Focus Level: Airplane Thinking

In a world where there is an increasing tendency for people to multitask the entire day away, the best prompts are going to come from quiet, rigorous thinking.

Does anyone else get their best thinking done on airplanes?

Regardless of where your best/deep thinking is done, there is going to be an increased need for that type of focused, deep work to maximize value from AI tools.

On a recent 17-hour flight (direct) to Singapore, I avoided paying for internet and, instead, took the opportunity to do some deep, focused work - much of which was building out draft Prompts and GPTs for a couple projects.

While we’ve been “trained” for years (decades?) to treat optimal internet search and research as “the quickest path to the answer”, winning with AI tools requires a much different mindset. It will require more than the rapid fire text into a search bar.

While quick pings to AI tools can be helpful for surface level tasks (e.g., proofread this document, create a grocery list for a particular meal, give me an analogy of this concept…), the long-term value is going to be derived from prompts and GPTs that have under good deep and critical thinking during their development.

To aid the deep work, I recommend using an outline for prompting that forces a holistic approach.

The FEAT framework I developed helps provide scaffolding for prompt development, and the level of thought that goes into really thinking through each element will be directly correlated with the quality of the output/interactions with the model.

A few other tips:

1️⃣ Think beyond the first interaction with a prompt.
What is important about the latter engagements? Should the AI let you go down rabbit holes? Or gently remind you to stay focused on your initially stated goals?

2️⃣ Define what the interaction should be like.
Should the AI provide thought provoking ideas? Or give specific guidance on how to accomplish something?

3️⃣ Specifics phases or stages for the interactions.
This is something I’ve been making standard in my prompts, but what should the interaction flow look like? What are the stage gates, and what are the requirements for proceeding with the interactions?

4️⃣ Define the use cases.
What are all of the ways you or others might be using the prompt? How can the prompt be structured to limit focus (increasing output relevancy) while still being flexible (allowing for multiple applications)?

If you’re leading the AI charge for your company, encourage people working on prompts or GPTs to block off time and turn off notifications to help prevent distractions that can limit the utility and impact of prompts.

In a world where there is an increasing tendency for people to multitask the entire day away, the best prompts are going to come from quiet, rigorous thinking*.

Would love to hear if others have experienced the same ROI from time invested into the deep thinking for AI prompts?

* Caveat: it’s still important to take a break to watch an in-flight show in between prompts :)