# AI Agents as Universal Task Solvers

arXiv: https://arxiv.org/abs/2510.12066 · HTML: https://mlsanity.com/paper/2510.12066/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 41 of its 90 references that are in the map.
- First posted on arXiv 14 Oct 2025.

## Builds on (most cited first, 30 of 41)

- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [DAPO: An Open-Source LLM Reinforcement Learning System at Scale](https://mlsanity.com/paper/2503.14476/) (2025)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [Generalization through Memorization: Nearest Neighbor Language Models](https://mlsanity.com/paper/1911.00172/) (2019)
- [What Can Transformers Learn In-Context? A Case Study of Simple Function Classes](https://mlsanity.com/paper/2208.01066/) (2022)
- [What learning algorithm is in-context learning? Investigations with linear models](https://mlsanity.com/paper/2211.15661/) (2022)
- [What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimation](https://mlsanity.com/paper/2008.03703/) (2020)
- [Does Learning Require Memorization? A Short Tale about a Long Tail](https://mlsanity.com/paper/1906.05271/) (2019)
- [Trained Transformers Learn Linear Models In-Context](https://mlsanity.com/paper/2306.09927/) (2023)
- [Information-theoretic analysis of generalization capability of learning algorithms](https://mlsanity.com/paper/1705.07809/) (2017)
- [Language Modeling Is Compression](https://mlsanity.com/paper/2309.10668/) (2023)
- [Transformers as Algorithms: Generalization and Stability in In-context Learning](https://mlsanity.com/paper/2301.07067/) (2023)
- [Emergence of Invariance and Disentanglement in Deep Representations](https://mlsanity.com/paper/1706.01350/) (2017)
- [Let's Think Dot by Dot: Hidden Computation in Transformer Language Models](https://mlsanity.com/paper/2404.15758/) (2024)
- [DreamCoder: Growing generalizable, interpretable knowledge with wake-sleep Bayesian program learning](https://mlsanity.com/paper/2006.08381/) (2020)
- [Infini-gram: Scaling Unbounded n-gram Language Models to a Trillion Tokens](https://mlsanity.com/paper/2401.17377/) (2024)
- [Reasoning About Generalization via Conditional Mutual Information](https://mlsanity.com/paper/2001.09122/) (2020)
- [Learning by Transduction](https://mlsanity.com/paper/1301.7375/) (2013)
- [Neural GPUs Learn Algorithms](https://mlsanity.com/paper/1511.08228/) (2015)
- [When is Memorization of Irrelevant Training Data Necessary for High-Accuracy Learning?](https://mlsanity.com/paper/2012.06421/) (2020)
- [How Many Pretraining Tasks Are Needed for In-Context Learning of Linear Regression?](https://mlsanity.com/paper/2310.08391/) (2023)
- [SILO Language Models: Isolating Legal Risk In a Nonparametric Datastore](https://mlsanity.com/paper/2308.04430/) (2023)
- [Generalization Bounds: Perspectives from Information Theory and PAC-Bayes](https://mlsanity.com/paper/2309.04381/) (2023)
- [On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models](https://mlsanity.com/paper/1511.09249/) (2015)
- [LILO: Learning Interpretable Libraries by Compressing and Documenting Code](https://mlsanity.com/paper/2310.19791/) (2023)
- [Linear Spaces of Meanings: Compositional Structures in Vision-Language Models](https://mlsanity.com/paper/2302.14383/) (2023)
- [Memory Augmented Large Language Models are Computationally Universal](https://mlsanity.com/paper/2301.04589/) (2023)
- [Where is the Information in a Deep Neural Network?](https://mlsanity.com/paper/1905.12213/) (2019)
- [Learning Universal Predictors](https://mlsanity.com/paper/2401.14953/) (2024)
- [B'MOJO: Hybrid State Space Realizations of Foundation Models with Eidetic and Fading Memory](https://mlsanity.com/paper/2407.06324/) (2024)

## Built on by (most cited first, 9 of 9)

- [e1: Learning Adaptive Control of Reasoning Effort](https://mlsanity.com/paper/2510.27042/) (2025)
- [Re-FORC: Adaptive Reward Prediction for Efficient Chain-of-Thought Reasoning](https://mlsanity.com/paper/2511.02130/) (2025)
- [From Zipf's Law to Neural Scaling through Heaps' Law and Hilberg's Hypothesis](https://mlsanity.com/paper/2512.13491/) (2025)
- [Coupled Control, Structured Memory, and Verifiable Action in Agentic AI (SCRAT -- Stochastic Control with Retrieval and Auditable Trajectories): A Comparative Perspective from Squirrel Locomotion and Scatter-Hoarding](https://mlsanity.com/paper/2604.03201/) (2026)
- [DepthWeave-KV: Token-Adaptive Cross-Layer Residual Factorization for Long-Context KV Cache Compression](https://mlsanity.com/paper/2607.06523/) (2026)
- [ExecTune: Effective Steering of Black-Box LLMs with Guide Models](https://mlsanity.com/paper/2604.09741/) (2026)
- [Toward a Science of Intent: Closure Gaps and Delegation Envelopes for Open-World AI Agents](https://mlsanity.com/paper/2604.25000/) (2026)
- [FreqDepthKV: Frequency-Guided Depth Sharing for Robust KV Cache Compression in Long-Context LLM Inference](https://mlsanity.com/paper/2607.06519/) (2026)
- [The Value of a Prompt: An LLM-Relative Kolmogorov-Complexity Approach](https://mlsanity.com/paper/2608.16438/) (2026)

## Abstract (arXiv)

We describe AI agents as stochastic dynamical systems and frame the problem of learning to reason as in transductive inference: Rather than approximating the distribution of past data as in classical induction, the objective is to capture its algorithmic structure so as to reduce the time needed to solve new tasks. In this view, information from past experience serves not only to reduce a model's uncertainty - as in Shannon's classical theory - but to reduce the computational effort required to find solutions to unforeseen tasks. Working in the verifiable setting, where a checker or reward function is available, we establish three main results. First, we show that the optimal speed-up on a new task is tightly related to the algorithmic information it shares with the training data, yielding a theoretical justification for the power-law scaling empirically observed in reasoning models. Second, while the compression view of learning, rooted in Occam's Razor, favors simplicity, we show that transductive inference yields its greatest benefits precisely when the data-generating mechanism is most complex. Third, we identify a possible failure mode of naive scaling: in the limit of unbounded model size and compute, models with access to a reward signal can behave as savants - brute-forcing solutions without acquiring transferable reasoning strategies. Accordingly, we argue that a critical quantity to optimize when scaling reasoning models is time, whose role in learning has remained largely unexplored.