AI Tools in My Research Workflow

AI has quietly become the backbone of how I research. This post is a transparent look at the tools I use, what they are good at, and the guardrails I keep.

Where AI helps most

• Synthesis. Reading 20 sources and producing a structured brief — minutes instead of hours.
• Drafting. Getting past the blank page, then rewriting heavily.
• Analysis. Spotting patterns in data and code that I would have missed.
• Exploration. Rapidly testing hypotheses in unfamiliar domains.

Where I draw the line

AI is a research assistant, not the researcher. I still:

• Read primary sources before trusting a summary
• Write the core argument myself
• Verify every number that goes into a post
• Disclose when an analysis was AI-assisted

A simple daily loop

Morning  → AI brief on saved sources, distilled into open questions
Midday  → Deep work: reading primary material, writing
Evening → Capture learnings as notes; feed them back to the system

The honest trade-off

The risk is not that AI makes us lazy — it is that we outsource judgment. The skill that matters now is asking better questions.

function askBetterQuestions(goal) {
  return ["Why", "What if", "What does the evidence say"]
    .map((q) => `${q} ${goal}?`);
}