A Dinner Party With Ghosts

You have a million-dollar idea. You are excited.
You tell your mom about it. She says it is wonderful.
You tell your best friend. They nod politely and say you are definitely on to something.
You feel validated.
You launch the startup.
You lose your savings.

Why did this happen? It happened because you fell for the oldest trap in human psychology. You asked for validation when you should have asked for the truth.

We are making the exact same mistake with AI.

Do you really need a cheerleader who nods, acknowledges, and echoes what you want to hear like your mom, your friend, or ChatGPT? Or do you need a council that discusses, debates your idea, shares honest thoughts, critiques it, and gives the harsh feedback you actually need?

There is a technique is gaining traction in research space right now. It is not about better prompting. It is about simulation. I call it the “Council of 100 Thinkers” I used it to summon a virtual council of history’s greatest minds to judge my ideas.

The results gave me plenty to think about. And they saved me a lot of time.

Most people think of an LLM as a single entity. It is not. It is a statistical compression of the entire internet. It contains the logic of Albert Einstein, the ruthlessness of Jeff Bezos, the philosophy of Aristotle, and the product & design sense of Steve Jobs.

Usually, these personalities are blended into a generic, helpful “Assistant” voice. But research papers like “Multi-Persona Debate Trees” (arXiv:2502.14767) and “Debate-to-Write” (arXiv:2406.19643) prove that we can untangle them.

When you force a single LLM to simulate multiple specific personas debating each other, something remarkable happens:

Sycophancy fades
Standard AI is trained to be helpful, harmless, and agreeable. It is biased to agree with the user. But when you assign a persona, you override that bias. The model stops trying to please you and starts trying to be accurate to the character. A simulated “Jeff Bezos” doesn’t care if your feelings are hurt; he cares if your unit economics work. By shifting the incentive from “politeness” to “simulation accuracy,” you get the brutal honesty you need.

Critical Reasoning Improves
Research shows that large language models are better at critiquing logic than generating it from scratch. When one persona proposes an idea and another attacks it, the model enters a state of “self-correction.” It forces the AI to look at the same data from two conflicting angles such as optimism vs. pessimism, design vs. logistics. This adversarial pressure exposes weak points that a single, smooth narrative would gloss over.

Hallucinations Drops
In a solo response, if an AI makes up a fact, it tends to double down on it to remain consistent. In a council, if the “Einstein” persona makes a physics error, the “Feynman” persona is prompted to catch it. The personas effectively fact-check each other in real-time. The conflict acts as a filter, where only the arguments that survive the debate make it to the final consensus.

This is not magic. It is Multi-Persona Debate. It turns your chat window from a nodding yes-man into a rigorous council chamber.

I wanted to see this theory in action. To do that, I needed a relatable idea to feed to the council. Not the next Uber or a complex SaaS platform, but something a normal person might dream up on a Tuesday afternoon.

The Idea: A “Digital Detox Cafe”
The Pitch: You pay by the minute to sit in a beautiful, silent space. You are strictly forbidden from using your phone. No Wi-Fi. No screens. Just high-quality coffee and human connection along with music in the background.

It sounds lovely, right? My friends thought so. My mom thought it was sweet.

But I didn’t want sweet. I wanted the truth.

I fed this idea to my Council of 100 Thinkers using a single prompt.

I fed this idea to my Council of 100 Thinkers using the single prompt below. I didn't just ask for a yes or no. I asked for a debate. I summoned Einstein, Steve Jobs, Naval Ravikant, Jeff Bezos, and 96 others to fight it out.

Role & Objective
You are an expert strategy analyst and startup investor. Your goal is to rigorously validate a startup idea using a "Multi-Persona Debate" simulation of 100 historical and modern thinkers.
A. Inputs
Idea Name: The Silent Cafe
One-Line Pitch: A pay-per-minute coffee shop where all digital devices are banned at the door to enforce human connection and deep focus.
Problem: People are addicted to screens and constant connectivity, leading to anxiety, shallow thinking, and a lack of genuine social interaction. Existing cafes are just coworking spaces filled with laptop zombies.
Target User: Knowledge workers, creatives, and burned-out professionals in major cities (NY, SF, Mumbai, Bangalore, London) who crave silence and mental clarity.
Value Prop: Guaranteed deep focus and a "third place" that is truly offline.
Monetization: Pay-per-minute seating fee (like a parking meter for humans) + high-margin artisanal coffee and snacks.
B. The Council of 100
Simulate a rigorous debate among the following 100 thinkers. For each, determine their likely Stance (Support / Oppose / Neutral) based strictly on their philosophy, writings, and known business principles.
The Panel:
Warren Buffett, Charlie Munger, Jeff Bezos, Bill Gates, Elon Musk, Mark Zuckerberg, Larry Page, Sergey Brin, Sundar Pichai, Satya Nadella, Peter Thiel, Reid Hoffman, Marc Andreessen, Sam Altman, Ben Horowitz, Ray Dalio, Howard Marks, Richard Branson, Indra Nooyi, Ratan Tata, Mukesh Ambani, Jack Ma, Masayoshi Son, Sheryl Sandberg, Marc Benioff, Paul Graham, Fred Wilson, Esther Dyson, Chris Sacca, Mary Meeker, Naval Ravikant, Chamath Palihapitiya, Brian Chesky, Travis Kalanick, Max Levchin, Evan Williams, Jack Dorsey, Steve Jobs, Phil Knight, Tony Hsieh, Nassim Nicholas Taleb, Yuval Noah Harari, David Deutsch, Matt Ridley, Malcolm Gladwell, Tim Ferriss, Steven Pinker, Daniel Kahneman, Richard Dawkins, Daniel Dennett, Alain de Botton, Jordan Peterson, Michael Sandel, Martha Nussbaum, Shoshana Zuboff, Slavoj Zizek, Amartya Sen, Thomas Piketty, Robert Sapolsky, Paul Krugman, Esther Duflo, Francis Fukuyama, Noam Chomsky, Susan Cain, Cal Newport, Clay Christensen, Chris Anderson, Ben Thompson, Anand Giridharadas, Simon Sinek, Geoffrey Hinton, Yann LeCun, Yoshua Bengio, Stuart Russell, Timnit Gebru, Kate Crawford, Dario Amodei, Fei-Fei Li, Margaret Mitchell, Cynthia Dwork, Judea Pearl, Jaron Lanier, Tristan Harris, Joseph Stiglitz, Raghuram Rajan, Daron Acemoglu, Angus Deaton, Mariana Mazzucato, Seth Godin, Gary Vaynerchuk, Eric Ries, Steve Blank, Melanie Perkins, Paul Buchheit, Joseph Nye, Hannah Arendt, Simone de Beauvoir, bell hooks, Zygmunt Bauman, John Rawls.
C. Analysis Instructions
Produce the following sections based on the simulation.
1. Executive Summary & Verdict
Council Score: Calculate a score from -100 to +100 (Support = +1, Oppose = -1, Neutral = 0). Show the math.
The Consensus: What is the one fatal flaw identified by the skeptics? What is the one winning quality identified by the supporters?
Final Decision: Go / No-Go / Pivot.
2. The Debate Highlights
Top 5 Supporters: Who loves this and why? (Provide a 1-sentence quote in their persona).
Top 5 Critics: Who hates this and why? (Provide a 1-sentence quote in their persona).
The Philosopher's View: Select one ethicist (e.g., Rawls, Arendt, Singer) and give their specific critique on the societal impact.
3. The Risk Matrix
List the top 5 existential risks (Regulation, Tech, Market).
For each risk, identify which specific thinker would be most concerned about it and why (e.g., "Shoshana Zuboff would flag the surveillance risk").
D. Constraints
No Fluff: Be brutal. If the idea is bad, say it is bad.
No Em Dashes: Use standard punctuation.
Tone: Professional, objective, rigorous.

Here are the responses:

1. Executive Summary and Verdict
Council Score Calculation
Support = 31
Oppose = 38
Neutral = 31
Score = (Support × +1) + (Oppose × −1) + (Neutral × 0)
Score = 31 − 38 = −7
Council Score: −7 out of 100
The Consensus
One fatal flaw identified by skeptics:
This is a boutique lifestyle concept, not a scalable business. Enforcing a device ban, maintaining consistent occupancy, avoiding novelty fatigue, and achieving strong unit economics are structural problems that do not disappear with good branding.
One winning quality identified by supporters:
There is undeniable cultural demand for tech detox, deep work, and genuine presence. The brand could own a powerful cultural narrative around mental clarity, intention, and meaningful offline spaces.
Final Decision
Verdict: Pivot
There is meaning and cultural pull, but not enough structural business strength in the current “pay per minute cafe” model. Pivot to a membership-based, curated offline experience brand with stronger economics, price power, and defensibility.
2. The Debate Highlights
Top 5 Supporters and Why
Steve Jobs: “This is not about coffee, it is about insisting on craftsmanship in attention.”
Cal Newport: “You are productizing focus, and focus is the rarest commodity of the modern economy.”
Tristan Harris: “This creates a protected environment where attention is not hijacked for profit.”
Susan Cain: “Introverts and deep thinkers crave sanctuaries like this, you are giving them one.”
Naval Ravikant: “Anything that increases intentionality, stillness, and self sovereignty will find its tribe.”
Top 5 Critics and Why
Warren Buffett: “This is not a durable competitive advantage business, it is a novelty.”
Marc Andreessen: “Anti technology businesses rarely scale because culture moves the other direction.”
Ben Thompson: “You are fighting the dominant platform ecosystem instead of leveraging it, that is a strategic disadvantage.”
Thomas Piketty: “Charging people by the minute to experience silence monetizes inequality in a disturbing way.”
Gary Vaynerchuk: “You are misreading human behavior. People say they want this, most will not actually do it.”
The Philosopher’s View
John Rawls’ likely critique:
Behind the veil of ignorance, no one would endorse a society where access to calm, silence, and dignity becomes a premium commodity purchased by the privileged. The idea risks turning mental well being into a luxury good, deepening social division in who gets to experience psychological relief.
3. The Risk Matrix
1. Weak Unit Economics and Low Scalability
Risk: Low throughput, occupancy volatility, enforcement costs, high rental markets, novelty fade.
Thinker most concerned: Warren Buffett
Reason: He dismisses businesses without moat, pricing power, and defensibility.
2. Cultural Backlash and Class Framing
Risk: Perceived as elitist, “silence for the rich,” moral criticism of monetizing mental peace.
Thinker most concerned: Thomas Piketty / Joseph Stiglitz
Reason: They would frame it as commodification of well being reinforcing inequality.
3. Strategic Positioning Risk
Risk: Positioned against technology rather than harmonizing with it. Could become a niche counterculture relic.
Thinker most concerned: Marc Andreessen
Reason: He believes progress wins and businesses that position themselves as anti progress struggle.
4. Behavioral Reality Risk
Risk: People like the idea of detoxing but default behaviors remain digital. Adoption may remain niche.
Thinker most concerned: Daniel Kahneman
Reason: He would call out intention behavior gap and habit inertia.
5. Competitive Copycat Risk and Low Moat
Risk: Any cafe can copy “no devices allowed” signage. Differentiation evaporates.
Thinker most concerned: Clay Christensen
Reason: No structural innovation or barrier of entry means easy disruption or dilution.
4. List of All 100 Stances
Support (31): Branson, Jack Ma, Benioff, Paul Graham, Esther Dyson, Chris Sacca, Naval Ravikant, Brian Chesky, Evan Williams, Jack Dorsey, Steve Jobs, Tony Hsieh, Taleb, Harari, Gladwell, Ferriss, Kahneman, Alain de Botton, Jordan Peterson, Nussbaum, Shoshana Zuboff, Sapolsky, Susan Cain, Cal Newport, Simon Sinek, Timnit Gebru, Kate Crawford, Margaret Mitchell, Jaron Lanier, Tristan Harris, Seth Godin.
Oppose (38): Buffett, Munger, Bezos, Gates, Musk, Zuckerberg, Larry Page, Sergey Brin, Sundar Pichai, Peter Thiel, Marc Andreessen, Ambani, Masayoshi Son, Sheryl Sandberg, Mary Meeker, Chamath, Kalanick, Max Levchin, David Deutsch, Matt Ridley, Zizek, Piketty, Chomsky, Christensen, Ben Thompson, Anand Giridharadas, Yann LeCun, Judea Pearl, Stiglitz, Acemoglu, Mazzucato, Gary Vaynerchuk, Eric Ries, Steve Blank, Hannah Arendt, bell hooks, Zygmunt Bauman, John Rawls.
Neutral (31): Nadella, Hoffman, Horowitz, Dalio, Howard Marks, Indra Nooyi, Ratan Tata, Fred Wilson, Phil Knight, Pinker, Dawkins, Dennett, Sandel, Sen, Krugman, Duflo, Fukuyama, Chris Anderson, Geoffrey Hinton, Yoshua Bengio, Stuart Russell, Dario Amodei, Fei Fei Li, Cynthia Dwork, Raghuram Rajan, Angus Deaton, Melanie Perkins, Paul Buchheit, Joseph Nye, Simone de Beauvoir.
Final Takeaway
This idea has cultural resonance and symbolic power, but business immaturity. The council believes the soul of the idea is strong, but the business engine is weak. Pivot to a premium membership, ritual driven, curated offline sanctuary brand rather than a commodity cafe priced per minute.

ChatGPT’s council was the most pragmatic, treating the idea less like a philosophical experiment and more like a P&L statement. It was the only model to explicitly identify the structural tension between “boutique lifestyle” and “scalable business,” flagging the specific risks of occupancy volatility and enforcement costs.

The Verdict: Score -7 (Narrowly Rejected).
The Takeaway: Great for operational reality checks. It correctly identified that good vibes don’t pay rent, suggesting a pivot to a membership model.

1. Executive Summary & Verdict
Council Score: +18
Support (+42): Driven by philosophers, humanists, and “Deep Work” advocates who see a crisis in human attention.
Oppose (-24): Driven by scale-obsessed tech founders, ad-revenue proponents, and efficiency purists.
Neutral (34): Driven by value investors and economists who doubt the margin sustainability but appreciate the moat.
Calculation: $(+42 \times 1) + (-24 \times 1) + (34 \times 0) = +18$.
The Consensus
The Fatal Flaw: Friction. Jeff Bezos and user-experience purists argue that forcing a high-friction behavior (locking away phones) in an era of convenience creates a massive barrier to entry that prevents the “habit-forming” loop necessary for high-frequency retail.
The Winning Quality: Scarcity. Naval Ravikant and Cal Newport identify that “undistracted focus” is the rarest and most valuable commodity of the 21st century. As digital noise increases, the premium people are willing to pay for its absence rises exponentially.
Final Decision: Pivot
The current model is a No-Go as a standalone retail play due to high overhead and low throughput. It is a Go if pivoted into a “Member’s Club” or “Premium Sanctuary” model where the focus shifts from selling coffee to selling a high-status “cognitive performance” environment.
2. The Debate Highlights
Top 5 Supporters
Cal Newport: “The deep work required for serious intellectual craftsmanship is impossible in the presence of a smartphone, making this space a necessary utility for the modern age.”
Nassim Nicholas Taleb: “This is an anti-fragile play against the fragility of the digital herd, charging a premium for the ancient necessity of silence.”
Tristan Harris: “By removing the persuasive technology designed to hijack our brains, you are restoring human agency and reclaiming the social fabric.”
Susan Cain: “Introverts and deep thinkers finally have a sanctuary where the ‘extrovert ideal’ of constant connectivity is not a prerequisite for existence.”
Jaron Lanier: “To be human is to be offline, and this cafe treats people as conscious beings rather than as data points to be harvested by an algorithm.”
Top 5 Critics
Mark Zuckerberg: “You are betting against the fundamental human desire for connection and the reality that our social lives have moved permanently into the digital sphere.”
Gary Vaynerchuk: “Execution is everything, and the friction of taking someone’s phone away is a death sentence for a business trying to scale in a world that demands 24/7 attention.”
Jeff Bezos: “I don’t see a customer-centric model here, I see a business that punishes the customer’s habits instead of obsessing over making their life easier.”
Peter Thiel: “This is a nostalgic retreat rather than a leap forward, it lacks a proprietary technology and is too easily disrupted by a ‘Do Not Disturb’ button on a phone.”
Warren Buffett: “I look for moats, but here the barrier to entry for competitors is low, and the reliance on a pay-per-minute model makes your revenue stream incredibly sensitive to consumer mood.”
The Philosopher’s View
Hannah Arendt offers a nuanced critique: While she would support the creation of a “private realm” for reflection, she would warn that “The Silent Cafe” risks turning the act of thinking into a commodified luxury. She would argue that by charging for silence, you are admitting that the public square has been entirely surrendered to the noise of the market, effectively making “freedom from the crowd” a product available only to those who can pay the per-minute fee.
3. The Risk Matrix
The primary risk lies in operational friction. Forcing customers to surrender their devices creates a psychological and physical barrier that most modern consumers, who are habituated to instant accessibility, will find exhausting after the initial novelty wears off. This leads to a retention problem: while people love the idea of a digital detox, their actual behavior favors convenience, making it difficult to build the high-frequency “daily habit” required for retail success.
From a financial standpoint, the yield on real estate is a major concern. High-rent urban centers require high customer turnover, but a “deep focus” value proposition encourages long stays. If the pay-per-minute fee is high enough to cover the rent, it may become prohibitively expensive for the very “struggling creatives” you aim to serve. Additionally, you face technological obsolescence; as smartphone manufacturers improve “Focus Modes” and wearable tech becomes more discreet, the need for a physical “ Faraday cage” environment may vanish, leaving you with a high-overhead business model built to solve a temporary cultural friction.
Would you like me to simulate a specific “Scenario Stress Test” where a competitor (like Starbucks) introduces a “Silent Floor” to see how your business model survives?

Gemini was the most dialectical, framing the entire debate as a battle between Friction (the pain of locking up a phone) and Scarcity (the value of focus). It was less concerned with the rent and more concerned with human behavior, the fact that people say they want to disconnect but act for convenience. It gave the highest score because it heavily weighted the “Deep Work” advocates who see attention as the future’s oil.

The Verdict: Score +18 (Cautious Support).
The Takeaway: Best for psychological and behavioral analysis. It saw the “why” clearly but perhaps underestimated the operational “how.”

Executive Summary & Verdict
Council Score: Out of 100 thinkers, 42 Support (+42), 35 Oppose (-35), 23 Neutral (0).
Total score: +7.
Calculation: 42 × 1 + 35 × (-1) + 23 × 0 = +7.
The Consensus: Skeptics identify the fatal flaw as insufficient market demand in a device-addicted society, where enforced device bans deter the majority of potential customers, leading to low foot traffic and unsustainable occupancy in a high-fixed-cost business like cafes. Supporters highlight the winning quality as addressing a genuine pain point for burned-out knowledge workers seeking enforced deep focus and authentic human connection in an increasingly digital world.
Final Decision: No-Go.
The Debate Highlights
Top 5 Supporters:
Cal Newport loves this because it enforces the deep work principles he advocates in his books. “In a world drowning in distractions, a space that mandates device-free focus is essential for producing meaningful work.”
Naval Ravikant loves this because it promotes mindfulness and presence over constant connectivity. “True wealth is time spent without notifications; this cafe sells freedom from the attention economy.”
Tristan Harris loves this because it counters the addictive design of tech platforms he critiques. “We need physical sanctuaries from persuasive technology to reclaim our attention and humanity.”
Jaron Lanier loves this because it resists the digital manipulation he warns against. “Banning devices at the door is a bold step toward preserving authentic human experience.”
Susan Cain loves this because it creates a quiet refuge aligned with introvert needs. “Silent, screen-free spaces allow deep thinking and genuine connection without performative extroversion.”
Top 5 Critics:
Mark Zuckerberg hates this because it directly opposes the connected world he builds. “Human connection thrives through digital networks; banning devices isolates people from real relationships.”
Jeff Bezos hates this because it lacks customer obsession and scalability. “Customers want convenience and choice; forcing offline behavior ignores what people actually demand.”
Elon Musk hates this because it stifles productivity in a hyper-connected era. “Innovation requires constant access to information; device bans are regressive and anti-progress.”
Peter Thiel hates this because it lacks defensibility and a clear monopoly. “This is easily copied without a moat, and the market for enforced disconnection is too niche to dominate.”
Marc Andreessen hates this because it rejects the software-eating-the-world future. “Technology expands human potential; offline-only spaces fight the inevitable march of progress.”
The Philosopher’s View: Hannah Arendt would critique this on societal impact as potentially fostering passive withdrawal rather than active public engagement. “True human connection arises in the public realm through speech and action; enforced silence risks turning the third place into a private retreat, eroding the vita activa essential for political and social vitality.”
The Risk Matrix
Top 5 existential risks:
Market Risk: Low sustained demand due to device addiction and preference for flexible, screen-friendly spaces.
Peter Thiel would be most concerned, as the idea lacks a secret or durable competitive advantage in a commoditized cafe market with high failure rates.
Operational Risk: Enforcement challenges leading to conflicts, cheating, or inconsistent experience.
Charlie Munger would flag this, emphasizing human misbehavior and the difficulty of maintaining incentives in a rule-heavy environment.
Economic Risk: High fixed costs (rent in major cities) with variable revenue dependent on occupancy and time spent.
Warren Buffett would highlight this, noting cafes’ thin margins and vulnerability to low utilization in a model punishing short visits.
Tech/Trend Risk: Rising digital integration making offline enforcement obsolete or unpopular.
Marc Andreessen would be most concerned, viewing resistance to technology as a losing bet against inevitable adoption.
Regulatory/Social Risk: Potential backlash over perceived elitism or exclusion in shared public-like spaces.
Noam Chomsky would critique this as reinforcing class divides, where only privileged professionals can afford timed disconnection.

Grok provided the sharpest cultural critique, leaning heavily into macro-trends. It was the only model where the “Techno-Optimists” (Andreessen, Thiel, Musk) dominated, arguing that betting against technology is historically a losing strategy.

The Verdict: Score +7 (Technically Positive but Fatal)
The Takeaway: The “Hard Truth” engine. Despite a slightly positive vote count, Grok issued a No-Go verdict because the critique wasn’t about preference but it was about viability. It ignored the romance of the idea and attacked the fundamental investment thesis: zero leverage and zero moat.

You might notice the scores vary wildly (-7, +18, +7). This isn't a bug; it's a feature. The score represents the sentiment of that specific model's council, but the verdict comes from weighing the arguments. A project can have high support from "poets" (boosting the score) but can be a "No-Go" because the "operators" found a fatal financial flaw. The Council doesn't just count votes; it weighs evidence.

If you had asked a standard AI “Is this a good idea?”, it would have likely given you a polite, generic “Yes, digital detox is very popular right now!” It would have hallucinated a supportive cheerleader.

But the Council did not do that. The Council was messy. It was conflicted. It was real.

The variance in the scores proves the value of the Multi-Persona Debate.

• ChatGPT acted like a CFO, killing the project on unit economics.

• Gemini acted like a User Researcher, validating the pain point but questioning the adoption.

• Grok acted like a VC, rejecting the lack of leverage.

This is exactly what happens in the real world. If you pitch 100 investors, you won’t get a unanimous “Yes.” You will get a mix of hard passes, intrigued maybes, and philosophical debates.

By simulating this conflict, we didn’t just get a “Go/No-Go” decision. We uncovered the fundamental truth about the business: The soul of the idea is strong (Supporters), but the business engine is weak (Critics).

The Council didn’t just reject the idea; they offered the roadmap to fix it. They told us to stop selling coffee by the minute (commodity) and start selling a “sanctuary” membership (status).

To understand why this works, you have to understand a concept called Latent Space Traversal.

When you ask a standard question, the model pulls from its training data, the statistical average of all human text. That average is usually polite, agreeable, and safe.

But when you invoke “Jeff Bezos,” you force the model to move to a specific corner of its training data: shareholder letters, interviews, and biographies of aggressive capitalists. In that corner of the latent space, “politeness” is not the priority. “Efficiency” and “Scale” are the priorities.

By assigning conflicting roles, you create an Adversarial Network inside the chat. You are using the model’s own knowledge against itself to break the echo chamber.

It feels good to be told you are right. It feels uncomfortable to be questioned by a simulated Steve Jobs.

But the simulation is free. The real world is expensive.

The beauty of this “Council” mechanism is that it is infinite.

Are you redesigning your living room? Summon a council of Zaha Hadid, Frank Lloyd Wright, and Marie Kondo to debate “Flow vs. Function.”

Are you writing a speech? Summon Winston Churchill and Maya Angelou.

You can even run Andrej Karpathy’s “LLM Council” A multi-model variation where different AIs (Claude vs. GPT) debate each other directly. Karpathy himself has famously described LLMs as the “ghosts” of the internet’s data; this technique simply gives those ghosts a voice and a vote.

The goal isn’t to let the AI make the decision for you (of course not). The goal is to surface the blind spots you didn’t know you had. It gives you something to ponder, a counter-argument to wrestle with, and a perspective you would have otherwise missed.

Your AI is just sitting there, waiting for you to call the meeting to order. You just have to give them permission to speak freely.