Design like Elon Musk using 6 fundamental principles
How Musk’s companies are able to beat existing industry giants using 6 fundamental design principles.
This article describes six fundamentals of what is — in my opinion — the best way to develop new products. Like so many, I’ve been inspired by people like Steve Jobs or Elon Musk and have spent countless hours watching interviews, reading books and experimenting first-hand with product development in my own company, and continue to do so at the companies where I work.
While studying Applied Physics in 2011, I invested some of my student aid in Tesla. Easy to say now that I didn’t do it for the money, but it’s true. It had always just made so much sense to me to have electric transportation, and I wanted to support this company. As I watched some interviews with the now-famous Musk, I felt immediately that the approach was different, and so I invested.
Now I know that this feeling of something just absolutely making sense, is when you hit the jackpot. And this goes for your own products too.
Now this article is not about investing (well, indirectly it is), but about product development. I love product development, and while I continue to improve my own game, I just felt like sharing the principles that I use on a daily basis. They are based mostly on my perspective on how Elon Musk develops products, hence the title.
Fundamental principles
Developing the best products takes a certain way of thinking, but is in no way comparable to a recipe. Following a recipe lacks the required creativity for any great product to be made. Instead, I like to refer to these ways of thinking as principles. You still have to do the thinking, though.
The fundamental principles that I will discuss in this article are:
- Question the question
- Reason from first principles
- Kill your darlings
- Undesign
- Ideas supersede hierarchy
- Everyone is a chief engineer
Question the question
Contrary to what we’ve been taught in school, the right answer to any question is not always to answer the question, but rather to question the question. In other words: question the question first, question your answer next.
Let me put it this way: if you were asked which hammer works best for a specific screw, you could probably select a reasonably suitable hammer and make it work. However, the right answer would have been to first question if it wouldn’t be better to use a screwdriver… or a nail.
This example seems trivial, yet it happens all the time. Unfortunately, not applying this principle often leads to added complexity in a design, which has a rippling effect on cost, time, and quality.
The main objective is to question constraints you are given to solve a problem because it is most likely that these are to some degree wrong. Elon questions nearly everything, most notably about what can or cannot be done in terms of cost to space, as I will elaborate in the next section. Also see Elon talk about this here.
The best way to apply this principle is whenever someone asks you to solve a problem, to dig deeper into the constraints you are given. Spend time to see if they are asking the wrong question. You may see a solution that does not require solving the problem but preventing it in the first place. Therefore, the benefits of this principle are strongly amplified if everyone is a chief engineer as well.
Reason from first principles
For anyone who’s heard or seen interviews with Musk, he commonly mentions reasoning from first principles as the best approach for solving problems. The approach is a subset of the scientific method and typically used by physicists to acquire knowledge correctly.
The idea is to boil a problem down to chunks of which you know are fundamentally true. This sounds easy, but it’s surprisingly difficult. It requires digging deep into your assumptions, disregarding legacy, and earlier failures.
A great example that Elon himself uses, is to estimate the cost of a rocket launch. Pre-SpaceX era, rocket launches would cost hundreds of millions of dollars. Questioning the question would be to challenge this statement of cost in the first place. Reasoning from first principles means to figure out the fundamental cost of a rocket launch.
Say that for whatever reason you need a lot of launches. One approach would be to look at the cost per launch, multiply by the number of launches and get an estimate. A typical business person would probably negotiate the price of multiple launches from a rocket company. A typical corporate manager or purchaser could say something like “make every part of the rocket X percent cheaper”.
These approaches sound reasonable, but they all miss the most crucial question: what is the fundamental cost? When you do the math on the cost of the raw materials of a rocket (some metals, fuel and oxidizer, hydraulic fluids, and so on), you realize that the propellant is only a tiny fraction of the total costs, which determines the fundamental lower limit of costs per launch. Conclusion: you need a reusable rocket.
Instead of immediately focusing on how something could be solved (as in the examples earlier), you first and foremost focus on what you fundamentally need or want. Only after that is when you start figuring out how to get that done. That is Musk’s reasoning.
Kill your darlings
This phrase, attributed to William Faulkner or Arthur Quiller-Couch, is an amazing statement that helps past difficult design decisions time and again. Originally it was recommended to writers, whenever they had a great passage that — although great — did not fit or even weakened the full story. As a writer, you’ve “fallen in love” with what you wrote and are reluctant to say goodbye to it, even though it’s better for the whole.
A good example of Musk killing a darling was back in 2016, when Starship was still called Interplanetary Transport System. They were pretty proud of their carbon fiber tank, even sharing footage of pressure testing. Today, the idea of a separate tank is abandoned altogether, and instead the body is the tank. They realized the carbon tank just didn’t work out, and they killed it.
I’ve created darlings on occasion myself, and it takes effort to put aside your ego and realizing that whatever you came up with just doesn’t work in the bigger scheme of things. The takeaway is: don’t chase a foolish design.
In fact, Musk goes even further and states that:
“If a design is taking too long, the design is wrong”
This is a bold statement, but it makes sense if you think about it. When you combine this with the undesign principle, you may conclude that any design that is complicated, will take long to iron out all the details. Unfortunately, complexity sells better and companies trap themselves, falsely believing the solution must be complicated. Without killing this darling, they are also trapped in the timeline that comes with it.
The solution is to be willing to kill your designs when you know it just doesn’t make sense, technically or timely. Be careful not to fall into the sunk cost fallacy. Realize that you’re not starting over. You’re starting with experience. When you have a design that works, you’ll know.
Undesign
This is my personal favorite principle, although I prefer the term “simplify”. The principle entails to simplify any design as much as possible. This seems like an open door, but really it’s not. As famous mathematician E. W. Dijkstra said:
“Simplicity is a great virtue but it requires hard work to achieve it and education to appreciate it. And to make matters worse: complexity sells better.”
And it’s so true. A deceivingly simple design like a paperclip, screw-on lid, safety matches or even most of Apple products almost seem not designed at all. Now let me clarify, though. When I talk about design, I don’t just mean aesthetics. I mean all the intrinsic features that make the design perform its function. In other words, the way it works is the design.
The recent Starship design by SpaceX is an example of amazing design in my opinion. There are so many things to take into account for a complex system like this: aerodynamics, extreme temperature gradients, weight, pitch-yaw-roll control, reliability. The list goes on and and on, it’s mind-blowing.
Then you see the design, and may think: “Yeah cool, makes sense”. And that’s exactly the point. They managed to design one of the most complex machines ever built into something that doesn’t even look like it’s purposely designed.
If you’d like to know more about why exactly this design is such a big deal, I recommend checking out Everyday Astronaut.
So why go through all this trouble to make a design as simple as possible? Earlier I mentioned the rippling effect of complexity on time, cost, and quality. Complexity grows exponentially with the number of parts and the number of features per part. In short: you save time, cost, and quality issues.
It’s hard to overstate the benefits of undesigning, especially for volume production. The design intent should always be less, but better. Don’t spend time optimizing something that shouldn’t exist. Or, as Elon puts it:
“The best part is no part. The best process is no process. It weighs nothing. Costs nothing. Can’t go wrong.”
One important thing to realize when applying this principle is that complexity is easy, and simplicity is hard. Nevertheless, don’t settle for complexity. Your future self will thank you; it will just make sense.
Ideas supersede hierarchy
For any company doing research, development, and or production, it is important that the best ideas win and keep winning. Sometimes, this takes setting aside egos. We spend 99% of our time defending assumptions of which 99% are probably wrong.
Company veterans that “know everything” should be especially vigilant, as it’s easy to become complacent towards new ideas. I always try to listen to a voice in the back of my head asking “what if you’re wrong?”. And sometimes, the next day, I realize I was. Equally important then, is to say I was wrong and continue doing the right thing.
Although this is worth an article in itself, many companies seem to struggle with innovation. Or commonly: repeating past success with new products. One crucial part of innovation is that ideas must supersede hierarchy. Decisions should be based on the best ideas, regardless of their source. It can be the CEO, the intern, the janitor or a stranger on the internet.
In early phases of product development and idea generation, progress is often chaotic and unpredictable. It is important to provide objectives and limits while preventing obstructions in this process. Limits in terms of time and money are easy to set and understand. Objectives are more difficult, but generally doable.
Obstructions often come from unrelated or irrelevant legacy company processes and can be difficult to remove. Most obstructions have a hierarchical nature, like purchase approvals or mandatory reports. I might write a dedicated article on this, though for now it should suffice to say that hierarchical obstructions are detrimental for productivity and innovation.
Everyone is a chief engineer
To aid the principles of questioning the question and undesigning, it is extremely helpful for every engineer in the company to be a chief engineer. This means that everyone should at least have a basic understanding of the entire system, its constraints and interfaces.
“Cross-pollination of ideas is the main objective”
This seems like a burden to put on especially new employees, but it is exactly this system-wide understanding that helps reduce unintended subsystem optimization, while increasing contextual understanding. It is similar to reasoning from first principle, but on a system or product level.
In fact, product interfaces and errors often reflect organizational interfaces and errors. An example Elon himself mentions is the optimization of the Tesla Model S battery, where originally (in the first Model S generations), the battery pack was designed by one team, and the chassis by another.
This is an example of a subsystem optimization, where two local optima were found, but an overall better optimum was overlooked. However, during the recent Battery Day presentation, he indeed shows a consolidated design.
The analogy of a wing-shape is convenient because the same shift happened in airplanes years before: first, there was a wing, and a tank. Nowadays, the tanks are just wing-shaped.
In practice, company-wide system-level thinking increases the chance of a simpler design because there is increased cross-module interaction, therefore more implicit or even explicit peer-reviews, which in turn just cross-pollinates ideas which is the main objective here.
Note that it is extremely important for the ideas supersede hierarchy principle to be actively exercised in order for this to work properly. There should be no room for ego-based decision making.
Well, this turned out longer than intended. Please let me know in the comments if you have feedback, questions, additions or requests. Thanks!
