Freestyle Cognition is a hands-on system for using AI as a creative partner. This site will guide you—step by step—through everything you need to get started.
Begin the JourneyWatch the slides to understand what Freestyle Cognition is and how it works.
▶️ ViewSee how prompts shape the conversation and guide the AI’s response.
💬 TryExplore frameworks like GROW and ALIGN that guide better interaction.
🔧 LearnUse the builder to chain techniques together for powerful results.
🧰 BuildSee how the system works in practice—real prompts, real outcomes.
📘 ViewVisual slideshows of each template, technique, and system loop.
🖼 SlidesPosted by Ernan Hughes Author of Freestyle Cognition
“You’ve probably already felt it. You gave the AI a messy thought… and it came back with something better. That little magic moment? That’s Freestyle Cognition.”
Freestyle Cognition is not just another AI prompt strategy. It’s a process — a new way to think with AI, not just ask it for help.
Read More“What if AI systems could learn how to improve themselves not just at the level of weights or prompts, but at the level of strategy itself? In this post, we show how to build such a system, powered by symbolic rules and reflection.
The paper Symbolic Agents: Symbolic Learning Enables Self-Evolving Agents introduces a framework where symbolic rules guide, evaluate, and evolve agent behavior.
Read MoreChain-of-thought is powerful, but which chain? Short explanations work for easy tasks, long reflections help on hard ones, and code sometimes beats them both. What if your model could adaptively pick the best strategy, per task, and improve as it learns?
The Adaptive Reasoning Model
(ARM) is a framework for teaching language models how to choose the right reasoning format direct answers, chain-of-thoughts, or code depending on the task. It works by evaluating responses, scoring them based on rarity, conciseness, and difficulty alignment, and then updating model behavior over time.
AI research tools today are often narrow: one generates summaries, another ranks models, a third suggests ideas. But real scientific discovery isn’t a single step—it’s a pipeline. It’s iterative, structured, and full of feedback loops.
In this post, I show how to build a modular AI system that mirrors this full research lifecycle. From initial idea generation to method planning, each phase is handled by a specialized agent working in concert.
Read MoreModern AI systems require more than just raw processing power they need contextual awareness, strategic foresight, and adaptive learning capabilities. In this post, we walk through how we implemented a self-aware pipeline system inspired by the Devil’s Advocate paper.
Unlike brittle, static workflows, this architecture empowers agents to reflect on their own steps, predict failure modes, and adapt their strategies in real time.
ReReST: Devil's Advocate: Anticipatory Reflection for LLM Agents introduces a self-training framework for LLM agents. The core idea is to have a “reflector” agent anticipate failures and revise the original plan before executing a powerful method for reducing hallucinations and improving sample quality. Our implementation draws heavily on these ideas to enable dynamic planning and feedback loops within the pipeline.
Read MoreThe General Reasoner paper shows how we can train LLMs to reason across domains using diverse data and a generative verifier. In this post, I walk through our open-source implementation showing how we built a modular reasoning agent capable of generating multiple hypotheses, evaluating them with an LLM-based judge, and selecting the best answer.
We built a GeneralReasonerAgent
that:
cot
, debate
, verify_then_answer
, etc.)All of this was integrated with our existing co_ai framework, which includes:
Read MoreErnan Hughes writes at the frontier of human-machine collaboration, building frameworks that help people think, create, and live more powerfully with AI.