Auren Hoffman, SafeGraph CEO. fmr LiveRamp CEO. Started & sold 5 companies
Answered Sep 16
https://www.quora.com/What-is-first-principles-thinking
There are tons of people that claim you need to use first principles for all things. People I follow and greatly respect (Naval Ravikant, Shane Parrish, Julia Galef, Eric Weinstein, Scott Alexander, Peter Thiel, Elon Musk, etc.) regular promote first-principle thinking.
The problem is that you cannot use first principles to determine everything. You don’t have the time to do that. You need to rely on proxies who you believed have figured things out and believe in them (until you eventually figure out that the proxies are wrong, frauds, etc.).
For instance, I have never actually done the full proof that the world is round. I don’t actually know, with 100% certainty, the shape of the earth. I use proxies to help me determine that. It might not be round. There might be a conspiracy. Or we might be living inside a simulation. I’m not 100% sure. But I rely on proxies and make an assumption that the world is round (at least for my purposes).
I don’t know (with certainty) that the moon landing in 1969 was real. Some people believe it was faked. But I use proxies who I respect and therefore adopt the belief that the moon landing was real. I believe this even though I have not taken the 100+ hours to prove it myself.
Therefore, I believe the world is round and also believe the moon landing was real. Am I 100% certain? No. But I live life believing it and know that I will likely never take the time to prove either to myself.
So when should one go to first principles?
Yes, you do not have time to go to first principles on everything. But you cannot just blindly follow the “experts” on everything — because the experts are not always right. So when do you dive in on the data yourself?
One heuristic you could use is to go to first principles when the decision REALLY matters to you.
Your life would not be much different if it turned out the world was flat, so you might as well believe that the world is round.
But what about your health? Or your child’s health?
But here too, going to first principles is very costly. Should you eat meat? Should you completely cut out sugar? These questions might take thousands of hours for you to thoroughly research via first principles.
Of course, if it only took one hour to go to first principles figure out an important decision, it would be worth it.
I originally misspelled this post “first principals” but then the grammar checker caught my mistake and I was able to quickly rename it “first principles.” That was a good use of time to get to the truth. Took about 3 seconds.
But what happens if something takes 10 hours, 100 hours, 1000 hours, or even 10,000 to “know” — what should you do?
The answer (and this is incredible unsatisfying) is, “I don’t know.” But I will try to give some guidance:
Most people’s problem is not they try to go to first principles on TOO MANY things. Yes, a few people overdo it. But most people go to first principles on too few things. In fact, most people go to first principles on NOTHING. They rely on proxies for EVERYTHING.
The first rule of thinking: at any given point in time, you should be working on something (at least one thing) via first principles thinking. It might take you a really long time to think through. It might take you the rest of your life. But having at least one thing to go deep on is a life well-lived.
The second rule of thinking: while you are likely going to rely on proxies for over 99% of what you believe, it is really important to vary your proxies. Relying on only a handful of proxies to tell you how to think is dangerous.
By increasing the number of proxies, you are more likely that the proxies end up disagreeing with each other on important truths.
And this comes to the third rule of thinking: make sure that many of the proxies that you use conflict with each other at least some of the time. If all the proxies that you use agree with each other, you might never have the opportunity to question your assumptions. And some (but likely less than 25%) of your proxies should be heretics in their field.