B
Offline
Step 1 of 8

Why Agent-Assisted?

1of 8~3 min
  • Why having your AI agent internalize methodology is high-leverage
  • The human-AI collaboration model for research
  • What distinguishes this from prompting an AI directly
What you'll do
  • Understand the agent-assisted approach
  • Learn the division of labor between human and AI
  • Set expectations for the tutorial

The Core Insight

AI agents can internalize methodology, not just follow instructions.

When you have Claude Code or Codex read the Brenner documentation, something different happens than when you paste a prompt into a chat interface. The agent builds a working understanding of the method, not just the task.

What Makes This Different

Direct Prompting

  • You write detailed prompts each time
  • AI follows instructions literally
  • Context resets with each conversation
  • You do the methodological thinking

Agent-Assisted

  • Agent internalizes the methodology once
  • AI thinks with the method, not just about it
  • Context persists across the session
  • Agent provides methodological discipline

The Human-AI Collaboration

This isn't about replacing human judgment. It's about division of labor:

You Provide

  • +Domain expertise and context
  • +The research question that matters to you
  • +Judgment on what's feasible
  • +Final decisions and direction

Agent Provides

  • +Methodological discipline (from Brenner)
  • +Systematic hypothesis generation
  • +Third alternatives you might miss
  • +Structured artifact creation

Why This Is the Highest-Leverage Path

Reusable across questions

Once the agent has internalized the method, you can apply it to any research question without re-explaining the methodology.

Faster iteration

The agent remembers context within a session, so you can refine hypotheses and tests without starting over.

Consistent methodology

The agent applies Brenner's operators systematically, catching blind spots you might miss under time pressure.

Pro Tip
This tutorial requires Claude Code (with Claude Max) or Codex (with GPT Pro). If you don't have either, try the Quick Start path first — it works with any AI chat interface including free tiers.

What You'll Produce

By the end of this tutorial, your agent will have generated a complete research artifact:

  • Hypothesis slate with third alternatives
  • Discriminative tests ranked by potency
  • Explicit assumption ledger
  • Adversarial critique of your framing

Next up: In Step 2, you'll verify that you have Claude Code or Codex installed and ready to go.