---
name: ai-research
description: Systematic AI research process. Use when the user is AI researcher and the project is involved in AI research project.
---

# AI Research

A modular skill distilled from Neel Nanda’s research process sequence (Posts 1–3).  
The goal is to help you **make progress under uncertainty** by identifying the right stage, the right north star, and the right next actions.

This skill is designed for empirical research with short feedback loops, but generalizes broadly.

---

## North star

Research progress looks different at different stages:

- **Explore:** maximize information gained per unit time
- **Understand:** gather evidence that convinces you a hypothesis is true or false
- **Distill:** compress findings into concise, defensible, communicable truth

A common failure mode is using the wrong standard for the current stage.

---

## How to use

### Common inputs

- A short description of the project
- Current artifacts (results, plots, notes, draft text)
- What feels confusing or stuck
- Any constraints (deadline, compute, scope)

(Use: [Intake template](templates/intake-template.md))

### Common outputs

- Current stage + north star
- Top uncertainties
- 1–3 prioritized next actions
- A fast-fail test (if the direction is wrong)
- Baselines / alternative explanations to check
- If applicable: hypothesis or claim distillation

(Default output format: [Triage output template](templates/triage-output-template.md))

---

## Workflows

### A) “I’m stuck / what should I do next?”

1. [Stage classifier](modules/01-stage-classifier.md)
2. Run the relevant stage module:
   - Explore → [Exploration: information gain](modules/03-exploration-info-gain.md)
   - Understand → [Understanding: hypotheses & evidence](modules/04-understanding-hypotheses.md)
   - Distill → [Distillation: communicable truth](modules/05-distillation-communicable-truth.md)
3. [Prioritisation](modules/07-prioritisation.md)
4. Produce triage output

### B) “I don’t trust my results”

1. [Truth-seeking](modules/06-truth-seeking.md)
2. [Understanding: hypotheses & evidence](modules/04-understanding-hypotheses.md)
3. [Red-team checklist](templates/red-team-template.md)

### C) “I need to move faster”

1. [Moving fast](modules/08-moving-fast.md)
2. [Tight feedback loops](modules/09-tight-feedback-loops.md)
3. [Fail fast](modules/10-fail-fast.md)

### D) “I need to write this up / distill”

1. [Distillation: communicable truth](modules/05-distillation-communicable-truth.md)
2. [Distillation claims template](templates/distillation-claims-template.md)
3. [Hypothesis ↔ evidence map](templates/hypothesis-evidence-map-template.md)

---

## Modules index

- [00 Overview](modules/00-overview.md)
- [01 Stage classifier](modules/01-stage-classifier.md)
- [02 Ideation & problem choice](modules/02-ideation-problem-choice.md)
- [03 Exploration: information gain](modules/03-exploration-info-gain.md)
- [04 Understanding: hypotheses & evidence](modules/04-understanding-hypotheses.md)
- [05 Distillation: communicable truth](modules/05-distillation-communicable-truth.md)
- [06 Truth-seeking](modules/06-truth-seeking.md)
- [07 Prioritisation](modules/07-prioritisation.md)
- [08 Moving fast](modules/08-moving-fast.md)
- [09 Tight feedback loops](modules/09-tight-feedback-loops.md)
- [10 Fail fast](modules/10-fail-fast.md)
- [11 Action under uncertainty](modules/11-action-under-uncertainty.md)
- [12 Research taste: overview](modules/12-research-taste-overview.md)
- [13 Research taste: ingredients](modules/13-taste-ingredients.md)
- [14 Cultivating research taste](modules/14-cultivating-taste.md)
- [15 Common failure modes](modules/15-common-failure-modes.md)

---

## Templates

- [Intake template](templates/intake-template.md)
- [Triage output template](templates/triage-output-template.md)
- [Experiment plan template](templates/experiment-plan-template.md)
- [Red-team template](templates/red-team-template.md)
- [Distillation claims template](templates/distillation-claims-template.md)
- [Weekly review template](templates/weekly-review-template.md)

---

## Acknowledgements

Compiled from Neel Nanda’s research process sequence (Posts 1–3, Apr–May 2025).