Free with a sign-in.
The best 14 min of 36
Jon McNeill: The Algorithm Behind Tesla and SpaceX | Podcast | In Good Company
Jon McNeill says companies should question assumptions, remove steps, perfect processes, build speed and automate only last.
In Good Company with Nicolai Tangen · Nicolai Tangen
- Jon McNeill is CEO and co-founder of DVx Ventures; author.
The best minutes
- 1:22Tesla cut online car orders from 300,000 configurations to twoPro pick 5 min
- 10:25Speed reveals broken processes before it improves themPro pick 6 min
- 15:47Tesla nearly went bankrupt by automating too early4 min
Every moment plays free. PRO marks a write-up that's in Pro.
Also worth watching
- 8:30Tesla targeted 30% gross margins against a 10% industry norm2 min
- 5:59Mapping the work reveals steps customers do not pay for2 min
- 20:24Ambitious targets force teams to rethink the operating system4 min
- 27:34Curiosity and financial levers guide McNeill's management calls3 min
Skip the factory-floor leadership stories, the remaining culture discussion, and the closing music and mentoring questions.
Full analysis · 6 min read
Real Alpha · Nov 8
Jon McNeill says Tesla cut over 300,000 combinations to two
- Tesla moved from over 300,000 online combinations to two models.
- Its near-bankruptcy account shows the cost of automating before testing a process in reality.
- For AI budgets, test manually first and track quality, cycle time, and cash impact.
Free with an account
Read the rest of the analysis
6 sections, the 5 other voices it tests McNeill against, and what it means for you. Every episode we cut gets one.
Check your inbox
We sent a link to your address. Click it to finish signing up. Not there in a minute? Look in spam.
or
Signed up before? The same buttons sign you in. The episode, its moments and the transcript stay free.
Tesla cut choice before scaling factory complexity
Tesla's online ordering moved from over 300,000 available combinations to two models, performance and long range. Buyers could still choose colors. That reduced choice where factory planning began.
The initial configurator took 64 clicks, McNeill said. He cited the e-commerce rule that more clicks generally reduce conversion, then asked whether Tesla needed true build-to-order. He said buyers clustered around performance or long-range cars, while color remained selectable.
For example, the change joined customer preference to manufacturability. McNeill said two models let Tesla scale production and size its supply chain more easily. The manufacturing head told him it would make his life 10x easier.
Yet the episode offers no conversion result showing the narrower menu sold more cars. McNeill is the author selling the operating framework in his book. His Tesla experience supplies operational detail, not an independent sales test.
“we had over 300,000 different combinations that we were trying to build as a first-stage manufacturer.”
Our read
The case supports testing whether every option serves buyers or just complicates production. It does not show that fewer options always raise sales.
Process maps exposed work customers do not fund
McNeill's next step is to map every process and substep. Managers then mark which work directly produces what customers buy. That reframes simplification as a customer test, not just a headcount or cost exercise.
He says about 90% of steps may be unnecessary. That is his broad estimate, not a measured result for each process. The map is a way to inspect the claim locally.
For instance, Tesla removed some intermediate quality checks and made workers responsible for passing on finished work. Downstream staff could return poor work, while the company measured those passbacks. That transfers detection rather than proving inspection has no value.
Tesla also combined battery and cabin cooling in one heat-pump system. Fewer systems meant fewer things to manufacture and one less potential failure point, McNeill said. Whether reliability improved is not quantified here.
Alex Danilowicz said on Lightcone in October 2026 that everyone at Magic Patterns handles support. Daily customer contact puts user friction before engineers. That backs the value of customer exposure, not the safety of removing checks.
Alex Danilowicz · Lightcone · Oct 22Daily customer support improves design without sacrificing speed
Our read
Process maps expose work that internal handoffs can disguise. But removal should be paired with a quality measure, as Tesla's passbacks were.
Manual tests should precede AI and automation
The sequence matters because speed and automation can lock in earlier choices. McNeill says teams should test a simplified process manually before optimizing or digitizing it.
He warns that AI can make a bad process faster and more expensive through token use. That is a specific cost warning, not a general case against AI.
To define great, McNeill says Tesla targeted gross margins around 30%, roughly Apple's level. He compared that with about 10% for the car industry. He says Tesla's gross margin stayed between 24% and 28% for years. That gave teams and investors a measurable bar, though the episode does not isolate which changes produced it.
Meerah Rajavel said this on The Information's TITV in October 2026. Palo Alto Networks used a harness to trace requirements and measure quality, security, speed, and efficiency. She reported three times as many new product introductions in six months, without adding software engineers. Those measures make output more legible than coding speed alone.
Woodson Martin said on The Cognitive Revolution in October 2026 that OutSystems' agent reads app telemetry and estimates time savings. It can build a first version for teams to validate. Both examples frame automation around measurement and review; neither proves every process deserves an agent.
“what we insist teams do is they manually run the process first”
Meerah Rajavel · The Information's TITV · Oct 16Palo Alto uses a harness to measure AI-assisted development
Woodson Martin · The Cognitive Revolution · Oct 7Agentic transformation starts with choosing what to automate
Our read
The budget test should be a changed workflow with a baseline, not an AI label. Measure quality and cost alongside cycle time before scaling.
Faster flow can free cash, but figures conflict
For Model 3 or Model Y, McNeill says the line began at 50 cars a week. He describes speeding it to 100, then 200, 500, and eventually 1,000 a week.
He says a speed target exposes downtime between steps, where work waits without moving. Repetition also builds worker and machine practice. That connects throughput with process learning, not simply pressure to hurry.
His cash-velocity comparison links production time to working capital. He first says about five days for Tesla, then corrects himself to 15; Toyota took four.
Fifteen divided by four is 3.75, by our calculation. That is a cycle-time ratio, not a working-capital ratio. McNeill says Toyota needed two and a half times less working capital, but gives no accounting basis. His stated 2.5 figure therefore does not reconcile directly with the durations.
Nik Storonsky said on Founders in October 2026 that easy funding can hide weak systems. He said Europe's constraints pushed Revolut to improve economics before scaling headcount. That supports attention to operating economics, but cannot verify Tesla's factory comparison.
Our read
Velocity is useful when tied to idle time and cash, not output alone. Here, the comparison raises a diligence question because its own figures conflict.
Model 3 exposed the cost of automating too early
Tesla's Model 3 line was designed digitally before a brick was laid. On the factory floor, McNeill saw machines six inches apart, needing hourly calibration by people with tools.
He says that line never entered production. Tesla had forecast cash flow that did not arrive, leaving the company close to bankruptcy.
Manual production was the recovery path in McNeill's account. Tesla built a tent outside and made 100 cars a week by hand, then 500. It later developed an automated line that reflected factory conditions.
For example, McNeill says DoorDash began with PDF menus and a phone number. Its founders picked up and paid for orders themselves while mapping the workflow. The example treats manual work as discovery, not a lasting rejection of automation.
This is one retrospective account from an executive who helped operate Tesla, not a quantified comparison of alternative designs. His story supports the sequence, but cannot establish its general return.
“We said, automate last.”
Our read
The failure makes the order of operations concrete. Test physical constraints before committing to automation that is costly to undo.
Stretch targets need learning and limits
McNeill says a 5% to 10% growth target gets teams that scale. By contrast, he says a 50% to 100% target forces a different operating formula.
He recalls teasing Elon Musk that 50% or 60% of a stated goal would satisfy him. Musk agreed, according to McNeill.
A stretch still needs a boundary. McNeill says goals must be close enough for people to try, not so extreme they give up. Their value is prompting a new plan and starting a learning loop.
The staffing model also has limits. McNeill described Tesla and SpaceX as special forces, deployed in short bursts, and said continuous intensity would burn people out. He names curiosity, bias to action, and intelligence as hiring signals. Culture matters too, because he says founder traits persist and can take a long time to change.
Will Bryk said on Knuckle Up with Nakul in August 2026 that Exa uses one- to two-day work trials. Candidates solve real creative problems, letting interviewers see how they handle tradeoffs. That complements McNeill's interview rubric, though neither episode reports comparative hiring outcomes.
Our read
Stretch targets can force teams to rethink a process, but sustained urgency is not a substitute for a workable system. Hiring and workload shape whether the learning loop can continue.
What it means for you
- Map customer-funded work. Map one workflow from request to delivery, mark steps the customer values, and record passbacks before removing checks.
- Baseline AI work. Run the proposed workflow manually first. Compare quality, security, cycle time, and cost with its automated version.
- Reconcile cash cycles. Track wait time and working capital together. Reconcile definitions when operating figures imply a different ratio than a speaker's comparison.
- Bound stretch goals. Set a target that changes the plan, then define review cadence and limits on high-intensity work. Treat misses as learning, not an excuse for permanent overwork.
What to watch
- Whether AI rollouts track quality and cost alongside speed.
- Whether cycle-time targets reduce idle work and working capital, not just increase output.
- Whether high-intensity work remains in short bursts rather than becoming continuous.
Send this to the COO setting automation priorities and the finance lead tracking cash.
Watch the minute where McNeill gives Tesla's count of over 300,000 combinations.
Written by Real Alpha from the episode's transcript and the other moments it links, then checked against the transcript. Quotes are the speaker's exact words.
Minute by minute
A Pro write-up, open to you here. Try Pro free for 7 days
A Pro write-up. The minute by minute is in Real Alpha Pro; the episode, its moments and the transcript stay free.
Read it with Pro, 7 days free- 0:00Tesla's production crisis taught automation last
McNeill says Tesla nearly went bankrupt when Model 3 cash flow fell short of forecasts. The team built a tent outside the factory and made cars by hand, first 100 a week and then 500 a week, until it could build an automated line that reflected reality.
- 1:22Question requirements before scaling
McNeill says Tesla’s online car-buying process had 64 clicks and more than 300,000 possible configurations. Data showed buyers mainly chose performance or long-range cars, so the team moved to two models with color choices. He also says a lawyer found none of the paragraphs in the 12-page financing documents were required by law or regulation.
- 5:50Delete steps customers do not value
McNeill says managers mapped processes with sticky notes and asked which steps customers paid for; he estimated that about 90% of steps might be unnecessary. Tesla removed quality checks between steps and measured work sent back for problems. He also says Tesla combined battery and cabin cooling in one heat-pump system.
- 8:20Simplify and test the process
McNeill says teams should run a simplified process manually before building a digital solution. He argues that speed reveals faults and that a strong process can be good, fast and cheap. Tesla aimed for gross margins of about 30%, matching Apple’s target, compared with about 10% for the car industry; Tesla’s margins stayed between 24% and 28% for years.
- 10:25Speed exposes process bottlenecks
McNeill says speeding up a flawed process only gets to a bad answer faster, and applying AI can add token costs. He describes raising Model 3 and Model Y production targets from 50 cars a week to 100, 200, 500 and 1,000, using speed goals to expose downtime. He contrasts Tesla’s 15 days from aluminum to finished car with Toyota’s four days, and says Tesla used a weekly heartbeat to track whether it would make its quarter.
- 15:47Automate only after learning manually
McNeill says Tesla designed its Model 3 production line digitally before building the factory, but machines that needed hourly calibration were installed only six inches apart; that line never went into production. He says the founders of DoorDash first took orders from menu PDFs and handled deliveries themselves to learn the workflow. He argues that adding AI to cumbersome processes before fixing them can speed up failure and expense.
- 19:56Ambitious goals drive innovation
McNeill says Musk drove weekly improvements that accumulated into major breakthroughs, and argues that innovation often requires control of key parts of production. He cites robot hands as an example: when needed actuators do not exist, a company may have to make them. He says targets of 50% to 100% growth force teams to rethink their methods, but should still feel within reach.
- 24:26High-intensity teams need recovery
McNeill says people at Tesla and SpaceX respond to difficult goals with humility about what they do not know and confidence that they can learn. He describes intense work in short bursts, with time to recover, and says teams must learn quickly from failed tests. In hiring, he looks for curiosity, a bias to action and intelligence.
- 28:46Find the financial levers first
McNeill says he starts with a company’s financial model, identifies two, three or four key levers, and looks for ways to double, triple or quadruple one. He says a culture that accepts the status quo and lacks curiosity or a bias to action can signal a business may be hard to repair. He adds that founder-shaped culture can persist for decades, citing Arthur Sloan’s lasting imprint on General Motors.
- 31:00Stay close to customers and mentors
McNeill says his biggest business mistakes came when he stopped using the product; he recommends spending time with customers and frontline employees, who often point to the same few fixes. He advises young people to find both experienced mentors and peers facing similar challenges. He also says music helped him understand advanced math and engineering.
Published Sep 16, 2026.
Transcript · 28 min read
In Good Company with Nicolai Tangen, hosted by Nicolai Tangen, with Jon McNeill.
Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.
We almost went bankrupt because we didn't have the cash flow that we had predicted coming off a Model 3. And the only way we saved ourselves was to go back to the manual process. We literally built a tent in the factory outside and produced cars by hand, first 100 a week, then 500 a week, etc., until we could actually build an automated line that reflected reality. When we looked back on that and said, look, we almost killed the company, what would we do differently? We said automate last. You've got to perfect the process before you automate, otherwise you just might bury yourself. And in that case, we almost did.
Hi everyone, I'm Nicolai Tangen, the CEO of the Norwegian Samhain Wealth Fund. And today I'm in really good company with John McNeil. Now, John spent three years reporting directly to Elon Musk as president of Tesla, then was Chief Operating Officer of Lyft, and now he builds companies at DVX Ventures. Lots of people have worked for Elon, but John actually wrote a book about how to do things, the algorithms, which basically lays out how Musk used to build Tesla and SpaceX. And so today, we are going to go through this, you know, point by point, so that you also can sort out your business and make it a huge success.
So, John, big thank you for coming on here. It's an honor to be on with you, Nikolai.
Read the whole transcript
“we had over 300,000 different combinations that we were trying to build as a first-stage manufacturer.”
proven otherwise yeah okay step number two in your book delete every step yeah so how does that what
“if it's not of high quality you can pass it back and we're measuring the passbacks”
Rule 3, simplify. Skur, you must be impressed that I read the book and I know it by heart, John.
Yes, this is wonderful.
“We set a goal for ourselves of gross margins equal to Apple.”
“if you speed up a bad process, you're just getting to the bad answer faster.”
“We literally built a tent in the factory outside and produced cars by hand, first 100 a week, then 500 a week”
What you describe in the book helps to speed things up, produce faster and cheaper. Does it help innovate? Yeah, that was really the point of the model, is that we use the algorithm to drive innovation and not incremental change, but quantum change. And this is really what's behind the kinds of innovation you see coming out of Tesla or SpaceX.
“When you set a goal of 50% to 100% growth, people have to rethink entirely how they're doing that.”
How much fear was there in the organization? I would say that the most common trait of people at SpaceX or Tesla is humility, believe it or not. And there's not much fear. And when Elon lays out a goal, the most common response is a response of humility and confidence at the same time. That sounds weird, but let me break that down.
The first response is a response of humility. I have no idea how to do that. I have no idea how to achieve that. The second response is, but we'll go figure it out. And that's this response of confidence that world-class people tend to have. I don't know how to do this, but let's go figure it out. And we're going to chip away, chip away, chip away at this problem until we figure it out. I think that's mindset number one. Mindset number two is just being ready. You're going to have a lot of failed launches and failed tests before you get to that final goal. So you've got to be able to absorb the failures, rapidly learn from them, not repeat them.
But again, the progress you're making is compounding against competitors because your competitors typically are too fearful to start that journey. So just starting gives you an edge in the race. How many nights did you sleep on the factory floor? I spent probably weeks worth of nights on the factory floor for the launch of Model X, for the launch of Model 3. And that was because we wanted to show that we were in the problem with the people and that the problems that were happening at the edge in manufacturing mattered. It mattered to cash flow. And we also considered ourselves teachers of this methodology.
So if something was so critical, we were trying to get the cash flow from a particular product and it wasn't coming, we wanted to show folks, here's the method and the formula and the framework you can use to break through these problems. And we'll be with you in the trenches as we do this. And that was part of the leadership model. Did your wife think it was a good idea that you stepped on the factory? She did not. In fact, she would tease me and say, I'm pretty sure the executives at GM or Toyota are not sleeping on the floor. How sustainable is this model? How do you burn out people? It's not sustainable in the sense that when we hired people, I would tell people, you are joining special forces, not regular army.
And here's the difference in those two models. Number one is special forces aren't deployed continuously. They're deployed in short bursts of time for key missions. And so there will be some short bursts of time where you're sleeping on the factory floor. But those are short bursts of time. But the trade-off for that is you're going to be working in a platoon with the best of the best. You're going to do the best work of your life. You're going to be thrilled when you break through these problems. But most nights, if you came into the office at 7 p.m., you could roll a bowling ball through this office and not hit anybody because we are a special forces model.
And we're training daily from 8 to 6. It's an intense environment. and there will be these periods of intensity, but they're only periods. Because to your point, Nikolai, you'd burn people out if that was continuous.
“curiosity, bias to action, and intelligence, it's a pretty good bet that they were really going to thrive at a place like SpaceX or Tesla.”
Can you not change a culture? I think changing a culture can happen over time, but it takes, in my experience, long periods of time to change a culture. Why is it so slow? Why is it so difficult? It's embedded in the DNA of the business. And most culture comes from founders. I've even seen at General Motors, a 150-year-old company, that the DNA that Arthur Sloan put into that company still exists.
And that DNA is not only injected into the company by the founder, it's perpetuated over time and really becomes very difficult to shake. So changing culture is not for the faint of heart, nor is it for the people who are short on time. So interesting. Sometimes I ask people to define the corporate culture, and they come up with some defining characteristics. And then you say, but hey, these are just your personality traits, and that's exactly what they are. Exactly. Yes, they're personality traits of the founder, typically of the leader. John, what is the biggest mistake you've made in your life?
And boy, where do we start? I think if we limit it to business mistakes, I would say eating my own dog food. There have been businesses where I have gotten lazy and not sampled the product every day. And it wasn't until I read Sam Walton's book made in America that I started to appreciate sampling your own product on a daily basis. And one of the things that Sam says in that book is he famously would call his customer service telephone number every day on his way into work to understand how customers were being greeted and treated. And the biggest mistakes I've made is when I've stepped away from the product and I haven't experienced what the customer is actually going through.
Using that product, whether they're experiencing frustration or joy. And so I would say my biggest failures have come from that, from being disconnected. So when you're on the board of Lululemon, you run around in Lululemon stuff? I do. I'm wearing Lululemon right now. I'm going to a Lululemon store later today. I drive General Motors cars. Every day I drove a Tesla off the line because I wanted to have that experience. And I would tell Elon, like, I've got a 20% rule that's a little different than like a Google 20% rule, where you can 20% of the time work on whatever you want. I told him that 20% of my time is going to be in the front lines because I want to experience what the customer is experiencing and what our employees are experiencing.
And our frontline employees, it's just one of the best hacks in management. Frontline employees can tell you exactly what's wrong with the product because they're hearing it from customers all day long. They can tell you exactly what the customers want in the product. And oftentimes they give you a very quick cheat sheet as a manager to go make really effective change. Because when you're out on the front lines asking people, what would you do if you had my job? You don't typically get 500 different answers. You typically get like three or four of the same answers over and over again. And it's very revealing.
And so that's helped me avoid a lot of mistakes. When I ran my own company, I called the switchboard every day to make sure it was picked up on ring number one. Yes. Yep. Very Sam Walton-esque. What is something about you that most people don't know? I have just an absolute admiration, joy, and appreciation of music of almost all types. I was lucky to have a musical mother who put this love into me early on. And I almost went to university to study music. I loved it so much. But what I discovered when I got to university and got into engineering was music had taught me a base eight math system.
And it was incredibly, incredibly helpful to me, understanding advanced math and engineering. So I'm grateful to it. But a lot of people don't know that I was almost a music major. So when you see a production line, do you think about it like a symphony? A little bit, yeah. When I see robots welding, you know, 300 robots welding a chassis, it does look like a symphony a bit to me. John, we have a lot of young listeners. What is your advice to young people? Grab a mentor as soon as you can. And two types of mentors. There's vertical mentors, people that are ahead of you in the journey, maybe by a generation.
They can provide wisdom and experience. And then what I call horizontal mentors. Grab people that are high potential, that are in the similar situation that you are. Maybe you're a first-time product manager. Maybe you're a first-time CEO. Grab five or six first-time CEOs who are non-competitive in a business similar to yours, size, maybe growth. Get together with them once a quarter and have a session where it's complete Chatham House rules and you bring your biggest problem to that group and learn from them, but also have mentors in place that have been in that journey before who can say, I've seen that movie, let me help you get through it.
I wouldn't be where I am today without a half dozen mentors, some of whom I mention in the book, all of whom I mention in the acknowledgements of the book because they've made me who I am. Really good piece of advice. Big thank you for being here today. Once again, love your book and thanks for sharing your secrets. Thank you, Nikolai. It's been a pleasure talking with you. Likewise.
