Just paint

My uncle and aunt travelled to India for a sabatical. They were there for a few months. My aunt, who is quite the artist, had an idea to "paint India" wherever she went. But when she was there in the moment, getting ready to paint, she would note that things weren't quite the way she wanted them to be. Maybe it wasn't the setting she had in mind. Or maybe the colors she had on the palette didn't quite match what she was seeing. So she wouldn't paint and wait for the next opportunity. And that one wouldn't be quite right either.

Later on in the trip, having painted far less than she would have liked, she realized the conditions for painting didn't have to be perfect, and the resulting output doesn't have to be perfect. What was more important for the goal of "painting India" was that she just paint. When she adopted this mentality, everything fell into place and she did what she set out to do.

I come from academia, where you don't publish until your work is as perfect as you can get it. And even then, the dreaded Reviewer 3 has some bone to pick with what you did. So the perfectionist mentality is still in me, slowing down if not stopping what I do. So I get in a similar mindset to my aunt's, where I keep second guessing myself when I'm in the early stages of a project.

The end result is up until recently either I wouldn't show what I'm working on because it's not to my liking, or I would decide not to do something because I didn't think I'll be good at it or produce something that is "good enough." As if Reviewer 3 is breathing down my shoulder at every waking moment.

This article attempts to address the problem my aunt faced in India. The problem of not starting. Moreover, I discuss the subtle art of going from not doing a thing to maximally doing a thing (the flow state). I will talk about this in terms of gaining psychological momentum. I then go into a hindrance to the momentum I often face when I have started something, where I convince myself that it will never be to my liking. It will never be sufficiently perfect so what's the point in continuing? I would call this the art of finishing what you started, except that I address it in terms of the concept of never finishing. I introduce the concept of Long Content: treating everything you produce as something that you will improve upon and evolve perhaps for the rest of your life. A side effect of this is that I focus more on the process rather than the outcome (I can't focus on the outcome because there will be no outcome).

"Just paint" is the first step toward gaining psychological momentum. I'll loosely define that as the probability that I will do the next step of a task given that I do this step. Psychological momentum is critial to my work as a programmer, as it is to my aunt as a painter. If I want to tackle some problem, I have to get started. I have to make a little bit of progress. This makes it easier for me to make a little bit more progress, which makes it easier, and so forth. Before you know it, several hours have gone by. I was in the zone.

Cognitive scientists often call this flow state. It is often associated with games, as there is a well-defined task you have to complete with quantifiable metrics for performance and deliberate feedback at every step. The brain likes flow state. Anyone who has played an addictive game knows this. The one I'm thinking about is 2048, which was release in 2014 the week before finals back at Stanford (and any school on the quarter system), much to the detriment of the undergrad population.

For my work, I like to think of this as momentum. Specifically, I like to think of it as getting a train up to speed. It takes a lot of energy and time, but once it's at cruising speed, it will be very difficult to stop. I only came across the term psychological momentum recently, in this paper. I think it describes my style of work even better than flow state.

For me, starting the process of "just painting" is the hardest part. Often the problem I'm trying to solve is very complex, or not particularly well defined, something that is also known as a wicked problem. How do I solve these problems? In software engineering, one of the first things you learn is problem decomposition. Importantly, for me in this regard, the goal in the early stages is not to solve the problem (although my approach pushes me in that direction). It is to get psychological momentum. It is to just paint. It is to get the train up to speed. It's a subtle difference in focus. How do I do that? If I can't quickly decompose the problem into some actionable piece (eg. it's a wicked problem), I just start working on a simpler but related problem.

I first saw this strategy used in 2010 by a friend of mine who was working in fasion design. Her job for the day was to draw out prototypes of dresses. She had a notebook with hundreds of them. The day we were both working together (on our respective projects) at a cafe, and she had a bit of a "drawing block" and couldn't get started. So she went on wikipedia and started reading about art and fasion in different time periods in history. She explained to me that she always does that when she's stuck. It's not her task but it is sufficiently related that it gets things started. It gives her enough psychological momentum to get the pencil to the paper.

This strategy is beautifully described in this beautiful lecture from Prof Emeritus Robert G. Gallager on Claude Shannon's problem solving style. Again: solve a simpler but related problem. That's it. One example from my life: when I took CS106A at Stanford, one of the problem sets had us design Atari's Breakout. That was a hard problem for a complete beginner like me. A simpler problem was to get a single ball bouncing around on the game board. Still too much for me. Maybe I could just get a ball moving anywhere on the game board even if it ends up off screen. Nope, too much. Ok fine, no movement. Just a single ball anywhere at all on the game board. No movement. Nothing else. Just board and ball. Ok that worked. I did things like that for quite a while in the beginning. Anyone watching me would say that I could be a bit more methodical. But that came later. The absolute most important thing for me is getting psychological momentum. Once I have that, then the proper "best practices" of problem solving show up naturally.

At the time of writing, I am 36 years old. I am one year older than my father lived to be. What does that feel like? Although statistically I'm expected to live several decades longer, it feels like I'm living on borrowed time. What would happen if I died tomorrow? A lot of my thoughts and my work, potential contributions, would die with me. Things I didn't make public because of the academic mentality of hiding things until they're perfect.

But I'm working in medical research and bioinformatics. These are fields where you have to maintain a high level of rigor. How do you reconcile this with the idea of something not needing to be perfect to show it to the world? The strategy I'm taking, inspired by this article by johnswentworth on LessWrong, is to continue being a perfectionist when it comes to finished products, but to share what I'm working on a bit more often. If it's not perfect, I'll talk about where I'm at in the project, what I've observed so far, how I interpret these observations in the present moment, and where I'm going with it.

We already do this in academia at lab meetings, internal departmental retreats, and conferences. I think the difference here for me is that I'm simply sharing my work publically a bit more often rather than signing up for a talk for some conference in two months. I'm no longer afraid of getting scooped (been there), because I am no longer in the academic "publish or perish" incentive structure. Even if all my ideas are stolen and published elsewhere, I can still point to them on my website to show my clients how I'm going to add value to their organization. Different incentive structure.

This section, and my general approach to producing and showing work, is inspired heavily by this piece by AI and tech writer Gwern on the concept of Long Content. In short, I consider everything I show to be constantly evolving. If you look at the date for this artcile and the others you'll see a date range. Chances are at least for a time, if you come back to this article or any other article on my website, you'll see it change. There is no law against this. We think of blogs, pubications, books, posts, as static documents. Snapshots in time. But what I'm thinking right now, at least in some contexts, is that it doesn't have to be that way.

A lot of my ideas now, that lead to my company's profit and my continued growth, can be traced back to my PhD thesis. This being said, I often cringe at my old work. As a thought experiment, what would happen if I spent an hour a day for the rest of my life simply revising my PhD thesis as my ideas matured, rather than cringing at old work and writing something related but completely new? What would my thesis look like at the end of my life? Not that I'm going to do that, but that is how I have been trying to treat everything I write and produce from here. I'm trying to get rid of the assumption that we have to finish something. It has to be sufficiently good before I show it to people, but that doesn't mean I have to finish it altogether.

Now to be clear, I am not arguing against having goals, or outcomes, that you want to achieve. Far from it. I'm also not saying to disregard deadlines and just set them for 50 years from now. My company has goals and deadlines to stay afloat. What I am saying is that I am trying to loosen the definition of "outcome" a bit such that I can "finish" a project (eg. a book, or software), but continue to revise after its done as I see fit, and as my ideas evolve.

What is an example of Long Content? Writer and entrepreneur Balaji Srinivasan just wrote a book called The Network State. He is selling it on Amazon, but he also has it online for his readers. The latter is significant because he plans to revise the book in real time as he gets more an more feedback, rather than simply writing another book.He talked about this on his most recent podcast with Tim Ferriss. His book is appropriate for the Long Content category because it is about a fascinating blockchain-based concept that is in its infancy but will envolve over time. As this happens, his book will co-evolve with it. Furthermore, if this concept of the Network State becomes a big deal down the line, the book may very well increase in value with time.

Gwern also discusses the concept of producing text that increases in value with time as central to Long Content. What does that look like? Here is a video of kids from 1966 describing what life will be like in the year 2000. In 1966 such a discussion probably seemed trivial. We all think about the future. But right now in 2022 this video is absolutely fascinating. What type of writing would be boring now but interesting down the line? Making predictions about Bitcoin might be a start. Right now it's at 20,000 USD. I'll predict that it will break the six figures in the next five years. Whether it does, or whether it goes to zero, this will be an interesting prediction in five years. What about journal entries? The boring day-to-day of your life? Yes, but also a snapshot of the 21st century to readers in the 31st century (insert the obligatory "if X doesn't destroy us first"). The Meditations by Marcus Aurelius is a gem for the modern-day Stoic. In reality, this book was a merely a journal to himself. He literally did not intend for it to be published to the masses. It is a snapshot of a man trying to motivate himself every day to keep going (which this article is turning out to be). There are probably plenty of people writing journal entires like that right now, but armed with 2000 additional years of education. Nonetheless, his journal is very much relevant and valuable today. What work of yours will be valuable in 2000 years? Or maybe the better rhetorical question: what work of yours WONT be valuable at all in 2000 years?

What's my point? Some arbitrary painting or text that is not interesting now could be interesting in a few years, and if not then, will be definitely interesteing in thousands of years when historians come across it. Do you critique ten thousand year old cave paintings for lack of skill or rigor or whatever else? You've seen better, so let's disregard it entirely? If your work is not appreciated now, it will be appreciated later. You might want to consider securing some storage space beneath the Arctic ice.

I encourage you to ask yourself long content specific questions, just as I have been doing for myself. It has helped me keep going. What would your work look like if you slowly revised it for the rest of your life, rather than considering it done forever and jumping to the next thing? That doesn't mean that J.K. Rowling should have stopped halfway through writing Harry Potter and the Sorcerer's Stone and just revised that for the rest of her life. I'm talking about instances where you see your work having a common thread. An idea that you developed years ago that seems to keep coming back again and again in your work. This is the instance where perhaps you should consider going back to the original documents with this idea and do the Long Content appraoch.

Treating my work as Long Content has helped me psychologically keep going in my projects. The obvious question here is why? Is there a neuroscience-based explanation why the Long Content approach has helped me? Is there a reason why focusing on the process and all but disregarding the outcome has helped me push forward in my projects when there is no one breathing down my neck and I have to motivate myself?

The Long Content appraoch has a neuroscience-based benefit: dopamine optimization. To clarify, I'm talking about dopamine in the context of our motivation circuitry, not the so-called "dopamine hits" that you get when someone likes your social media post. Dopamine affects motivation to do a thing rather than reward from doing a thing. Quite a lot of work has been done here. I envourage you to listen to this podcast by Andrew Huberman where he sums up the recent research and take-home points around dopamine.

One of the key findings he explains is the idea that dopamine is effectively maximized when one seeks reward from the process rather than the outcome. For example, a runner who is training for a marathon would experience high levels of dopamine prior to the event, but then a dip below baseline after the event that would potentially manifest as depression. On the other hand, a runner who simply enjoys running for its own sake would be less likely to experience this dip.

I have direct experience here. I go to the gym for its own sake, an it has led to me playing a much longer game: I've worked out consistently since 1999, and I am going to guess that my dopamine pathways are properly optimized here. The reward is in the act of working out itself, not the fitness goals. I haven't talked to any dopamine experts about this exact "long content" context yet, but the idea here is if you completely remove the idea of "finishing" your work, then you might be able to tap into this "seek reward from the process" heuristic and therefore have higher and/or more stable dopamine levels, leading to more general motivation to do what you do.

Here, I attempted to address the problem of starting a project and the problem of staying motivated to continue. To start a project, I introduced the concept of psychological momentum and talked about the concept of solving a simpler but related problem whenever I'm stuck for the sake of simply maintaining and building momentum, regardless of whether that is the most efficient path for a given project. I then talk about focusing on the process rather than the outcome from the standpoint of writing and showing something that is not yet perfect, and from the standpoint of Long Content, in which you really downplay the outcome and build/revise something for years if not decades. I talk about this from the standpoint of dopamine optimization, where I suggest that this process focus is a neuroscience-backed way to maintain motivation.

Now regardless of whether we're talking about your work in the near term or thousands of years from now: just about every idea we have builds off of the ideas of the past. So I think part of moving humanity forward is sharing what we've got, even if its not perfect. You really never know who it will influence, even if it's only one person who influences someone else, who influences someone else, who influences someone else, who ends climate change by solving cold fusion. It makes me really think about the higher order effects of every action we take. I know there is a lot of randomness in the world, but I at least at take it as a hypothesis that our actions can have profound higher-order influences tomorrow, ten years down the line, or ten thousand years down the line. It will take a lifetime to test this. What do you have to lose if you think this way?

Just paint.