# Flow State: Humans Enabling AI Systems to Program Themselves

arXiv: https://arxiv.org/abs/2504.03771 · HTML: https://mlsanity.com/paper/2504.03771/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 51 references that are in the map.
- First posted on arXiv 3 Apr 2025.

## Builds on (most cited first, 15 of 15)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Sparks of Artificial General Intelligence: Early experiments with GPT-4](https://mlsanity.com/paper/2303.12712/) (2023)
- [Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection](https://mlsanity.com/paper/2310.11511/) (2023)
- [Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks](https://mlsanity.com/paper/2211.12588/) (2022)
- [Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions](https://mlsanity.com/paper/2212.10509/) (2022)
- [CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis](https://mlsanity.com/paper/2203.13474/) (2022)
- [DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines](https://mlsanity.com/paper/2310.03714/) (2023)
- [Textbooks Are All You Need](https://mlsanity.com/paper/2306.11644/) (2023)
- [Decomposed Prompting: A Modular Approach for Solving Complex Tasks](https://mlsanity.com/paper/2210.02406/) (2022)
- [CoAuthor: Designing a Human-AI Collaborative Writing Dataset for Exploring Language Model Capabilities](https://mlsanity.com/paper/2201.06796/) (2022)
- [ExpertPrompting: Instructing Large Language Models to be Distinguished Experts](https://mlsanity.com/paper/2305.14688/) (2023)
- [Understanding the Capabilities of Large Language Models for Automated Planning](https://mlsanity.com/paper/2305.16151/) (2023)
- [Understanding Nonlinear Collaboration between Human and AI Agents: A Co-design Framework for Creative Design](https://mlsanity.com/paper/2401.07312/) (2024)
- [Alto: Orchestrating Distributed Compound AI Systems with Nested Ancestry](https://mlsanity.com/paper/2403.04311/) (2024)

## Built on by (most cited first, 3 of 3)

- [A Novel Compound AI Model for 6G Networks in 3D Continuum](https://mlsanity.com/paper/2505.15821/) (2025)
- [ReusStdFlow: A Standardized Reusability Framework for Dynamic Workflow Construction in Agentic AI](https://mlsanity.com/paper/2602.14922/) (2026)
- [LLM-Based Code Documentation Generation and Multi-Judge Evaluation](https://mlsanity.com/paper/2606.09852/) (2026)

## Abstract (arXiv)

Compound AI systems, orchestrating multiple AI components and external APIs, are increasingly vital but face challenges in managing complexity, handling ambiguity, and enabling effective development workflows. Existing frameworks often introduce significant overhead, implicit complexity, or restrictive abstractions, hindering maintainability and iterative refinement, especially in Human-AI collaborative settings. We argue that overcoming these hurdles requires a foundational architecture prioritizing structural clarity and explicit control. To this end, we introduce Pocketflow, a platform centered on Human-AI co-design, enabled by Pocketflow. Pocketflow is a Python framework built upon a deliberately minimal yet synergistic set of core abstractions: modular Nodes with a strict lifecycle, declarative Flow orchestration, native hierarchical nesting (Flow-as-Node), and explicit action-based conditional logic. This unique combination provides a robust, vendor-agnostic foundation with very little code that demonstrably reduces overhead while offering the expressiveness needed for complex patterns like agentic workflows and RAG. Complemented by Pocket AI, an assistant leveraging this structure for system design, Pocketflow provides an effective environment for iteratively prototyping, refining, and deploying the adaptable, scalable AI systems demanded by modern enterprises.