I’m working on a startup in MIT Sandbox and signed up for an expert session recently. As we were nearing the end of the call, I asked about the unknown unknowns when starting a company as these are risks that one has to be mitigate. Put another way, you don’t know what you don’t know and thus can be easily caught off guard. He mentioned three points, of which the one I want to dive deep on here is the hypothesis driven mindset. This got me thinking about how I think once again. I thought about how a given problem that one needs to solve can be approached from two angles: a hypothesis driven mindset and first principles thinking. Some might call this inductive reasoning and deductive reasoning, or bottom-up and top-down. I personally don’t like those definitions - you have to memorize those definitions and what’s bottom-up to you may be top-down to someone else. I have been looking into some finance applications recently and thus will structure the examples around finance.

Hypothesis Driven Mindset

When studying for a class, it is often useful to do many practice questions - the question is sort of the hypothesis in this case. You are given a question and have to dive deep into the textbooks, the lectures, the notes, and figure out principles that can help you solve the problem. In the context of finance, one question you might ask yourself would be: What trades can I make this week? Well, that’s a really broad question, and those who have done management consulting interviews would be familiar with these types of questions and you have to “drill down” and not “boil the ocean” (I promise this is only used once in this post!). So you might start drawing a tree and asking yourself lots of questions. You think about what might affect trades this week - earnings announcements, federal reserve meetings, inflation data being released, job market reports, etc. You think about what kind of risk you’re willing to take. You think about any downside protections you might want to have. The list goes on. Eventually, you arrive at a trade that you can make and you go through that strategy thoroughly before making that trade.

First Principles Thinking

I learnt first principles thinking from my mother as she taught that to me from a young age. I’m not sure what the prevailing definition of it is now - I am not here to offer one in any case. Again, in the context of finance, a first principles question you might ask yourself would be: What return distribution does a given stock follow? I don’t actually know the answer to this, and I can imagine the answer is really complex. I have brief exposure to this, and I do know that there was some random walk paper a long time ago (40, 50 years maybe?). Ultimately, it’s a really hard question and the answer is probably “it depends”. In some contexts, it might be a random walk. It’s also probably a nonstationary process. The list of technical terms you can use goes on and on. The value of answering a question like that, is that you get to understand the underlying processes of something deeply.

Application to Life in General

As you would probably have guessed, you can’t just stick to one mindset and apply it religiously. You have to use a mix of both. First principles thinking enables you to acquire a set of tools. A hypothesis driven mindset gives you opportunities to explore questions and potentially leverage these tools. The important catch here is to understand the tools you have in your toolbox. When you’re in the moment and deep in a problem, you might forget that you have a certain tool in your toolbox. For example, when you are deep into coding something from scratch, you might forget that you saw a framework or codebase 6 months ago that could be helpful. Or when you’re cooking something on the stove, you grab the nearest tool that does the job, but there might be a better tool in the drawer that you forgot was there. Or you may realize that you don’t have the tool and you need to acquire it! It can be useful to take a step back sometimes and try and meet in the middle through both forms of thinking.

While these are tools for thinking, I missed out the most important motivating factor - what are you doing this for? If you’re cooking, it’s probably to have some delicious food. If you’re coding something, it’s probably to build good software. If you’re writing a paper, it’s probably to propose a new method and publish a paper. The framework of thinking helps you to achieve a goal. I intentionally did not put this at the start, as that’s probably implicit and only worth mentioning at this point.


I’ve been writing more on thinking about how to think recently. I already somewhat apply this to my research, but I will be doing this more intentionally. Specifically, I start with a hypothesis: Can I use X to solve Y? I then dive deeper into the relevant papers, and try and confirm or disprove this hypothesis. At the same time, while diving into the process, I try and recall the mathematical tools that I have learnt, and also how to implement the code. More importantly, I need to think about how to structure the paper, run the experiments and present the results such that the paper will have a higher chance of getting accepted. I think it can be useful to start with the goal and draw a mind map from there. Perhaps I will try this.