Working: Fast and Slow
Let’s say you need to plow a large field, and for the task, you are given a machine. How would you plan your work to be as effective as possible? I assume the first questions you want to ask is — What is this machine I have? How does it work? What can it do? I think these are great questions, ones that we should ask ourselves whenever we have a project to do, for which we are going to use the machine we call our brains.
This post will try to use cognitive science insights to create a methodology for tackling problems that require a lot of thinking.
In the book Thinking, Fast and Slow, Daniel Kahneman argues that people have two “systems” of thought which we alternate in our daily lives:
- System 1 (the fast one): It’s when you instinctively know the answer to something or have an emotional response.
- System 2 (the slow one): It’s when you deliberate something with yourself logically and rigorously, and intentionally focus your attention on it.
Each system is better at different things and prone to different kinds of problems. Knowing this, and choosing which system you should use for different tasks, can be very beneficial.
System 1 is much more prone to cognitive biases, which could lead us on the wrong path or expose us to be exploited. For example, the availability heuristic: when we try to approximate the probability of something, we instead evaluate “how easily do I recall this thing happening”. This is clearly inaccurate and can easily be exploited: think of a car insurance salesperson who offers people insurance a day after they crashed their car.
Not that system 2 is as foolproof as we may hope. Some biases only seem to be relevant to System 1 but are relevant to both. My favorite example is Confirmation Bias, which is when people know that they’re right, and automatically (but allegedly logically) reject all data that conflicts with their initial opinion, and embrace every datum that supports it. For a prime example of how this operates see: Politics.
So how should we choose which system to use?
Some 30 years before Kahneman’s book, the Dreyfus Model of Skill Acquisition was published. According to this model, the more experience you have at a skill, the more you can rely on intuition (in our terms: System 1) because your intuition has already integrated so much knowledge, and implements a holistic view on the topic. Until then, you need to use analytical thinking, and decompose the topic into small chunks (or: use System 2). You do that by watching others, working methodically, and thinking about every step of the way.
I argue that working on a new project that has a lot of unknowns is a lot like learning a new skill. At first, you’re wide-eyed, trying to absorb information, and need a lot of guidance, but over time you become something of an expert in this (very limited) domain, and at that point, you have a pretty solid intuition regarding the solutions to the problem at hand.
Under these circumstances, we might be able to speed up the process of becoming experts, by leveraging what we know about the two systems. What’s following is a system I use that I think does just that. I call it Working, Fast and Slow.
Working, Fast and Slow, in 3 steps:
- Work fast, using System 1. We might not be experts yet, but we pretend that we are and try working on the problem holistically — as if we’re going to solve all the problems at once. This can mean different things for different projects, but the important thing is looking at the big picture and not the fine details.
The outcome of this step can be a document, a presentation, or a diagram with a full solution for the problem, even if it’s a bad one. Work on a best-effort basis: try to have the full thing ready in a few hours or so. If you get stuck or feel that you really don’t know what you’re doing, don’t worry. Just save whatever you have — you’ll be able to fix it in the next steps. - Work slow, using System 2. Here you want to take some time to reflect on your plan. Think about the problem from different angles, get opinions from peers, read about similar problems, etc. Sleep on it for at least one night. You’ll probably start to see the plan you’ve made in (1) fall apart, if it didn’t already, at least if it’s the first time around, because the next step is:
- Repeat. Re-read your plan after the slow work phase. If it sounds perfect, you’re done. If not, just put it aside, go back to (1) and create a new document, presentation, diagram, or whatever you’ve made, using the previous one as a reference.
This method will hopefully utilize both systems for what they’re good at. In (1) you are exercising your intuition, letting it be creative and, possibly, fail. In (2) you are methodological, you focus your attention and learn rigorously. The systems feed each other. System 1 will work better after each iteration because you know more, And System 2 will work better because it will have something better to work with.
I’ve had projects which took 5 or more iteration to get to the final result, including system designs, presentations, and even blog posts (wink). Over time I got used to labeling the first one as “-v0”, then “-v1” etc. I usually only use the latest version as a reference, but sometimes I use the previous ones as well.
I like to think of this process as making my plans “anti-fragile” —Creating plans and treating them as if they were final makes it easier to find the faults in them. After a few iterations of improvements, while keeping the advantages of the previous versions, I’m able to come up with something pretty resilient. (More about anti-fragility in another post)
All this might sound like a lot of work. I think it’s worth it: At the end of the process, you not only have a plan, but you feel like an expert on the problem.
There’s a saying that goes:
“An expert is someone who has made all the mistakes which can be made, in a narrow field” / Niels Bohr
Which is a good indication of what this method does. After a few iterations of pretending to know what you’re doing, you may not have made “all the mistakes which can be made”, but you came a long way in this direction.
