"As the agent gets better, the evaluation also\ngets **harder**, and the bar keeps rising.\"\n\n### The Red Queen Gödel Machine\n\nAlex is the first author on the pre-print paper the research team\nhas just uploaded to arXiv. **[The Red Queen Gödel Machine:\nCo-Evolving Agents and Their Evaluators](https://arxiv.org/abs/2606.26294)** shares the\ntechnical details of their work, along with some impressive results\nfrom using this framework across a number of tasks.\n\n*Figure: Agents and evaluators improving together. The Red Queen\nGödel Machine searches through many possible versions of an AI agent,\nwhile also improving the evaluator that judges those agents. During\neach phase, the evaluator is kept fixed so progress can be measured\nreliably. At checkpoints, a stronger evaluator can replace the old one\nif it performs better on trusted ground-truth examples. Scores from\nthe old evaluator are then removed, so the next phase is guided by the\nnew, more demanding standard. This creates a self-improving loop in\nwhich agents and evaluators improve together, while the system remains\nanchored to reliable checks.*\n\nIn both scientific paper writing and reviewing, and (Maths) Olympiad-level proof writing and grading, the Red Queen Gödel Machine improved performance over previous self-improving AI agents.\n\n\"Co-evolved scientific paper writers reach 1.78×–1.86× higher acceptance rates under a diverse agent-as-a-judge panel,\" report the researchers in the paper, \"while co-evolved graders reach 9% higher ground-truth accuracy.\"\n\n### The Red Queen hypothesis\n\nThe framework's curious title references the Red Queen – a\nfictional character in Lewis Carroll's children's novel, *Through\nthe Looking-Glass* – who famously tells Alice, the novel's\nheroine, that \"it takes all the running you can do, to keep in the\nsame place.\" The Red Queen hypothesis, a hypothesis in evolutionary\nbiology put forward in 1973, was named after the character as it\nproposes that species must constantly adapt, evolve, and proliferate\nin order to survive while pitted against other species that are also\ncontinually evolving.\n\nThis hypothesis has now been applied to AI.\n\n\"Instead of improving an agent against a fixed test, we let the evaluation evolve alongside the agent. As the agent gets better, the evaluation also gets harder, and the bar keeps rising.\"\n\nAlex Iacob, PhD student\n\nThe work was carried out by a team of researchers here and with the\nsupport of collaborators including NVIDIA, Departmental spin-out\n[**Flower Labs**](https://flower.ai/), MBZUAI\nand Inria.\n\n### A path towards more capable open agent systems at lower cost\n\nAnd another key finding that emerged was that using open source models to carry out some of the work, in addition to computationally expensive tools like ChatGPT, could be a way of significantly lowering the costs of developing self-improving AI systems.\n\nIn one experiment by\nthe researchers, to co-evolve AI reviewers and writers of scientific\npapers, the researchers used the **[NVIDIA\nNemotron 3 Ultra](https://www.nvidia.com/en-us/ai-data-science/foundation-models/nemotron/),** alongside ChatGPT-5.5. In the\npaper-reviewing task, this hybrid setup approached the performance\nachieved by ChatGPT on its own, while reducing search-token costs by\naround 13 times.\n\nCo-author Professor Nic Lane says: \"This is a narrow result, and we should be careful not to overstate it. But it also indicates where this method can go. If open models can carry the bulk of the search while stronger frontier models guide the higher-level improvement process, then we have a plausible path toward much more capable open agent systems at far lower cost.\"\n\nAnd Daniel Burkhardt, Developer Relations Manager at NVIDIA, says: \"This effort shows how open models can play an important role in advanced agentic systems. NVIDIA Nemotron models are designed to be efficient and capable for reasoning and agent workloads, and this work shows how they can be used as part of a broader self-improvement loop.\"\n\nThe research team is at pains to point out that the work is preliminary, and that longer search horizons will be needed to understand how far the approach can scale. But the results point to a new class of self-improving systems in which agents and evaluators recursively bootstrap one another, opening a path toward more capable, efficient and open agentic AI.\n\nThe paper, The Red Queen Gödel Machine: Co-Evolving Agents and Their Evaluators, is available on arXiv. The underlying method will be open sourced to support wider adoption, while the team continues to investigate how this self-improving approach performs across a broader range of AI tasks, including experiments being scaled significantly beyond these initial results.\n\n- [**The Red Queen\nGödel Machine: Co-Evolving Agents and Their Evaluators**](https://arxiv.org/pdf/2606.26294) :\nAlex Iacob, Andrej Jovanović, William F. Shen, Daniel\nBurkhardt, Meghdad Kurmanji , Nurbek Tastan, Lorenzo Sani,\nNiccolò Alberto Elia Venanzi, Ambroise Odonnat, Zeyu Cao, Bill Marino,\nXinchi Qiu, Nicholas D. Lane", "url": "https://wpnews.pro/news/the-red-queen-hypothesis-a-new-way-forward-for-self-improving-ai", "canonical_source": "https://www.cl.cam.ac.uk/news/2026/07-21-red-queen-hypothesis/", "published_at": "2026-11-10 03:57:01+00:00", "updated_at": "2026-11-10 04:22:17.798508+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-research", "ai-agents", "ai-safety"], "entities": ["Alex Iacob", "Nic Lane", "NVIDIA", "Flower Labs", "Machine Learning Systems Lab", "University of Cambridge", "The Red Queen Gödel Machine", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/the-red-queen-hypothesis-a-new-way-forward-for-self-improving-ai", "markdown": "https://wpnews.pro/news/the-red-queen-hypothesis-a-new-way-forward-for-self-improving-ai.md", "text": "https://wpnews.pro/news/the-red-queen-hypothesis-a-new-way-forward-for-self-improving-ai.txt", "jsonld": "https://wpnews.pro/news/the-red-queen-hypothesis-a-new-way-forward-for-self-improving-ai.jsonld"}}