Done projects are better than perfect projects

I started a lot of projects in high school that I didn't end up finishing:

• An app to make personal science (think quantified self) much easier. I actually attempted something like this three or four times over middle + high school, but never finished an attempt.
• A roguelike game that you play with a MIDI keyboard I was working on the summer after high school. Made a fair amount of initial headway, but the project ballooned and I abandoned it after a month.
• Learning piano and guitar. Started and learned rapidly at the beginning, but as the surface area of possible skills to pick up increased (sight-reading, playing by ear, improvisation, arranging), I attempted to master all at once and made frustratingly slow progress along any dimension.
• Initial attempts at starting a blog. I would sit down to write and start getting my thoughts out, but always feel like what I had was kinda sucky and not worth putting on the internet. This led to writing feeling very hard, and something I had to force myself to do.

While at the time I often considered abandoning projects to be due to a lack of discipline, I now think this is the wrong framing. I started these projects because I was very excited, and I initially had a ton of fun working on them.

Here's a more accurate narrative:

1. I had some vision in my head of how I wanted the project to end up, and my excitement led to me adding on more and more requirements ("wouldn't it be cool if X?").
2. Because of this, and the inevitable reality that there are many more problems to solve than first envisioned, the distance between the current state of the project and the vision seemed only to grow.
3. This wore down my initial excitement. Working on the project shifted from being the default action to being something that I had to push myself to do.
4. After enough time, I abandoned the project to move on to some other idea that I became excited about (for which I'm blissfully unaware of the impending scope creep).
5. Once this cycle had happened enough times, I began to lose confidence that I would follow through on a project more quickly.

This is super suboptimal!

When this happens, it's a big source of frustration for me, and yet in the past I've conceptualized this as a lack of discipline. I thought the issue was myself being lazy and not being able to stick with projects long enough to finish them. I now think this is silly! My decrease in motivation to work on these projects arose as a symptom of the underlying problem: not getting closer to the goal due to moving the goalposts and moving in too many directions.

But it doesn't have to be like this! I've finished plenty of things (see if you can notice a pattern):

• I made websites for several local businesses. Notably, these gigs came with deadlines!
• All of my open-ended final projects. These always get done, and even though there are always things I wish I could improve, I tend to learn a lot from them.
• I've finished videogames and apps for hackathons, as well as ones for which I imposed an artificial deadline and deliberately had lower standards (emulating a hackathon).
• ATA, an edtech project I did with some friends. Notably, I was expected by my friends and our users to finish features by a particular date.
• I've published 13 blog posts over the past few months (~1 a week). I'm using beeminder to keep me accountable (if I don't publish this blog post by 2am, I'll get charged $30), and I've deliberately lowered my standards to focus on kickstarting the habit.

With these two sets of examples laid out, I hope it feels obvious what the mistakes I was making were:

• Anchoring on an everything-is-amazing scenario and underestimating how hard it would be to get there.
• Spending too much time in the planning / design phase and not enough time in the execution and make-it-happen phase.

Strategies

• Notice when this is happening and understand the opportunity cost (from a Neel Nanda post). I'm working on a 2-month research project for my degree requirements and I spent the first two or three weeks reading papers and thinking about several possible projects, without fully committing to an idea. This is crazy! The time I spent in the planning / design phase has a significant cost: it's time I could have spent working on an idea¹. It's useful to build the skill of noticing when you're falling into a trap like this, and viscerally feeling the opportunity cost.
• Learn how to scope things.
• Make success inevitable.
• Convince yourself that finishing something imperfect quickly is probably raising long-term quality rather than hurting it.
  • Quality comes from iteration and learning rather than an all-out effort on a masterpiece. Zero-shotting quality is really hard (and cases where it looks like someone zero-shot quality are typically hiding a lot of previous iteration). Quick iteration is typically a much better approach.
  • A lot of the feedback/lessons that fuel learning come from a finished product, as you're able to see the result in action, get feedback from others, etc.². Undertaking and finishing many projects quickly allows you to capture a lot of learning per unit time.