Founders Podcast · EXTRACTED

How Elon Works

The algorithm, the idiot index, and the five-step system behind three decades of company building. Complexity is the enemy of scaling, and Elon has one answer for it.

▶564K views on YouTube

"I became a broken record on the algorithm. I think it's helpful to say it to an annoyingly degree." — Elon Musk
David Senior, on the episode

This is a solo episode from Founders, in which host David Senior spent over 60 hours reading and rereading Walter Isaacson's 615-page biography of Elon Musk to extract the operating system underneath the noise. The pop framing of Elon is politics, controversy, and personality. The actual operating system is a small set of company-building principles he has applied, repeated, and iterated across at least seven companies over three decades. Senior strips away everything that is not how Elon works and presents the enduring principles in chronological order, showing how they appear, reappear, and get refined across Zip 2, PayPal, SpaceX, Tesla, and beyond. This protocol pulls the three highest-leverage plays from that extraction.

Tactic 01

Run the Algorithm

By the time Tesla was fighting for survival in 2018, Elon had distilled everything he learned from production hell at the Nevada and Fremont factories into five sequential steps he called the algorithm. His executives had heard it so many times they would mouth the words as he spoke. He said, I became a broken record on the algorithm. I think it's helpful to say it to an annoyingly degree. The five steps are: one, question every requirement, and attach the name of the specific person who made it, never a department. Requirements from smart people are the most dangerous because people are less likely to question them. Two, delete any part or process you can. If you do not end up adding back at least 10% of what you deleted, you did not delete enough. Three, simplify and optimize, but only after step two, because a common mistake is to simplify a process that should not exist at all. Four, accelerate cycle time, but only after the first three steps. Elon said he wasted enormous time accelerating processes he later realized should have been deleted entirely. Five, automate last. The catastrophic mistake at the Nevada and Fremont factories was beginning with automation before requirements had been questioned, parts deleted, and bugs shaken out. A deleted plastic pin cap illustrates the whole system. Battery packs shipped from Nevada with plastic caps on the prongs to prevent bending. When the batteries arrived in Fremont, the caps were removed and discarded. Sometimes Nevada ran out of caps and had to hold entire shipments. Elon asked who specified the caps. The factory team could not find a name. So he deleted them. There was never a problem with bent pins.

The play

Take one process, checklist, or requirement in your company this week that has existed long enough that no one questions it. Find the name of the actual person who created it, not the department. Ask whether it would break a law of physics if you removed it. If it would not, delete it and run for two weeks without it. Document what breaks, if anything. Apply this to one new item each week until it is a reflex.

Tactic 02

Use the Idiot Index to Kill Cost Creep

Before Elon built a single rocket, he went to first principles on why rockets cost what they cost. He calculated the raw price of the carbon fiber, metal, and fuel that went into a finished rocket, then compared that to the price of the finished product. The finished rocket using standard manufacturing methods cost at least 50 times more than its raw materials. He called this ratio the idiot index. If a product had a high idiot index, its cost could be reduced significantly by devising more efficient manufacturing techniques. He applied it relentlessly and expected everyone around him to know the numbers cold. In one meeting he asked an engineer the idiot index on a specific Raptor engine part. The engineer said he thought the part cost $13,000 and that the raw material was a few thousand dollars. Elon told him the part was made of steel and the steel cost $200. He told the engineer that if he came into a future meeting without knowing the idiot and best parts off the top of his head, his resignation would be accepted immediately. The principle showed up in every company. A supplier quoted $120,000 for a part Elon judged no more complicated than a garage door opener. He told an engineer to make it instead. The engineer made it for $5,000. NASA latches cost $1,500 each. A SpaceX engineer modified a bathroom stall latch and produced a locking mechanism for $30. An air cooling system for the Falcon 9 was quoted at $3 million. Elon asked Gwynne Shotwell what a house air conditioning unit cost. About $6,000, she said. They bought commercial units, modified the pumps, and made them work on top of the rocket.

The play

Pick five parts, services, or line items your company pays for regularly. For each one, find or estimate the raw material or commodity cost underneath it. Divide the price you pay by that raw cost. Any ratio above 10 is your first target. Bring that ratio to your next team meeting and ask the person responsible for it to either justify the gap or find a way to close it within 30 days.

Tactic 03

Showmanship Is Salesmanship: Build the Demonstration First

In 2009, Tesla was weeks from bankruptcy. Daimler executives were visiting the US and invited Tesla to present a proposal for an electric version of their Smart car. The standard move would have been a PowerPoint deck. Elon's team bought a Smart car, pulled out the interior, and dropped in a Tesla Roadster electric motor and battery pack. When the Daimler executives arrived expecting slides, Elon asked them if they wanted to drive the car instead. The car bolted to 60 mph in four seconds. The executives were shocked. Elon asked them to invest. Daimler agreed to a $50 million equity stake. Elon said later, if Daimler had not invested in Tesla at that time, we would have died. This was not a one-time improvisation. It was a principle Elon had been running since Zip 2, where the team put a small computer inside a large empty rack so that visiting investors would think they had a massive server. He ran the same play when unveiling the first Tesla Roadster prototype, where he personally took over planning the event, oversaw the guest list, chose the menu, and approved the cost and design of the napkins. He designed the Model S door handles to sense the driver approaching, light up, and pop out. Everyone told him a standard door handle would work just as well. He said it would send a chirpy signal of friendliness, it was fun and unexpected, and that when you surprise and delight customers they form a deeper emotional bond with your company and tell other people. The principle behind all of it: belief is transferable, but only if you give people something to feel, not something to read. The money flows as a function of the stories, and the most powerful story is a working thing in someone's hands.

The play

Before your next important pitch, investor meeting, or product review, identify the one claim at the center of your case and build the smallest working version of it that the other person can experience directly, touch, drive, or use in the room. Replace the slide that makes the claim with the object that proves it. If you cannot build the object in time, find the closest existing proxy and put the person inside it. A reaction beats a slide every time.

Tactic 04

Think In Burn Rate, Not Budget

At SpaceX, Elon held a view of money that none of the aerospace engineers around him were used to. An early employee recalled him laying it out plainly: everything the company did was a function of burn rate, and they were burning through $100,000 per day. That framing changed how every spending decision looked. It was not a question of whether a purchase was expensive in the abstract. It was a question of what a day of lost progress actually cost. The logic ran in both directions. Sometimes Elon would refuse to let someone buy a part for $2,000 because he expected them to find it cheaper or build something cheaper themselves. Other times he would not flinch at renting a plane for $90,000 to get a component to the island launch site, because it saved an entire workday, and that was worth it. The calculus was always the same: measure the cost of the action against the cost of the delay, not against the sticker price alone. He placed this urgency on the team by telling them he expected revenue in ten years to be $10 million a day, and that every day they were slower to hit their goals was a day of missing out on that money. Most people treat spending as a budget question. Elon treats it as a time question. The $90,000 plane is cheap if the alternative is a lost day at $100,000 in burn. The $2,000 part is expensive if it can be built for $500 and the process of building it in-house teaches the team something. The unit of measurement is not the dollar, it is the day.

The play

Calculate your actual daily burn rate and write it somewhere visible. Then run every non-trivial spending decision through a single question: does this cost less than a day of lost progress? If paying more gets the work done today instead of next week, it is often the cheaper choice. If a cheaper option exists and finding it takes one afternoon, find it. The goal is not to spend less in total but to stop treating time and money as separate ledgers when they are the same ledger.

Tactic 05

Isolate Yourself In Silent Darkness To Think

Elon spends an unusual amount of time in complete silence. Not meditating, not journaling, not talking through problems with a team. Just lying flat in a dark, silent room with no stimuli, letting his mind work. The author notes this pattern appearing repeatedly across the book. At one point during the most brutal stretch of the Tesla production crisis, Elon told the people around him: "I've got to lay down. I've got to shut off the lights. I need some time alone." At other points he would sit motionless on the edge of his bed for an entire night, staring into space. His girlfriend at the time described waking up every couple of hours to find him still in the same position, completely still, not asleep, just thinking. The practice mirrors something Jim Simons, founder of Renaissance Technologies, used throughout his career. Simons would lie flat in complete darkness and silence with no stimuli at all and think through problems without trying to control the direction of his thoughts. No phone, no music, no background noise. Just the mind running on its own. What both men found is that removing all external input creates the conditions for the brain to surface solutions it cannot reach when it is constantly processing incoming information. The host started using this practice after reading the Simons biography and describes the result: after 20 to 60 minutes of lying flat with earplugs and a sleep mask in complete silence, without trying to direct his thoughts, the solutions come and the energy afterward is remarkable.

The play

The next time you are stuck on a hard problem, do not schedule a meeting, open a browser, or talk it through with someone. Instead, lie flat on a couch, a bed, or the floor in a completely dark and silent room. Put in foam earplugs. Add a sleep mask if needed. Do not try to solve the problem. Do not try to control your thoughts. Just remove all stimuli and let your mind run for at least 20 minutes, longer if you can. The absence of incoming information is what makes this work. Your brain keeps processing in the background whether you direct it or not, and silence gives it the conditions to surface what it cannot reach when it is being constantly interrupted.

Tactic 06

Question, Delete, Simplify, Accelerate, Automate, In That Order

Elon did not arrive at his five-step algorithm by sitting down and thinking it through. He arrived at it by making every mistake in the wrong order, most expensively by automating too early. At the Nevada and Fremont factories, he pushed hard for robots before the underlying process had been cleaned up. The robots went in, and the line broke. "We had this enormously automated production line that used tons of robots," he said. "There was one problem. It didn't work." They ended up sawing robots out of the production line and throwing them into the parking lot, cutting a hole in the side of the building to remove the equipment. The lesson he pulled out of that wreckage became the skeleton of the algorithm. The five steps are: question every requirement, delete any part or process you can, simplify and organize, accelerate cycle time, and then, only then, automate. The sequence matters as much as the steps. Simplifying before deleting means you optimize something that should not exist. Accelerating before simplifying means you run a broken process faster. Automating before all of the above means you permanently lock in every flaw at machine speed. "Always wait until the end of designing a process," Elon said, "after you have questioned all the requirements and deleted unnecessary parts, before you introduce automation." He also built in a self-check on deletion: if you do not end up adding back at least ten percent of what you cut, you did not delete enough. The algorithm came with postcripts he repeated just as often. Technical managers must have hands-on experience, spending at least twenty percent of their time doing the actual work, not supervising it. When there are problems, skip the managers and meet directly with the level below them, the people with their hands on the thing. And the only rules that cannot be questioned are the ones dictated by the laws of physics. Everything else is a recommendation.

The play

Take one process in your company that feels slow or expensive. Walk it through the five steps in strict order. Step one: list every requirement and find the name of the actual person who created each one, not the department. Question each one regardless of how senior the source. Step two: delete every part and step you possibly can, knowing you may add some back. If you do not restore at least ten percent, you cut too lightly. Step three: only after deleting, simplify and reorganize what remains. Step four: now speed up the cycle. Step five: only now consider automation. Running the steps out of order, especially automating early, will scale your problems instead of your output.

Tactic 07

Walk To The Red

During the Model 3 production crisis, Elon needed a way to move through a factory floor with hundreds of stations and immediately know where his time was needed. He installed monitors showing each station on the assembly line with a green or red light indicating whether it was flowing properly. The system had a name: walking to the red. He would head straight to any red light he saw. The method was blunt and fast. What's the problem? A part was missing. Well, who's in charge of the part? Get him over here. A sensor keeps tripping. Who calibrated it? Find someone who can open the console. Can we adjust the settings? Why do we even need that sensor? He would then resume his procession through the factory looking for the red lights. The monitor did not tell him what to think. It told him where to go. Everything else followed from showing up at the right station and asking the right questions. Elon calculated that he made a hundred command decisions a day as he walked the floor. At least 20% are going to be wrong, he said, and we're going to alter them later. But if I don't make decisions, we die. The red lights gave him a way to concentrate those decisions where they actually mattered, in real time, on the physical line, with the people doing the work standing right there.

The play

Install a visible, real-time status display for every critical stage in your operation. Each stage gets a simple indicator: flowing or blocked. Then walk the line and move directly to every blocked stage. Do not schedule a review meeting. Do not wait for a report. Ask what the problem is, who owns the relevant part or process, and whether the requirement that created the bottleneck actually needs to exist. Resolve it or escalate it on the spot. The goal is to convert a floor walk into a hundred small decisions per day, each made at the source of the problem rather than three levels away from it.

Tactic 08

Combine Urgency With A Working Demonstration

When Daimler executives visited Tesla, they were expecting a PowerPoint. They got a car. Tesla had taken a stock smart car, pulled out the drivetrain, and dropped in a Roadster electric motor and battery pack. When the Daimler team arrived, Elon asked if they wanted to drive it. They did. The car bolted forward and hit 60 mph in 4 seconds. Daimler agreed to a $50 million equity stake shortly after. As Elon put it: 'If Daimler had not invested in Tesla at that time, we would have died.' What makes this more than a clever sales story is the combination of two principles Elon uses across every company. The first is that showmanship is salesmanship and that a working physical demonstration carries an authority no slide deck can match. The second is that maniacal urgency is an operating principle, not just a motivational phrase. Tesla scrambled to build the prototype specifically because Daimler's visit had a hard date. The demonstration existed because the deadline forced it into existence. Remove either element and the outcome changes. No urgency, no prototype ready in time. No prototype, no deal. Elon uses this combination repeatedly. The goal is never to explain what a product will do. The goal is to make someone feel it. When the Roadster was first unveiled, Elon personally oversaw the event, approved the guest list, chose the menu, and signed off on the cost and design of the napkins. The demonstration is the pitch. The pitch is the demonstration. Belief, as Elon understood from his earliest days at Zip 2, transfers through experience, not description.

The play

Before your next important pitch, investor meeting, or customer conversation, identify the single most visceral thing your product or service can do and build a working version of that moment first. Do not present a roadmap. Do not show a mockup. Put the buyer inside the experience. If resources are tight, strip the prototype to the one capability that will make the biggest impression and build only that. Set a hard deadline tied to the meeting date and treat it as non-negotiable. A working demonstration delivered under time pressure closes deals that months of polished decks cannot.

Tactic 09

Set Binary Survival Milestones And Repeat Them Until Everyone Knows

During Tesla's production hell, Elon did not hand his executives a strategic plan or a roadmap with phases. He did the math on costs, overhead, and cash flow, and he reduced the company's survival to a single number. Five thousand Model 3s per week. That was the line. Hit it and Tesla lives. Miss it and the company runs out of money. He repeated this to every executive like a mantra. He installed monitors on the factory floor showing the up-to-the-minute output of cars and components. The target was never more than a glance away. The power of this is not just clarity, it is compression. A company fighting for its life cannot afford ambiguity about what winning looks like. When the goal is simple enough to repeat in one sentence, every decision on the floor has a test it can run against. Does this get us to five thousand? Yes or no? Elon had done the calculations. He shared them. He moved himself physically onto the factory floor and led what he called an all-in search. He called one of his employees and said meet me at the Van Nuys airstrip in an hour. They flew to Reno and the employee ended up staying there for four months. The milestone was not a management abstraction. It was a countdown. Elon describes this approach as part of how he drives urgency across an entire organization. He became a broken record on the goal. His executives learned to repeat it back. He believed, and the record bears this out, that if the people around you do not already know what you are about to say, you are not saying it enough. The monitor on the factory wall was not motivational decoration. It was a live feed of whether the company would exist next month.

The play

Identify the single metric that determines whether your company or project survives the next six to twelve months. Do the actual math on cash flow, unit economics, or output rate until you land on one number. State it in one sentence: we must hit X by Y or we are dead. Repeat it to every executive and team lead, not once in a kickoff meeting but in every meeting, every update, every floor walk, until they mouth the words before you finish saying them. Install a visible, real-time display of progress toward that number where the people doing the work can see it. Move your own desk or schedule closer to the constraint. The goal of the milestone is not to motivate. It is to make every daily decision self-sorting.

Tactic 10

Test Requirements Before You Trust Them

During production hell at Tesla, Elon noticed the assembly line slowing at a station where strips of fiberglass were being glued to battery packs by an expensive, slow robot. He asked the most basic question he could: what the hell are these strips for? The engineering team said the noise reduction team had specified them to cut down on vibrations. The noise reduction team said the specification came from the engineering team to reduce fire risk. Nobody owned it. Nobody could trace it back to a person or a reason that held up under pressure. It was, as Elon put it, like being in a Dilbert cartoon. Rather than debate it in a conference room, he ordered a simple test. Record the sound inside the car without the fiberglass. Then record it with the fiberglass. See if you can tell the difference. They couldn't. The requirement disappeared along with the robot, the bottleneck, and the cost. The lesson Elon drew from this was not that the engineers were stupid. It was that requirements accumulate and calcify, and the only way to find out which ones are real is to go test them. "Step one should be to question the requirements," he said. "Make them less wrong and dumb, because all requirements are somewhat wrong and dumb." The same logic applied to plastic caps placed on battery pack pins to prevent bending during transit from Nevada to Fremont. When he asked who specified the caps, nobody could name the person. He deleted them. The pins never bent.

The play

The next time your team defends a specification by citing a department rather than a person, treat it as untested. Assign someone to run the simplest possible empirical check: record with and without, measure before and after, build with and without the part. Do it this week, not next quarter. If the test shows no measurable difference, delete the requirement. If nobody can name the person who made the requirement in the first place, that is your signal to run the test immediately.

Tactic 11

Find The Limit Through Empirical Testing

During Starship construction, Elon needed to know how thin the stainless steel walls of the rocket's tanks could actually be. Rather than asking company executives or relying on engineering documents, he went straight to the workers doing the welding. That is a rule he applies across every company: go as close to the source as possible for information. The people with their hands on the work know things the org chart does not. The welders told him they thought the tank walls could safely get as thin as 4.8 millimeters. Most managers would have written that down and moved on. Elon heard a boundary stated in confidence and immediately pressed on it. 'What about four?' he asked. The workers said that would make them nervous. 'Okay,' Elon said. 'Let's do four millimeters. Let's give it a try.' It worked. The assumption that 4.8 was the floor turned out to be a comfortable estimate, not a physical limit. This is the same instinct that drives his obsession with deletion and the idiot index. You cannot delete aggressively unless you actually know where the real limit is. Most teams operate well inside that limit because no one ever tested it. Elon's version of first principles is not just a thinking exercise. It is an empirical one. Stop thinking you have limits, he says. Stop giving into imaginary delays.

The play

The next time your team gives you a specification, a timeline, or a cost figure, ask who set it and whether it was ever actually tested. Pick one constraint you have accepted as fixed and push it by 15 to 20 percent. Ask the people doing the hands-on work, not the managers above them, what they think is the real boundary. Then test it. The limit is almost always further out than the number you have been given, and the only way to know is to run the experiment.

Tactic 12

Source Non-Aerospace Parts From Other Industries

When Elon started building rockets, he found that aerospace suppliers charged prices that bore almost no relationship to the actual cost of materials. He challenged those prices constantly, and his method was simple: compare the finished component to a similar part from an unrelated industry, then ask whether the gap made any sense. The latches NASA used cost $1,500 each. A SpaceX engineer modified a latch from a bathroom stall and built a locking mechanism that cost $30. When an engineer told Elon the air cooling system for the Falcon 9 would cost $3 million, he called over Gwynne Shotwell and asked what a home air conditioning unit cost. About $6,000, she said. The SpaceX team bought commercial air conditioning units, modified their pumps, and made them work. The $120,000 supplier quote is the clearest example. Elon looked at the part, decided it was no more complicated than a garage-door opener, and told one of his engineers to make it instead. The engineer built it for $5,000. That is a 24-to-1 ratio between what a supplier charged and what the part actually cost to produce. Elon had a name for that ratio: the idiot index. If a finished product costs far more than its raw materials, the manufacturing process is the problem, and the problem is solvable. Rockets had an extremely high idiot index. Elon calculated that the finished product, using existing manufacturing methods, cost at least 50 times more than the raw materials going into it. What makes this principle durable across companies is that it forces a specific question rather than a general posture about frugality. You are not just trying to spend less. You are calculating a ratio, identifying where the ratio is most absurd, and then targeting that gap directly. Elon pressed his teams to source components from non-aerospace companies as a standing practice. The logic was not that aerospace suppliers were dishonest. It was that any industry insulated from competition will let costs drift, and the idiot index tells you exactly where the drift has gone furthest.

The play

The next time a supplier quotes a price for a component, ask your engineer two questions: what are the raw materials in this part, and what does a comparable part cost in an unrelated industry? If the supplier price is more than five to ten times the raw material cost, treat that as a signal to build the part in house or source it elsewhere before accepting the quote. Assign someone to find the non-aerospace, non-specialized equivalent and price it out. The bathroom-stall latch and the home air conditioner both existed before SpaceX needed them. The savings were there; no one had gone looking.

Your action plan

the plays, one checklist.

01
Run the Algorithm
Pick one requirement no one questions. Find who made it. Ask if removing it breaks a law of physics. If not, delete it and run two weeks without it.
02
Use the Idiot Index to Kill Cost Creep
List five things you pay for. Estimate the raw cost underneath each. Divide. Any ratio above 10 is your first cost target. Assign an owner and a 30-day deadline.
03
Showmanship Is Salesmanship: Build the Demonstration First
Replace the slide that makes your central claim with the smallest working version of the thing itself. Put the other person inside the experience. A reaction closes faster than a deck.
04
Think In Burn Rate, Not Budget
Calculate your daily burn rate and post it. Run every spending decision against that number: a $90,000 plane rental is cheap if it saves a full day of progress. Time and money are the same ledger.
05
Isolate Yourself In Silent Darkness To Think
When stuck on a hard problem, lie flat in complete darkness and silence, earplugs in, sleep mask on, for at least 20 minutes. Do not try to solve anything or control your thoughts. Remove all stimuli and let your mind run. Solutions surface in silence that cannot be reached any other way.
06
Question, Delete, Simplify, Accelerate, Automate, In That Order
Pick one slow or expensive process. List every requirement with the name of the person who made it, then question each one. Delete as much as possible before simplifying what remains. Speed it up only after it is clean. Automate last, never first. If you do not add back at least ten percent of what you cut, you did not delete enough.
07
Walk To The Red
Set up a live display showing green or red status for every key stage in your operation. Walk the floor daily and go straight to every red. Ask what caused it, who owns it, and whether the requirement behind it needs to exist. Resolve it on the spot. Decision-making at the source of the problem is faster and more accurate than any report.
08
Combine Urgency With A Working Demonstration
Skip the presentation. Build a working prototype of your product's most impressive single capability and put your buyer inside the experience directly. Set the meeting date first, then build backward. The deadline is what forces the demonstration into existence.
09
Set Binary Survival Milestones And Repeat Them Until Everyone Knows
Do the math on survival. Find the one number. State it in one sentence. Repeat it to every executive and team lead until they finish the sentence for you, and post a live readout of progress where the people doing the work can see it every day.
10
Test Requirements Before You Trust Them
When a requirement can't be traced to a named person, test it empirically before defending it. Record with and without. Measure before and after. If the test shows no difference, delete it.
11
Find The Limit Through Empirical Testing
Pick one constraint your team treats as fixed. Go to the people doing the actual work and ask them where the real limit is. Then push past their first answer and test it. Most limits are comfortable estimates, not physical laws.
12
Source Non-Aerospace Parts From Other Industries
When a supplier quote arrives, calculate the ratio of finished-product price to raw material cost. If the gap is large, find the equivalent part from an unrelated industry or build it yourself before accepting the price. The savings already exist; the only question is whether you go looking for them.