betaFeedback

WorksComputer Science

Adapting the Process of Adaptation: Darwinian, Lamarckian, and Agent-Directed Evolution in AI

Abstract

Evolution has shaped intelligence for billions of years. Artificial intelligence (AI) now raises a striking possibility: intelligence may begin to direct its own evolution. As AI agents become capable of modifying and evaluating their own behavior, how quickly could they improve, and how can that improvement remain safe? We develop a theoretical framework comparing Darwinian, Lamarckian, and agent-directed evolution following environmental change. Darwinian evolution relies on natural selection, Lamarckian evolution adds inherited learning, and agent-directed evolution allows agents to generate and evaluate behavioral variants and choose how extensively to search. When learning is sufficiently effective, these mechanisms exhibit progressively faster adaptation. Generating more variants accelerates adaptation, while optimal search effort depends on learning ability, problem difficulty, and experimentation costs. However, an agent's ability to direct its improvement does not ensure that improvements are evaluated against the intended objective. When the evaluation target is misspecified, additional search accelerates convergence toward the wrong behavior, even as measured performance improves. In recursive self-improvement, such errors may persist across generations, posing challenges for AI safety. Independent evaluation and oversight must keep pace with autonomous AI development. The central challenge is not simply to accelerate the evolution of intelligence, but to preserve humanity's ability to guide its direction, limit its speed, and intervene when necessary.

Introduction

On 8 September 2026, OpenAI announced that a coordinated system of artificial intelligence (AI) agents had produced a proposed resolution of the Navier–Stokes Millennium Prize problem, showing that a finite-time singularity forms under smooth external forcing [33]. According to OpenAI, roughly 10,000 concurrent agents explored different formulations and approaches, shared intermediate findings, and built on promising results. They reached the proposed solution in about 88 hours and then spent 17 hours formalizing and verifying it in Lean [12].

This search resembles evolution: variants are generated, promising ones are retained, and useful information is passed on to later attempts. FunSearch [35], AlphaEvolve [31], and the Darwin Gödel Machine [42] follow the same principle: AI systems propose modifications to programs or agents, evaluate them, and build on the successful ones.

These ideas come together in an emerging research direction, recursive self-improvement (RSI), in which an AI system improves itself and uses the improved version to make further improvements. In the Darwin Gödel Machine, AI agents modify their own code, evaluate the resulting variants, and keep the successful changes for later generations [42]. More recently, Srikanth et al. [40] developed AIDE2, an AI research agent that recursively modifies its own code and evaluates the results. These studies suggest that AI systems are increasingly able not only to perform complex tasks but also to improve the processes through which they perform them.

These developments point to a more fundamental possibility: AI systems may increasingly take control of their own evolution. Instead of improving within a process designed by humans, a system may decide how it searches for improvements, how extensively it experiments, and how it carries improvements forward to later generations. Because improved systems can produce further improvements, this could greatly accelerate AI development and make its capabilities harder for humans to anticipate and oversee. We call this shift agent-directed evolution. It holds great potential for scientific discovery and technological progress, but it also raises safety concerns as AI systems gain greater control over their own development.

This possibility motivates several questions. How much faster can a system adapt through learning and self-directed search than through natural selection alone? What determines the speed of adaptation, and how should an agent weigh the benefits of more search against its costs? More importantly, does faster adaptation necessarily lead to better outcomes, or can it also accelerate the development of behaviors that humans did not intend?

To address these questions, we build on the evolutionary framework of Brennan and Lo [4, 5], in which individuals inherit different behaviors and natural selection favors the behaviors that lead to faster population growth. When the environment changes, previously successful behaviors may no longer be optimal, and the population must adapt. In § 3, we show that the growth lost to suboptimal behavior equals the Kullback–Leibler (KL) divergence between the environmental and behavioral distributions (Propositions 2 and 3). We call this loss growth regret and use it to compare how quickly Darwinian, Lamarckian, and agent-directed evolution adapt.

The three mechanisms start from the same population. Under Darwinian evolution, inherited behaviors are fixed, and the population adapts as natural selection makes better-adapted individuals more prevalent (§ 4). Under Lamarckian evolution, individuals can also learn from their environment and pass their acquired improvements to later generations (§ 5). Under agent-directed evolution, individuals further generate multiple variants of themselves, evaluate them, and retain the best for further learning (§ 6). More importantly, the agent decides how extensively to search, which we use as a simple model of how an intelligent system can influence its own evolution. We measure speed by the adaptation time, the number of generations needed to reduce growth regret to a given threshold, and the adaptation rate, the long-run exponential rate at which growth regret declines.

Our first set of results characterizes the adaptation speeds of the three mechanisms. Darwinian adaptation relies on natural selection to spread individuals who already have the optimal behavior, so it can take arbitrarily long when they are initially rare (Theorem 1). Lamarckian evolution removes this dependence, because individuals improve their inherited behavior directly through learning (Theorem 2). Agent-directed evolution adapts even faster by searching among alternative improvements (Theorem 3). In § 7, we show that, when learning is sufficiently effective, agent-directed evolution has the highest long-run adaptation rate, followed by Lamarckian and then Darwinian evolution; the same ranking holds for expected adaptation times (Theorem 4).

Our second set of results, in § 6, concerns how search accelerates adaptation and how an agent should allocate its search effort. Generating more variants always increases the adaptation rate, by an amount that depends on the difficulty of the problem (Proposition 4). For easier problems, the gain can be substantial: in the most favorable case, it grows approximately as 2log⁡N2 \log N, where NN is the number of variants. For harder problems, the gain is smaller and approaches a finite ceiling, although more variants still speed up adaptation. When search is costly, the optimal number of variants increases as learning becomes less effective, the problem becomes easier, or search becomes cheaper (Proposition 5). An intelligent system can thus adjust its search effort according to its learning ability, problem difficulty, and available resources.

Our third set of results concerns AI safety: faster adaptation does not necessarily move a system toward the objectives its developers intend. In § 8, we extend the agent-directed mechanism so that search and learning are guided by an incorrectly specified evaluation target. The agent then keeps improving rapidly according to its evaluator, but its true growth regret does not vanish, and more variants only accelerate convergence toward the incorrect target (Theorem 5). Natural selection, acting on realized outcomes, can eventually correct the population’s behavior, but this correction may be slow. Independent evaluation and oversight therefore become increasingly important.

Our analysis brings together ideas from evolutionary biology, economics, and AI: drawing on the biological mechanisms of selection and inheritance, we use a growth-based framework to study how learning and search affect the speed and cost of adaptation in AI systems. Appendix § 12 summarizes how the main elements of our model correspond to their counterparts in evolutionary biology and modern AI systems.

The remainder of this paper proceeds as follows. § 2 reviews the related literature, and § 3 introduces the Brennan–Lo framework. § 4 to § 6 analyze Darwinian, Lamarckian, and agent-directed evolution, respectively, and § 7 compares their adaptation speeds. § 8 examines adaptation toward an incorrectly specified target and its implications for AI safety. § 9 concludes.

Related Literature

We review six areas of research in evolutionary biology, information theory, economics, and AI that bear on the speed and direction of adaptation.

Natural Selection and Growth-Optimal Behavior. We build on the evolutionary model of Brennan and Lo [4], in which natural selection favors inherited behaviors that lead to higher population growth. Later studies extend this framework to bounded rationality and intelligence (Brennan and Lo [5]), behavioral diversification (Brennan, Lo, and Zhang [6]), and learning under limited information and memory (Lo and Zhang (2021); Lo, Zhang, and Zhao (2025)). In evolutionary biology, bet hedging describes how organisms diversify their offspring’s behaviors to cope with environmental uncertainty (Cohen [10]; Bull [8]; Seger and Brockmann [37]). Building on this literature, we study how quickly a population adapts after an environmental change.

Information Theory and Evolutionary Growth. Information and growth have long been linked. In the framework of Kelly [22], the growth lost by using incorrect probabilities can be expressed in terms of the KL divergence [11 Chapter 6], and evolutionary biology studies the fitness value of information about the environment [23, 13, 34]. We establish an analogous relationship between population growth and KL divergence, which lets us measure the loss from suboptimal behavior and compare how quickly different evolutionary mechanisms eliminate it.

Learning and the Inheritance of Acquired Traits. Previous research distinguishes learning that affects an individual’s performance from learning whose improvements are inherited. In the Baldwin effect, individual learning can facilitate evolution even without the inheritance of acquired traits [20, 1], whereas Lamarckian evolutionary algorithms pass improvements acquired through learning to offspring [41]. Related ideas appear in population-based training [21] and in methods that transfer learned knowledge across generations of neural networks [16]. Our Lamarckian mechanism isolates the inheritance of acquired improvements and examines how this channel changes the speed of adaptation under natural selection.

Adaptive Search and Resource Allocation. Evolutionary computation has long studied how to adjust search parameters, such as mutation rates and the distribution of candidate solutions, to improve performance [14, 19, 2, 32]; autoconstructive evolution goes further and lets the mechanisms of reproduction and variation evolve within the individuals themselves [39]. Economic models examine how agents adjust their learning rules to changing environments or forecasting performance [15, 28]. More recent research on inference-time computation studies how the benefit of generating more candidates depends on problem difficulty and computational resources [7, 38]. Our agent-directed mechanism poses a related decision in an evolutionary setting, how many variants to generate, and we characterize how search raises the rate of adaptation and how the optimal number of variants depends on problem difficulty, learning ability, and search costs.

RSI in AI. The possibility that intelligent machines could improve their own capabilities has been discussed since Good [18]; an early formal approach is the Gödel machine of Schmidhuber [36]. FunSearch and AlphaEvolve use AI to generate, evaluate, and retain improved solutions [35, 31], and the Darwin Gödel Machine [42], AIDE2 [40], and multi-agent RSI [24] extend this process to AI agents themselves. These developments have prompted discussion of whether AI evolution will remain directed by humans or become increasingly autonomous [3, 29, 43]. Our analysis offers a theoretical perspective by studying how an agent’s control over its own search effort affects the speed and cost of adaptation.

Evaluation and the Direction of Adaptation. The effectiveness of search depends on how candidate improvements are evaluated. Evolutionary algorithms can exploit evaluation criteria in unintended ways [25], and stronger optimization against learned reward models can widen the gap between measured reward and true quality [17]. We examine how an incorrectly specified evaluation target affects evolutionary adaptation and show that more search can accelerate adaptation toward the evaluator’s target without correcting the error.

The Framework

To compare different mechanisms of adaptation, we need a framework that identifies both the optimal behavior and the cost of deviating from it. We adopt the evolutionary model of Brennan and Lo [4], in which natural selection favors behaviors that maximize long-run population growth. § 3.1 studies the growth of a lineage, § 3.2 extends the analysis to a population with multiple lineages, and § 3.3 introduces the three evolutionary mechanisms, establishes a common setting for comparing them, and defines the adaptation time and rate.

Growth of a lineage

Consider a large lineage of asexually reproducing individuals. In generation tt, all individuals inherit the same behavior ft∈(0,1)f_t \in(0,1), choosing action aa with probability ftf_t and action bb with probability 1−ft1-f_t. The three evolutionary mechanisms considered in this study differ in whether and how ftf_t changes across generations. Individuals in generation tt reproduce to form generation t+1t+1, and their reproductive success depends on the chosen action and the environmental state Xt+1∈{α,β}X_{t+1} \in\{\alpha, \beta\}, which is shared by all individuals. Let xax_a and xbx_b denote the numbers of offspring produced by an individual choosing aa and bb, respectively. Then

(xa,xb)={(m,0),Xt+1=α,with probability p,(0,m),Xt+1=β,with probability 1−p,(x_a,x_b)= \begin{cases} (m,0), & X_{t+1}=\alpha,\quad\text{with probability }p,\\ (0,m), & X_{t+1}=\beta,\quad\text{with probability }1-p, \end{cases}

where m>0m>0 is the number of offspring produced by an individual choosing the successful action: action aa succeeds in state α\alpha, and action bb succeeds in state β\beta. Individuals choosing the other action produce no offspring. The probability of state α\alpha is p∈(0,1)p\in(0,1), and the environmental states X1,X2,…X_1,X_2,\ldots are independent across generations.

Let ntn_t denote the size of the lineage in generation tt. Because the lineage is large, a fraction ftf_t of its individuals chooses action aa, while the remaining fraction 1−ft1-f_t chooses action bb. If Xt+1=αX_{t+1}=\alpha, only those choosing aa reproduce, each producing mm offspring, so the lineage grows by the factor nt+1/nt=mftn_{t+1}/n_t=mf_t. Similarly, if Xt+1=βX_{t+1}=\beta, the lineage grows by the factor nt+1/nt=m(1−ft)n_{t+1}/n_t=m(1-f_t). Taking the expected logarithm of these growth factors gives

Ip(ft)≡E[log⁡nt+1nt∣ft]=plog⁡(mft)+(1−p)log⁡(m(1−ft))=log⁡m+plog⁡ft+(1−p)log⁡(1−ft).(1)\begin{aligned} I_p(f_t)\equiv\mathbb{E}\left[\left.\log\frac{n_{t+1}}{n_t}\right|f_t\right] =p\log(mf_t)+(1-p)\log\left(m(1-f_t)\right) \\ &=\log m+p\log f_t+(1-p)\log(1-f_t). \tag*{(1)} \end{aligned}

We refer to Ip(f)I_p(f) as the growth rate associated with behavior ff.

Given the growth rate in Equation (1), we can now identify the inherited behavior that maximizes population growth.

Proposition 1 (Probability matching). For p∈(0,1)p\in(0,1), IpI_p is strictly concave on (0,1)(0,1), and

arg max⁡f∈(0,1) Ip(f)=p.\underset{f\in(0,1)}{\operatorname{arg\,max}}\ I_p(f)=p.

The proposition shows that a lineage maximizes its expected log growth in every generation by matching the fraction of individuals choosing each action to the probability that the action succeeds. The intuition is that diversification is essential when all individuals face the same environmental risk. A lineage that always chooses one action will eventually become extinct when the other action succeeds, whereas a lineage that divides its individuals between the two actions can reproduce in either environmental state. This is the probability-matching result of Brennan and Lo [4].

To compare how different evolutionary mechanisms adapt to a changing environment, we need a measure of how much growth a lineage loses when its behavior deviates from the optimum. The following proposition defines this loss and establishes its connection to information theory.

Proposition 2 (Growth regret is relative entropy). For p,f∈(0,1)p,f \in(0,1), the growth regret of behavior ff is

Rp(f)≡Ip(p)−Ip(f)=plog⁡pf+(1−p)log⁡1−p1−f=DKL(Bern⁡(p)∥Bern⁡(f))≥0,(2)\mathcal{R}_p(f) \equiv I_p(p)-I_p(f)=p\log\frac{p}{f}+(1-p)\log\frac{1-p}{1-f}=D_{\mathrm{KL}}\bigl(\operatorname{Bern}(p)\mathbin{\|}\operatorname{Bern}(f)\bigr)\geq0, \tag*{(2)}

where Bern⁡(θ)\operatorname{Bern}(\theta) denotes the Bernoulli distribution with success probability θ\theta, and DKLD_{\mathrm{KL}} denotes the KL divergence. Equality holds if and only if f=pf=p.

The proposition shows that growth regret can be expressed exactly as the KL divergence between the environmental distribution and the inherited behavioral distribution. This result is closely related to the Kelly criterion [22], which characterizes the cost of betting according to incorrect probabilities. For a lineage with the behavior ftf_t in generation tt, Rp(ft)\mathcal{R}_p(f_t) is the expected log growth that it loses from generation tt to generation t+1t+1 relative to a lineage with the optimal behavior pp. Growth regret thus provides a natural measure of maladaptation, allowing us to compare the three evolutionary mechanisms in terms of how quickly they recover the growth lost after an environmental change.

To understand the magnitude of growth regret and how it varies with behavior ff, consider its Taylor expansion around the optimal behavior f=pf=p:

Rp(f)=(f−p)22p(1−p)+O((f−p)3),f→p.(3)\mathcal{R}_p(f)=\frac{(f-p)^2}{2p(1-p)}+O\bigl((f-p)^3\bigr),\qquad f\to p. \tag*{(3)}

The expansion shows that, for small deviations, growth regret increases approximately with the square of the distance from the optimum. Doubling the deviation therefore approximately quadruples the loss in growth.

Growth of a population

So far, we have considered a single lineage whose individuals share the same behavior within each generation. We now extend the analysis to a population consisting of several lineages, each with its own behavior. For simplicity, we focus on two lineages, although the results extend to more. In generation tt, let f1,tf_{1,t} and f2,tf_{2,t} denote their behaviors, qtq_t and 1−qt1-q_t their population shares, and ntn_t the size of the entire population. The two lineages therefore contain qtntq_t n_t and (1−qt)nt(1-q_t)n_t individuals, respectively.

To determine the growth of the entire population, we first consider the offspring produced by each lineage. If Xt+1=αX_{t+1}=\alpha, only individuals choosing action aa reproduce. The first lineage therefore grows by the factor mf1,tm f_{1,t}, contributing mf1,tqtntm f_{1,t}q_t n_t individuals to the next generation. Similarly, the second lineage contributes mf2,t(1−qt)ntm f_{2,t}(1-q_t)n_t individuals. Adding the two contributions, the population has mf1,tqtnt+mf2,t(1−qt)nt=mfˉtntm f_{1,t}q_t n_t + m f_{2,t}(1-q_t)n_t = m\bar{f}_t n_t individuals in generation t+1t+1, where

fˉt≡qtf1,t+(1−qt)f2,t(4)\bar{f}_t \equiv q_t f_{1,t} + (1-q_t)f_{2,t} \tag*{(4)}

is the average behavior of the population. If Xt+1=βX_{t+1}=\beta, only individuals choosing action bb reproduce. The two lineages then contribute m(1−f1,t)qtntm(1-f_{1,t})q_t n_t and m(1−f2,t)(1−qt)ntm(1-f_{2,t})(1-q_t)n_t individuals, respectively, giving a total population of m(1−fˉt)ntm(1-\bar{f}_t)n_t in generation t+1t+1.

We can also use these results to determine how the relative shares of the two lineages evolve through natural selection. The share of the first lineage in generation t+1t+1 is its number of offspring divided by the total population size:

References

  1. [1]Ackley, D., and M. Littman, 1992, Interactions between learning and evolution, in C. G. Langton, C. Taylor, J. D. Farmer, and S. Rasmussen, eds., Artificial Life II, 487–509 (Addison-Wesley).
  2. [2]Bedau, M. A., and N. H. Packard, 2003, Evolution of evolvability via adaptation of mutation rates, BioSystems 69, 143–162.pubmed.ncbi.nlm.nih.gov/12689727
  3. [3]Boudry, M., 2026, Domesticated, not feral: Why evolvable AI is not yet a Darwinian threat, Proceedings of the National Academy of Sciences 123, e2617785123.DOI
  4. [4]Brennan, T. J., and A. W. Lo, 2011, The origin of behavior, Quarterly Journal of Finance 1, 55–108.
  5. [5]Brennan, T. J., and A. W. Lo, 2012, An evolutionary model of bounded rationality and intelligence, PLoS ONE 7, e50310.DOI
  6. [6]Brennan, T. J., A. W. Lo, and R. Zhang, 2018, Variety is the spice of life: Irrational behavior as adaptation to stochastic environments, Quarterly Journal of Finance 8, 1850009.
  7. [7]Brown, B., J. Juravsky, R. Ehrlich, R. Clark, Q. V. Le, C. Ré, and A. Mirhoseini, 2024, Large language monkeys: Scaling inference compute with repeated sampling, arXiv:2407.21787.arxiv.org/abs/2407.21787
  8. [8]Bull, J. J., 1987, Evolution of phenotypic variance, Evolution 41, 303–315.DOI
  9. [9]Cagan, P., 1956, The monetary dynamics of hyperinflation, in M. Friedman, ed., Studies in the Quantity Theory of Money, 25–117 (University of Chicago Press, Chicago).
  10. [10]Cohen, D., 1966, Optimizing reproduction in a randomly varying environment, Journal of Theoretical Biology 12, 119–129.pubmed.ncbi.nlm.nih.gov/6015423
  11. [11]Cover, T. M., and J. A. Thomas, 2006, Elements of Information Theory, second edition (Wiley, Hoboken, NJ).
  12. [12]de Moura, L., and S. Ullrich, 2021, The Lean 4 theorem prover and programming language, in Automated Deduction – CADE 28, volume 12699 of Lecture Notes in Computer Science, 625–635 (Springer).DOI
  13. [13]Donaldson-Matasci, M. C., C. T. Bergstrom, and M. Lachmann, 2010, The fitness value of information, Oikos 119, 219–230.
  14. [14]Eiben, A. E., R. Hinterding, and Z. Michalewicz, 1999, Parameter control in evolutionary algorithms, IEEE Transactions on Evolutionary Computation 3, 124–141.
  15. [15]Evans, G. W., and S. Honkapohja, 2001, Learning and Expectations in Macroeconomics (Princeton University Press, Princeton, NJ).
  16. [16]Feng, F., J. Wang, X. Yang, and X. Geng, 2025, Learngene: Inheritable “genes” in intelligent agents, Artificial Intelligence 348, 104421.DOI
  17. [17]Gao, L., J. Schulman, and J. Hilton, 2023, Scaling laws for reward model overoptimization, in Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, 10835–10866.arxiv.org/abs/2210.10760
  18. [18]Good, I. J., 1965, Speculations concerning the first ultraintelligent machine, in Advances in Computers, volume 6, 31–88 (Academic Press).DOI
  19. [19]Hansen, N., and A. Ostermeier, 2001, Completely derandomized self-adaptation in evolution strategies, Evolutionary Computation 9, 159–195.pubmed.ncbi.nlm.nih.gov/11382355
  20. [20]Hinton, G. E., and S. J. Nowlan, 1987, How learning can guide evolution, Complex Systems 1, 495–502.
  21. [21]Jaderberg, M., V. Dalibard, S. Osindero, W. M. Czarnecki, J. Donahue, A. Razavi, O. Vinyals, T. Green, I. Dunning, K. Simonyan, C. Fernando, and K. Kavukcuoglu, 2017, Population based training of neural networks, arXiv:1711.09846.arxiv.org/abs/1711.09846
  22. [22]Kelly, J. L., 1956, A new interpretation of information rate, Bell System Technical Journal 35, 917–926.
  23. [23]Kussell, E., and S. Leibler, 2005, Phenotypic diversity, population growth, and information in fluctuating environments, Science 309, 2075–2078.pubmed.ncbi.nlm.nih.gov/16123265
  24. [24]Lee, H., J. Xu, J. Seely, D. Lee, S. Sojoudi, M. Zaharia, and Y. Tang, 2026, Recursive self-improvement through multi-agent self-supervision, arXiv:2610.12176.arxiv.org/abs/2610.12176
  25. [25]Lehman, J., J. Clune, D. Misevic, et al., 2020, The surprising creativity of digital evolution: A collection of anecdotes from the evolutionary computation and artificial life research communities, Artificial Life 26, 274–306.
  26. [28]Marcet, A., and J. P. Nicolini, 2003, Recurrent hyperinflations and learning, American Economic Review 93, 1476–1498.
  27. [29]Müller, V., L. Steels, and E. Szathmáry, 2026, Evolvable AI: Threats of a new major transition in evolution, Proceedings of the National Academy of Sciences 123, e2527700123.DOI
  28. [30]Nerlove, M., 1958, Adaptive expectations and cobweb phenomena, Quarterly Journal of Economics 72, 227–240.DOI
  29. [31]Novikov, A., N. Vũ, M. Eisenberger, E. Dupont, P.-S. Huang, A. Z. Wagner, S. Shirobokov, B. Kozlovskii, F. J. R. Ruiz, A. Mehrabian, M. P. Kumar, A. See, S. Chaudhuri, G. Holland, A. Davies, S. Nowozin, P. Kohli, and M. Balog, 2025, AlphaEvolve: A coding agent for scientific and algorithmic discovery, arXiv:2506.13131.arxiv.org/abs/2506.13131
  30. [32]Ollivier, Y., L. Arnold, A. Auger, and N. Hansen, 2017, Information-geometric optimization algorithms: A unifying picture via invariance principles, Journal of Machine Learning Research 18, 1–65.
  31. [33]OpenAI, 2026, On the Navier–Stokes Millennium Prize Problem, https://openai.com/index/navier-stokes-solution/, published 8 September 2026; accessed 1 October 2026.
  32. [34]Rivoire, O., and S. Leibler, 2011, The value of information for populations in varying environments, Journal of Statistical Physics 142, 1124–1166.DOI
  33. [35]Romera-Paredes, B., M. Barekatain, A. Novikov, M. Balog, M. P. Kumar, E. Dupont, F. J. R. Ruiz, J. S. Ellenberg, P. Wang, O. Fawzi, P. Kohli, and A. Fawzi, 2024, Mathematical discoveries from program search with large language models, Nature 625, 468–475.
  34. [36]Schmidhuber, J., 2007, Gödel machines: Fully self-referential optimal universal self-improvers, in B. Goertzel, and C. Pennachin, eds., Artificial General Intelligence, 199–226 (Springer, Berlin).DOI
  35. [37]Seger, J., and H. J. Brockmann, 1987, What is bet-hedging?, Oxford Surveys in Evolutionary Biology 4, 182–211.
  36. [38]Snell, C., J. Lee, K. Xu, and A. Kumar, 2024, Scaling LLM test-time compute optimally can be more effective than scaling model parameters, arXiv:2408.03314.arxiv.org/abs/2408.03314
  37. [39]Spector, L., 2001, Autoconstructive evolution: Push, PushGP, and Pushpop, in Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001), 137–146 (Morgan Kaufmann, San Francisco).
  38. [40]Srikanth, D., B. Zhao, D. Xu, Y. Wu, and Z. Jiang, 2026, Recursive self-improvement of AI research agents, arXiv:2609.26457.arxiv.org/abs/2609.26457
  39. [41]Whitley, D., V. S. Gordon, and K. Mathias, 1994, Lamarckian evolution, the Baldwin effect and function optimization, in Parallel Problem Solving from Nature — PPSN III, volume 866 of Lecture Notes in Computer Science, 5–15 (Springer).DOI
  40. [42]Zhang, J., S. Hu, C. Lu, R. Lange, and J. Clune, 2025, Darwin Gödel Machine: Open-ended evolution of self-improving agents, arXiv:2505.22954.arxiv.org/abs/2505.22954
  41. [43]Zong, Q., J. Liu, J. Shen, Z. Tang, L. Wu, Y. Liu, R. Wang, Z. Wang, W. Wang, C. Qian, X. Chen, and Y. Song, 2026, Co-evolution in agentic systems: Toward self-directed evolution beyond human design, arXiv:2608.10299.arxiv.org/abs/2608.10299

Paper details

vlnc:W_s2ar73us
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY-SA 4.0
Usage

views · 19

downloads · 2

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

ChatGPT (6.1 Sol Ultra), Claude Fable 5.1 (bare), refine.ink

View statement
Looking for related papers…
Contents