A guide to the inventor's mindset

The big idea

The writers who survive the AI shift will be the ones who treat their practice like an inventor’s workshop: try something new, keep what works, drop what doesn’t, and don’t treat any failed experiment as a verdict on your ability. The goal is a working rhythm of tinkering, not a one-shot plan.

The framework

1. Start with a concept, not a finished idea

An inventor doesn’t start with a completed blueprint. They start with a rough concept they want to explore. For a writer, that might be ​“can AI help me outline faster?” — not ​“build a complete AI writing workflow.”

2. Tinker in small bets

Try one change at a time. Use ChatGPT for headlines this week. Try Perplexity for source-hunting on the next piece. Measure each experiment in hours of your time, not in stakes.

3. Keep what works, drop what doesn’t

If a change saves time without compromising quality, keep it. If it doesn’t, move on. The key move: don’t frame a failed experiment as a failure. It’s data about what doesn’t fit your practice.

4. Expect to land somewhere you didn’t plan

Gregory describes the pattern as ​“they get a lightbulb moment then they make six more changes till it’s the way that they want it.” The finished system usually looks different from what you imagined on day one.

When to use this

• You’re staring at a new tool and can’t decide whether to try it on a live project.
• You’ve been avoiding AI because the policy questions feel too big to resolve.
• You want to build a personal workflow but don’t know where to start.
• You’re in the ​“messy middle” of AI adoption and want a stance that doesn’t require certainty.

Common mistakes

• Treating a failed experiment as a verdict on your skills → treat it as information about what doesn’t fit your practice.
• Waiting to adopt AI until you have a complete ethics framework → form your ethics through actual usage, one project at a time.
• Using AI to write → use it to think, outline, and research, then write it yourself.
• Copying AI output into your draft → keep AI outputs in a separate tab; let them spark your own words.

Quick-start

Pick one task this week — brainstorming, outlining, or source-hunting — and try one AI tool on it. Compare the output to how you’d normally do the work. If it saves time without lowering quality, keep it. If not, move on to the next experiment.