The host rejects the expectation that every generated recommendation should be correct or immediately actionable. As with a real board, the user has more context than the advisers and must decide which ideas deserve attention, deeper investigation, or rejection.

Expect uneven relevance across the generated responses
Retain responsibility for the final decision
Use advice to stimulate thinking rather than outsource judgment
Explore the few ideas that reveal genuinely new angles

“You, as the individual, the CEO, whomever in this situation who has a board of directors, are always going to have more context.”
05:30

“The purpose of having any board of directors is to get your brain thinking in a different way, get yourself seeing angles to the problem…”
06:00

#judgment#ai-limitations#decision-making#critical-thinking

Explainer· 2

Explainer00:30

Build a Personal Board of Directors With AI

The episode proposes using an LLM to simulate feedback from accomplished entrepreneurs, leaders, and founders. This gives people without access to elite advisers a fast, free way to examine difficult problems from multiple perspectives.

Choose advisers with distinct strengths and experiences
Ask the LLM to respond to a specific problem as each adviser
Use simulated perspectives to uncover angles you may have missed
Treat the board as a thinking aid rather than literal access to the people

“Today I'm gonna show you how to build that with artificial intelligence for free.”
00:00

“So you can pick anybody, whether you know them or not.”
01:30

#ai#decision-making#advisers#chatgpt

Explainer10:00

Model the Questions Your Stakeholders Might Ask

The technique is not limited to famous founders. With enough source material, users can model a CEO, manager, coworker, or parent to anticipate likely reactions and questions before a conversation or decision.

Create personas for people you actually work with
Ground each persona in available recordings or transcripts
Use role-play to anticipate objections and questions
Treat the output as preparation rather than a definitive prediction

“And, you know, you could do this for your CEO, you could do this for your manager, your co workers, your parents, right?”
10:00

“As long as you have enough information as this person, you can have AI basically role-play with you what that person might think or do…”
10:00

#stakeholders#role-play#communication#preparation

Story· 1

Story04:00

What Steve Jobs, Dyson, and Zell Say About YouTube Growth

The host tests the board on growing Marketing Against The Grain faster on YouTube. The generated responses emphasize simplifying the viewer journey, relentless iteration, and finding an overlooked audience niche rather than relying on one generic recommendation.

Steve Jobs' perspective emphasizes intuitive visual and narrative design
James Dyson's perspective emphasizes testing and iteration
Sam Zell's perspective emphasizes overlooked niches and hidden value
Different advisers illuminate different parts of the same growth problem

“Ask yourself does each element simplify the viewer's journey and spark curiosity?”
05:00

“Look for an overlooked niche in your audience.”
06:30

#youtube#growth#marketing#steve-jobs#experimentation

Tool· 2

Tool01:00

Ask AI to Help Write the Prompt You Need

Rather than drafting the entire board prompt manually, the host asks the model to create it. He then improves the result by specifying the desired advisers, the feedback format, and the problem the board should address.

Start by asking the model to help write the prompt
Name the advisers whose perspectives you want represented
Specify that each adviser should respond to a hard problem
Leave a clear place to insert the idea, writing, or challenge

“Just a reminder: one of the best things to do is ask AI to help you write a prompt.”
01:00

“And then I said, I want a prompt that an LLM can use to give me feedback as if it were each of these people,…”
02:00

#prompting#llm#chatgpt#workflow

Tool02:30

Ground Each AI Adviser With Deep Research

The initial prompt produces short descriptions of each chosen person, but the host considers that insufficient. He uses deep research to add cited summaries focused on the experiences and qualities that make each adviser distinctive, giving the model richer context for its responses.

Do not settle for shallow personality summaries
Research the experiences and qualities that distinguish each adviser
Add the resulting evidence directly to the reusable prompt
Use citations and detailed context when the model supports research

“I was like, that's good, but it's not good enough.”
02:30

“I want summaries that focus on the very unique experience and qualities that make each person great.”
02:30

#deep-research#personas#prompt-engineering#context

Takeaway· 3

Takeaway01:00

Choose AI Advisers With Complementary Strengths

The usefulness of the simulated board depends partly on selecting people who bring different capabilities to the same problem. The host chooses figures associated with product design, invention, investing, branding, media, mental models, and luxury to create a broad range of viewpoints.

Select people known for different areas of expertise
Include perspectives relevant to both the problem and its wider context
Do not limit the board to people you personally know
Change the lineup when a different problem requires different experience

“So these are the people I picked because I think they're amazing entrepreneurs, amazing leaders, amazing founders, and they all have some different skills.”
01:30

“And uh you can pick anyone you want.”
01:30

#advisory-board#perspectives#entrepreneurship#problem-solving

Takeaway08:30

Better Context Produces More Useful Board Feedback

The broad demonstrations are designed to show the range of possible responses, but the host says a real consultation should contain more detail. Supplying a memo, presentation, transcript, recording, or other relevant material can make the simulated advice more specific and grounded.

Use broad questions for demonstrations, not important decisions
Provide a memo or deck for more detailed analysis
Add transcripts, recordings, or videos when modeling known people
Match the available context to the complexity of the problem

“I think in actuality, you would give it like a memo, a deck, more context that would be even better.”
08:30

“And if you just have some video, some transcripts, call recordings, things like that. You can add them.”
10:00

#context#prompting#ai-advice#documents

Takeaway11:00

Spend 10 Minutes on Your Biggest Problem Each Day

The closing recommendation is to make the board a recurring problem-solving habit rather than a one-time novelty. Users can save the prompt or turn it into a custom GPT, then apply it whenever they feel stuck or to the most important problem of each day.

Save the prompt for repeated use
Optionally package it as a custom GPT
Use it whenever progress on an important problem stalls
Dedicate ten minutes each day to reviewing diverse feedback

“And I think what matters more is building a habit of doing this anytime you have a big problem.”
11:30

“I'm going to use this approach and dedicate the 10 minutes of doing this prompt and reading the feedback and thinking about it to help…”
11:30

#habits#problem-solving#productivity#custom-gpt

Frameworks in this episode

AI Personal Board of Directors

Personal Board Advice Filter

Ten-Minute Daily Board Habit

Insights & moments

⚡Myth Buster· 1

A Simulated Board Is Not Supposed to Be Perfect

✶Explainer· 2