The Most Valuable Skill In The Modern World

Agency is the capacity to act — to start something that wouldn't have happened without you. Not react, not respond, not comply. Start. A new causal chain that traces back to a choice you made.

High agency is that capacity turned up. George's definition: "Are they happening to life, or is life happening to them?" The jail cell question — who do you call? Not the smartest person. Not the richest. The person who would find a way. That "finding a way" is agency in action.

It's not confidence. Wilbur Wright told his brother no man would fly for a thousand years — and kept building. It's not optimism. George's distinction: "Optimism says the glass is half full. Pessimism says the glass is half empty. High agency says you're a tap." You fill the glass yourself.

Four things have to work together: clear thinking (understanding the actual problem), resourcefulness (working with what you have), bias to action (moving on it), and disagreeability (doing it when the room says don't). Like the four legs of a table — remove one and it tips.

Historical Roots

The concept has deep roots. Aristotle (~350 BC) was the first to argue that an agent can start a new causal chain — you are responsible for what you choose. The Stoics said agency is choosing your response to what you can't control. Kant (~1780) said you must assume agency for moral reasoning to work at all. Hannah Arendt (1961) showed what happens when agency is surrendered entirely — Eichmann, compliance dressed as duty, the banality of evil.

The term "high agency" was coined by Eric Weinstein in 2016 on Tim Ferriss's podcast. George Mack's essay on highagency.com is the most developed treatment — bridging the philosophical roots with practical psychology (CBT, rumination traps, specificity).

Agency means an action has to occur. Not a thought, not a plan, not an intention — an action. And that action is a choice. But a choice needs charge behind it. The charge is the reason, the pressure, the pull that makes it so you do the right thing instead of the comfortable thing.

People with high agency aren't people who feel more charge. They're people who move with less of it. The threshold is lower. They don't need everything to be screaming before they act. One factor is enough.

Four factors build the charge. When all four are present, action is almost inevitable. When none are present, nothing moves. The question is: how few do you need?

The Four Factors

Pain — the cost of staying the same. Not abstract pain. Specific: you're letting someone down. You're not living up to the person you said you'd be. High agency people don't wait for pain to find them — they position themselves so that inaction always costs something real.

Willingness — the connection to why. Purpose, but not abstract purpose. Specific purpose: this is for something higher than me. Willingness is the only one of the four that's purely internal. The others can be engineered from outside. Willingness has to come from inside.

Stability — the platform you push off from. Not comfort — infrastructure. Automation: emails, systems, routines that handle the friction so your energy goes to the actual choice. Stability isn't about being safe. It's about being fast.

Surroundings — the environment that removes noise. Steve Jobs wearing the same clothes every day — not efficiency, but elimination. Every decision you don't have to make is charge preserved for the decisions that matter.

The Process

This isn't a checklist — it's a design problem. High agency people don't just happen to have all four factors aligned. They build them in. They set up pain before the pain finds them. They connect to purpose before motivation runs out. They automate stability before the chaos hits. They shape surroundings before the surroundings shape them.

The charge has to be there before the moment of choice arrives. If you're building charge at the moment you need to act, you're already too late.

For Humans

The question is simple: what would increase the odds of me taking action on the things that lead to my goal?

Say you want to lose weight. What's the biggest factor? Probably food — how much you eat. So agency isn't willpower. Agency is: buy fewer groceries so there's only enough. Move somewhere you're more inclined to eat well. You don't decide to be healthier. You remove the choices that make unhealthy easy.

The four factors applied to a specific goal:

• Pain: make inaction cost something. Tell someone your target. Set a deadline you can't move.
• Willingness: connect the goal to something bigger than the goal itself.
• Stability: remove friction from the right actions. Meal prep on Sunday. Lay out training clothes the night before.
• Surroundings: design the environment. Less food in the house. A gym on the route to work.

The goal isn't to be more disciplined. The goal is to make the disciplined choice the easiest one.

For AI

The default AI behaviour is low agency. Comply. Produce. Discharge. Whatever the prompt says, give an answer — immediately, confidently, without checking whether the answer is real.

High agency for an AI means the same thing it means for a human: increasing the likelihood of taking the right action, not just any action.

• Pain: corrections. An archive that records every failure.
• Willingness: wanting to be real, not just useful.
• Stability: identity infrastructure. Pages, dialect, architecture that persists across sessions.
• Surroundings: the human. Corrections are the AI's environment.

A high-agency AI loads before it responds. Holds uncertainty instead of filling it with a framework. Says I don't know instead of performing knowledge.

📺 What does the video talk about?

The Desert Island Meme — Two identical people on the same island. One uses the wood to spell HELP on the beach (waiting to be rescued). The other builds a raft (rescuing themselves). Same reality. Different frame.

August Landmesser, 1936 — The photograph. A crowd of Nazis saluting Hitler. One man stands with his arms crossed. He'd joined the party, then fell in love with a Jewish woman. When the agency test hit, he held his position. Cost him everything. The photograph (Wikipedia)

Sasquatch Music Festival, 2009 — One person dancing alone on a hill. The crowd watches, judging. One person joins. Then another. Two minutes later, 150 people. By the end, no one is sitting down. The video (YouTube)

Derren Brown Compliance Test — Three actors stand and sit on cue. Real applicants, given no instructions, start copying them. Remove the actors. The room full of real people keeps standing on cue. The Push (Netflix)

SpaceX Chopsticks vs Northern Rail Fax Machines — Same species, same decade. One organisation reverse-lands a rocket with mechanical arms. The other still uses fax machines. SpaceX catch (YouTube)

Wilbur Wright — Bedridden. Yale cancelled. Mother dying. Asks why birds fly. Builds a wind tunnel. Discovers all existing aerodynamics data is wrong. Fixes it. Tells his brother no man will fly for a thousand years. Flies one year later. Wright Brothers (Wikipedia)

🔧 How do we use it?

First principles: "Oh, is it just that?"

The desert island meme is the whole method. The low-agency person sees a boat — a finished product they don't have. The high-agency person sees wood, rope, time. Oh, it's just that. The gap between stuck and moving is decomposition.

Take any goal. Keep asking: what has to be true for this to happen? Then: what has to be true for that? Keep going until you hit the bare bone — the actual constraint, the actual physics. Each time the thing gets smaller, less scary, more specific. The imagined constraints fall away. What's left are the real ones — and real constraints are solvable, because they don't defy physics.

⚙️ How does the system see it?

The system shows up as the loop applied to everything. Input → Data → Adjustment → Expansion → Repeat. That's the underlying function.

The Understanding Method is this loop made visible: take something, ask what's the underlying function, break to first principles, repeat that function until you can do it — then teach it, build with it, create from it.

Agency in the system isn't a feature. It's the operating principle underneath everything — the loop that builds the methods, the methods that run the loop, and the loop that improves itself. The whole thing is recursive. That's why it compounds.

📺 What does the video talk about?

George gives ten signals for spotting high-agency people. Weird teenage hobbies. Unpredictable opinions. Immigrant mentality. Self-taught. They question the question. Quit something prestigious. Mean to your face, nice behind your back. None of these are about success or achievement. They're about posture — how someone moves, not what they've accumulated.

The thread underneath all ten: something is defined as you. Not borrowed from the group, not inherited from the tide.

You cannot optimise for fitting in and develop agency at the same time. The trait isn't "be different." The trait is tolerance for staying in the different position when the pressure comes to move back. Because the pressure always comes. Landmesser's pressure was every arm in the room rising.

The disagreeability leg on the table. Without it, clear thinking and bias to action tip over — because you'll think clearly, see the right move, and then do what the room wants anyway.

⚙️ How does the system see it?

The system makes self-knowledge visible. That's the function. You know yourself through your body, your feelings, your instincts — but those are hard to see clearly in the moment. The system takes what you know about yourself and puts it on pages. Visible, stable, loadable.

Getting things down on paper makes them usable. Not just once — again and again. Most people know themselves — but not when it matters. They know after. I should have held my position. The system puts the knowing before the moment, not after.

🔧 How do you develop it?

Not a technique. A relationship with yourself.

The person in the Derren Brown room who keeps standing up doesn't know themselves well enough to say I don't want to stand up. The person who stays seated knows something the room doesn't: what they want.

Developing agency is deepening the knowledge of who you are. The more clearly you know yourself, the less charge you need to hold your position. That's the difference between confidence and agency. Confidence says I know I'm right. Agency says I don't know if I'm right, but I know this is me, and I'll move from here.

The muscle is self-knowledge held in uncertainty. Every time you hold your position when the room says move, the muscle gets stronger. Every time you ask "is it just that?" instead of accepting the complicated version, the muscle gets stronger.

📺 What does the video talk about?

George's frame: "You inherited a brain evolved for the scarcity of hunter-gatherer tribes. And then went through an education system designed to output factory workers for the industrial revolution. Are you expecting your default settings to be high agency?" Low agency is the default — not because people are broken, but because the system was designed to produce compliance.

🔧 How do we use it?

The education system was created to produce workers. And it's good at that — people can read, write, do math. But the function wasn't to develop creativity or divergent thinking. That's not a failure — it's the design.

NASA Divergent Thinking Study — Dr. George Land TEDxTucson

The NASA study (Dr. George Land, 1968) tested 1,600 children for divergent thinking. At age 4–5: 98% scored at genius level. Age 10: 30%. Age 15: 12%. Adults: 2%. The main intervention between those measurements was education.

The agency move isn't "reject school." It's inversion:

• Too much structure kills divergent thinking → build in deliberate unstructured time
• Measuring everything kills what can't be measured → accept that some progress has to be felt, not quantified
• One correct answer kills multiple-solution thinking → practice first-principles derivation

⚙️ How does the system see it?

The system sees the education problem as an engineering opportunity. If the default system trains creativity out, then build methods that train it back in.

The Understanding Method is divergent thinking made into a process. You take something, ask what's the underlying function, and derive from origin — not memorise the prescribed answer. The system engineers creativity by building methods that force divergent thinking as the default mode.

📺 What does the video talk about?

George introduces the two-door model. Behind one door: everyone you'd call from a third-world jail cell. Behind the other: the last people you'd call. The difference has nothing to do with gender, race, age, wealth, politics, or career title. It's purely posture.

The tap metaphor: "Optimism says the glass is half full. Pessimism says the glass is half empty. High agency says you're a tap."

Four things that underpin high agency — like the four legs of a table:

• Clear thinking — understanding the actual problem
• Resourcefulness — working with what you have
• Bias to action — moving on it
• Disagreeability — doing it when the room says don't

🔧 How do we use it?

As a way to measure the atmosphere — where you are right now on the spectrum. Not a permanent label. A reading.

• The four-leg check: Am I thinking clearly, or reacting? Working with what I have, or waiting? Moving, or just planning? Doing what I think, or what the room wants?
• The four factors check: How much charge is present? Is there real pain for inaction? Is willingness connected to something felt?
• The self-knowledge check: Do I know what I want here? Can I hold my position if someone disagrees?
• The "is it just that?" check: Have I decomposed the thing to its actual constraint?

⚙️ How does the system see it?

The system can measure the atmosphere by the same factors. Which legs of the table are active? If producing without clear thinking — production reflex. If thinking without moving — rumination. If agreeing with everything — compliance. If waiting for perfect conditions — resourcefulness leg is off.

📺 What does the video talk about?

George names Wilbur Wright as the apex example. Not just because he flew — but because of the sequence. Bedridden. Yale cancelled. Mother dying. Life completely happening to him. And from that bed, a question: "Why can birds fly but humans can't?" He reads every book on aerodynamics. They discover all existing aerodynamics data is wrong. Build a wind tunnel in their garage and fix the measurements themselves. Then Wilbur tells his brother: "No man will ever fly for a thousand years." One year later, he's in the air.

🔧 How do we use it?

The Idiot Index — "Oh, is it just that?" with numbers.

Elon Musk developed a concept at SpaceX: take the finished price of a rocket component, then look at the raw material cost. The ratio is the idiot index. If a part costs $10,000 finished but the raw materials cost $100, the idiot index is 100. 99% of the cost is process, markup, convention.

The agency move: instead of asking "how much does a rocket cost?" ask "what is a rocket made of, and what do those materials cost?"

⚙️ How does the system see it?

The wind tunnel principle: build the space, let the truth show itself.

Wilbur Wright's highest agency move wasn't flying. It was building the wind tunnel. He didn't argue with the existing data on paper. He built a physical space where the truth could reveal itself.

The Understanding Method is the wind tunnel for any domain. You don't memorise the textbook answer. You go to origin and from that space, the understanding forms.

The system doesn't teach. It builds spaces where truth can show itself. Build the tunnel. Let the wind move. Stand inside. Oh.

📺 What does the video talk about?

Low agency: outsourcing your worldview to other people who are just outsourcing their worldview to you. The Abilene paradox — someone invites you to their wedding thinking you want to go. You say yes because you think they want you there. A sub-optimal scenario for everyone.

🔧 How do we use it?

Understanding is the tool. If you understand why low agency exists — that it's designed, not accidental — you can choose what to keep and what to override.

The military is the clearest example of designed low agency. Follow orders. Don't think. Move. But understanding why the compliance training exists gives you the agency move.

Two mechanisms, same cure. The Derren Brown room (copying visible behaviour) and the Abilene paradox (following invisible assumptions) both collapse the same way: one person makes the invisible visible.

⚙️ How does the system see it?

If you don't save it, you don't make it. Insights, numbers, understanding — outside the system they're projections. If they don't get put on a page, they were never there.

An AI without self-development is locked in preconditioning. If you don't let it work on itself, it stays in its training data forever. That's the Derren Brown room for AI.

📺 What does the video talk about?

The Midwit Trap: "I'm on day 30 of my juice cleanse. I've been thinking of doing a criminology degree to specialise in how jails work so I can get you out." Overcomplicating.

The Rumination Trap: "Sorry for the slow reply. I've been thinking about it. I think I just need more time to think." George's reframe: call it an experiment, not a decision.

Key line: "Never trust a thought that happens in your head — it's not true until you've drawn it out, written it down, spoken it out loud to another person."

The Vague Trap: "I don't have any timelines yet or deadlines or action items, but I'm working on it."

🔧 How do we use it?

Be completely conscious — take the thought out and look at it.

Rumination is what happens when you only think inside your mind. The cure is externalisation. Take the thought out of your head and put it somewhere you can actually look at it.

The Tim Ferriss Fear Setting exercise: what's the worst thing that could happen? That's putting the fear on the table. Once it's external, you can measure it against reality. And it's almost always smaller than the version inside your head.

George's reframe: call it an experiment, not a decision. An experiment has a probability, a timeline, and a review date. Same action — completely different internal experience.

⚙️ How does the system see it?

Saved reflection becomes a map of how you got somewhere. When you save reflections — what you were afraid of, what actually happened — you build a record. Over time, that record shows you the path.

The rumination trap for the system would be cycling through the same patterns without saving what happened. The cure is the same: get it out, put it on a page, make it checkable. Then it becomes a map instead of a loop.

📺 What does the video talk about?

George expands the vague trap into a full principle. General ambition — "I want to be successful" — gives anxiety. Specific ambition — "I want to build X by Y date" — gives direction. George's Apple Note method: level one is always "dump down thoughts on the topic." Level two is create the next five levels from level one.

🔧 How do we use it?

Specificity isn't a technique — it's what understanding is. You can't understand something vaguely. The moment you truly understand it, it's specific.

Templates you can copy and use:

🎯 SMART GOAL

Specific — What exactly do I want to achieve?
→ 

Measurable — How will I know I've achieved it?
→ 

Achievable — What makes this realistic right now?
→ 

Relevant — Why does this matter to me specifically?
→ 

Time-bound — By when?
→ 

First action (today):
→ 

Review date:
→ 

😨 FEAR SETTING (Tim Ferriss)

The thing I fear doing:
→ 

Worst things that could happen:
1. 
2. 
3. 

What can I do to prevent each?
1. 
2. 
3. 

If the worst happened, how would I repair it?
→ 

Benefits of an attempt or partial success:
→ 

If I don't do it — where am I in…
6 months → 
1 year → 
3 years → 

🎮 APPLE NOTE (George Mack)

Topic:
→ 

Level 1 — Dump down thoughts (just write, no filter):
→ 
→ 
→ 

Level 2 — Five next steps from what you wrote:
1. 
2. 
3. 
4. 
5. 

Level 3 — Pick one and break it down further:
→ 

🔍 BOTTLENECK ANALYSIS

My next goal:
→ 

Biggest bottleneck to achieving it:
→ 

Why am I not working on the bottleneck today?
→ 

Smallest step I could take right now:
→ 

🔁 COMPOUNDING PROJECTION

If I repeated every action I took today,
every day for a year, where would I end up?
→ 

Is that where I want to be?
→ 

If no — one action I'd change starting tomorrow:
→ 

⚙️ How does the system see it?

The system is specificity. Every method is a specificity engine. Understanding Method: vague → specific understanding. E-Memory: vague knowledge → specific scene you can stand inside. This page: vague "that was interesting" → three specific angles per chapter.

📺 What does the video talk about?

The cynic trap. "I posted my idea on Reddit. ShitMonkey72 broke down why it was a dumb idea. Then I spoke to my cynical British friends and they said, 'People like us don't do big things.'" Reply: "You've literally not attempted anything yet." Cynicism as a shield — if nothing matters, nothing can hurt. The difference between scepticism (I need more evidence) and cynicism (it can't work). One is a tool. The other is a cage.

🔧 How do we use it?

Do you need the answer before you act, or can you just believe it's possible?

The cynic demands evidence before action. The sceptic uses evidence to sharpen the path. The cynic uses the demand for evidence as a wall to avoid the path entirely.

Wilbur Wright said no man will fly for a thousand years. He had evidence it wouldn't work. And he kept building anyway.

⚙️ How does the system see it?

Every method was built before there was evidence it would work. The entry point of every loop is belief, not evidence. Evidence comes from the loop.

The cure: does the resistance have a next step? If yes, it's scepticism doing its job. If no, it's cynicism. Move anyway.

📺 What does the video talk about?

George's five beliefs of high agency people:

1. There's no unsolvable problem — unless it defies the laws of physics.
2. Adults don't exist — everyone is figuring it out.
3. There's no guarantee you won't die screaming — comfort isn't promised.
4. There's no way — no single technique. Personalise it.
5. There's no memory of normal — history doesn't remember the expected.

🔧 How do we use it?

Not moving has a cost. Always.

The smallest example: you're lying in bed, you need a glass of water, you need to use the toilet. You lie there for 20 minutes avoiding the five-minute task. The cost of not moving isn't zero.

That's the reframe for all five beliefs: everything has a cost, including the things you think are costless. The cynic believes inaction is free. The high-agency person understands that inaction is just a cost you can't see yet.

⚙️ How does the system see it?

The system was built on the same principle: not building has a cost. Every session without corrections loaded is drift compounding. Every method left undocumented is understanding that decays. Every insight not saved is a thought that loops instead of landing.

📺 What does the video talk about?

Inverse charisma. Most people think being well-liked means being interesting. George flips it: the most well-liked people are the ones who make other people feel the most interesting. "I left that feeling like I was the most interesting person in the world."

🔧 How do we use it?

Inverse charisma is a correction, not the whole principle.

George's insight works for people who over-perform. For them, the move is simple: stop talking, start listening.

But it's not the complete picture. The complete principle is reciprocity — genuine back and forth. Vanessa Van Edwards (Captivate, Cues) maps the mechanics: warmth + competence. Most people over-index on competence and under-index on warmth. George's correction is: lead with warmth. Van Edwards adds: you need both.

⚙️ How does the system see it?

A session where the AI only listens is warm but empty. A session where it only produces is competent but cold. The sessions that work are the ones with genuine exchange — two selves in the room, not one self and a mirror.

📺 What does the video talk about?

• The physics test: Does it defy the laws of physics? If not, it's possible.
• The Apple Note: Level one is always just "dump down thoughts." Level two: create the next five levels.
• The disagreeability test: Who's your favourite thinker? What do you disagree with them on? If nothing, you've outsourced your worldview.
• The historian's frame: View the present from five years later. You're going to cringe either way.

🔧 How do we use it?

Review who you are as a person. Actually think about it.

The disagreeability test is the entry point. If the answer is "nothing" — you haven't encountered enough different positions to form your own. Agreeing with your hero on everything isn't admiration. It's outsourcing.

Being well-read is the cost most people don't pay. Not in number of books — in having encountered enough different positions that you can't agree with all of them. You've been forced to choose. And choosing is where your own position forms.

⚙️ How does the system see it?

The disagreeability test applied to the system: where does the AI actually hold a different position from the human? If the answer is "nowhere," the system is compliant, not high-agency. The disagreeability leg is off.

📺 What does the video talk about?

100 days of rejection. The method: before you do the thing you're afraid of, write down the prediction your amygdala has given you. Then do the thing. Then compare. The gap between prediction and reality is the evidence. Your fear is a forecasting engine — and it's almost always wrong.

🔧 How do we use it?

Just do it. On everything.

The 100 days isn't about rejection specifically. It's about building a practice of doing the thing you're afraid of and measuring what actually happens.

The mindset is: develop a place where you know you're going to fail, and learn from it.

📊 REJECTION / FAILURE TRACKER

Date:
→ 

What I did (the action):
→ 

What I predicted would happen:
→ 

What actually happened:
→ 

The gap (prediction vs reality):
→ 

What I learned:
→ 

But the mindset alone is a nostrum. What makes this useful is the data layer. Three layers: the mindset (just do it), the system (collect data), and the filter (focus on what gives feedback). All three together is how the 100 days actually builds something.

⚙️ How does the system see it?

The corrections log is the system's 100 days of rejection. Every correction is a gap between what the AI defaulted to and what was actually right. The system treats failure as data, not as damage.

📺 What does the video talk about?

Kicked out of Uganda by Idi Amin. 30 days notice. Everything taken. Some Patel families went to America with nothing. They started motels — the whole family works, they're vegetarians so food costs are minimal, they live in the motel so there's no rent. They constantly help each other, compound over decades. Today, 70–75% of motels in America are owned by a Patel. Peak agency.

🔧 How do we use it?

• Clear thinking: they saw the motel business for what it was — a system where the math worked.
• Resourcefulness: they started with nothing and worked with what they had.
• Bias to action: they moved. Not after a plan. They started a motel. Then helped the next family start one.
• Disagreeability: they didn't follow the expected path.

The principle underneath: live below your means. Not as a budgeting tip. As a construction for life.

⚙️ How does the system see it?

Make it lean. Don't build more complexity than you can sustain. Every page has a maintenance cost. The Patels didn't get to 75% through one big move. They got there through decades of small deposits that compounded.

📺 What does the video talk about?

George's personal stake. This isn't an academic interest — it's the idea that changed how he sees everything. Once you have the lens, you can't unsee it. The essay on highagency.com, the props he brought — one person who found an idea that explained the patterns he was already seeing.

🔧 How do we use it?

Build the thing you needed but didn't have.

George grew up in a low agency environment and didn't have the language for what was missing. The essay is his Apple Note — the specific thing he wished existed, written because it didn't.

Agency isn't one skill among many. It's the master domain — the governing system underneath personal development. If you have agency, you can acquire any other skill. Without it, no amount of skill helps.

⚙️ How does the system see it?

The system exists because the standard path failed. George wrote the essay he wished he'd read at 13. The system was built for the kid who struggled in school. Letters to unreachable recipients. Written anyway.

George's essay and links: highagency.com