Imagine you have a really tough problem to solve in your life. It’s a little overwhelming, and feels next to impossible to overcome. Google isn’t giving any good advice either. But you adhere to the motivational advice that says to keep failing, never give up, establish goals, and whatnot. You even get that there may be a time that you should give up and try something else. Just not yet, you aren’t ready to give up. But the traditional motivational advice isn’t getting you anywhere. You’re working harder than everyone and aren’t seeing any real progress. How do you find a way forward?
Building a Spaceship
We know that spaceships can be built. Rocket scientists exist to do just that. Although it’s a distant and vast goal to build a spaceship, we at least know it can be done. If we were to start on a journey to build one, the steps to get started would probably be overwhelming for the majority of us. We’d need to know all about propulsion, gravitational forces, material sciences, astrophysics, and much, much, more. Eventually we’d need the resources to have a huge team, work spaces, labs and materials. The steps involved to make the goal a reality would be massive.
However, no matter how massive the goal, we know it can always be broken into smaller and smaller pieces. At some point, the pieces would be small enough for even a young child to understand and learn about. The idea of breaking concepts down into their most basic form is first principles. Elon Musk, who happens to have a company building spaceships, is a big proponent of this way of thinking. Whenever we are stuck, if we aim to break our problems into first principles, we should see a way through. That is great advice in general, but it’s only part of the answer.
What Is a First Principle?
In the scientific and mathematical fields, first principles refer to basic truths: Areas that can’t or don’t need to be broken into further components. We base the foundation of all of our knowledge and understanding on these first principles. Without these facts as a common frame of reference, we wouldn’t be able to create much of anything.
For example, if we use first principles in conjunction with deductive reasoning we can reach consistently valid conclusions. Those simple word problems we used to hate as a kid are examples of this type of reasoning. For example, we have a rocket that costs $50 million (premise one). Our budget is $60 million (premise two). Conclusion: If we purchased the rocket, we would have $10 million left.
If at any point, the premise is fuzzy or not necessarily 100% true, we can break the premise down into smaller pieces. For example, why does the rocket cost $50 million? How much are the individual materials? How much is the labor? How much time does it take to put it together? This type of questioning is what we are referring to when we speak of first principles. It’s about questioning the premises until we can’t any more.
First Problem with First Principles
This brings us to the first issue. It’s not logically plausible to base everything off of only 100% true information, especially not in everyday life. Asking why is great, but those questions can also lead us down a never ending path in the wrong direction. We could get stuck, trying to delve into the materials of the rocket, when instead the real win was the reusable factor. Life has a never-ending number of things that need to be analyzed, but it’s just too much to analyze it all.
Ultimately, it may be true that, if we can find first principles, we can answer and solve any problem. However, we have to be sure we are solving the right problems. We would need to have a way of determining when we should or shouldn’t go deeper down any line of questioning.
Micro and Macro
First principles give us an ability to break things into a very specific micro-perspective. However, it’s a misleading assumption that breaking things down into smaller pieces is the best way to figure out the next step, to innovate, or find a new solution. In reality, we don’t always need or want to break things down further. We may need to see a bigger picture first before continuing a particular line of questioning.
In our rocket scenario, we question why the rocket costs $50 million, but it takes a different line of questioning to entertain the possibility of reusing the rocket. On one hand, finding the solution as to how to reuse the rocket requires first principle thinking. Yet, to ask the question and identify it as a worthwhile endeavor is a form of big-picture thinking that sees possible key solutions in a sea of information that hasn’t been proven or isn’t necessarily true.
It seems that to be able to find a solution in a sea of information, to have a macro-perspective takes an ability to recognize patterns, while at the same time, choosing appropriate areas to delve deeper into. With first principle thinking and deductive reasoning, there are tangible and explicit links to every next step. There aren’t any cognitive leaps or assumptions.
When we take a macro approach, we are making cognitive leaps; we are making assumptions. Generally, this is more of an inductive process in that we are taking in patterns and drawing a conclusion based on those patterns.
What this describes, however, is how we all already think every day. One of the possible reasons Elon underscores first principles is because making assumptions and drawing conclusions from those assumptions is easy, and we all do too much of it. As such, a form of going back and forth between a macro view and first principles is needed.
It’s easy to target our assumptions or premises and break them into more fundamental first principles. Similarly, it’s easy to recognize patterns when all the information is there, laid out for us to see. However, the most difficult problems exist not because we make bad assumptions or that we miss patterns. The most difficult problems are those we have no data or experiences for.
For example, to find a pattern, we need to see several iterations before it can be identified. If we’re asked what comes after the sequence: 2, 4, 6, 8, we’d easily say 10. But what if we could only see the numbers 2 and 8? We’d have no clue. We wouldn’t have enough information, even though the pattern is simple.
Fill In the Blank
Generally this means that to solve problems, we have to smartly fill in the blanks. We zoom out to find possible missing pieces of the puzzle. Then, when we find them, we zoom in using first principles to fill the blanks. If, at some point, it doesn’t seem as if the newly filled-in blank will yield a positive solution, then we zoom out again and search for another missing piece.
For example, in our rocket scenario, we know we want to save money on rocket costs some way. The original pattern was that a rocket costs $50 million and that every launch destroys the rocket. If there was no evidence of any other patterns, there would be no immediate reason to conclude otherwise. This is a blank space, a hole that could form a pattern which doesn’t currently exist in our own understanding. If we look at existing patterns, we can always imagine what an opposite or relating pattern could look like, even though we don’t have the data.
We always have to start from somewhere. Something has to be our catalyst to create or imagine a new pattern. In the vast majority of our thought (possibly all thought), we have to base things off of either existing truths or existing experiences. For instance, maybe we just need to extract an existing pattern farther out. Going back to numbers for a second, the pattern 2, 4, 6, 8 firstly has two directions it could go: -4, -2, 0… as well as 10, 12, 14. Also, without seeing any other pattern, we could assume there’s a pattern 100, 102, 104 as well.
What if we altered the pattern artificially? We could take the opposite: -2, -4, -6, -8. We could multiply by a factor: 6, 12, 18, 24. We can do just about anything to existing patterns to find new possible patterns.
Now that sounds easy with respect to numbers, but what if we imagine these numbers as concepts or as information that we are putting together? For example, what is the opposite of destroying a rocket every launch? What are the effects of increasing or decreasing the size of the rocket, or scaling the processes, resources? This all is common sense to some, but not all of us are intuitively adept at seeing possibilities. Notice that this way of playing with patterns isn’t first principles. We’re finding possible patterns from existing patterns, without regard to whether they are first principles or not.
Problem Solving Process
Let’s put it all together. Generally speaking, everyone knows how to solve problems. We identify a problem, figure out all the information relating to the problem, and then draw conclusions about possible solutions.
From the start, we could be solving the wrong problem, so the first thing is to figure out not only the assumed problem but any related problems. In the example of our rocket, the assumed problem was that we need to save money on the rocket. But how did that problem even become a problem? Was it due to an immediate budget need? Or was it actually related to being affordable for smaller investors? Flushing out those details by prioritizing and understanding the why behind the problem is step one.
Next, we have to capture all the information that relates to the problem. We figure out everything we can at the same micro or macro level of the problem itself. For example, if our problem is about budget, we’d want to know what is driving our budget: why it is or isn’t a particular size. What impacts the budget the most, the least, etc.? We’d be looking at all the information relative to the significance of the problem first.
Then, we can start to identify areas that are assumptions, areas that are more factual, and areas that have blanks or no evidence or data. Based on all this information, we can identify possible solutions and areas that we need to look into further. The cycle would continue back and forth between smaller and smaller problem sets to bigger and bigger problem sets. We’d zoom in and out, resolving blanks, tweaking our problem sets, and identifying solutions until we find an optimal path.
Today’s Problem Solving
The issue with problem solving seems to be our incessant reliance on Google, YouTube, and DIY blogs. Research and problem solving are now relegated to searching and following someone else’s patterns. It’s the easy way to solve most day to day issues. But there are still some things that you simply can’t google. There are things that have no consistent checklist, no how-tos.
We have to approach our problems with patience and discipline. Problem solving can be tough and time consuming. It requires trying things and failing over and over again. It’s likely that some of us are much more intuitively adept at creating and finding new possible patterns, whereas others may be more adept at deep diving and first principles. Find out which one you are and be sure to account for the other perspective by either working on it or partnering with people who can balance things out.
In all cases, it will take keeping track of lots of information and zooming in and out without getting too fixated on any one issue. If possible, it’s best to find some aspect of enjoyment out of working toward a solution. There may come a time in the near future that our ability to solve problems is the only thing that keeps us employed by the AI that will eventually take our jobs… (Just kidding…a little.)