Simplifying

2026 has been the most eventful year of my life, and through it all I’ve been working more than I think I ever have. In bouts of mental fatigue, I often pause and think “what is all this for?” This can refer to beliefs or ideas, material possessions, or habits depending on the context. In any case, there’s merit to thinking about the answer to that question, but I’m too tired to think sometimes! Another way to an answer is to ruthlessly simplify.


For example, let’s look at beliefs or ideas. Businesses pay billions of dollars to force their products into our minds. Social and political leaders wage cultural wars for a chance to win our hearts, leading us to reproduce their ideology through our day-to-day activity. Even family and friends offer their opinions on how we should live, sometimes indirectly, other times not. And, of course, we learn from and form normative opinions on our own experiences and goals. It’s difficult enough to bear all these views of the world at once. To try to live according to all of them is outright dizzying!

So I find it useful from time to time to let go of my beliefs about how things are or ideas about how things should work. If the idea doesn’t return, then whatever originally gave rise to it must no longer be around, and the belief was obsolete. If the idea does return, I trace its path back to me; then I can tend to it based on its origin.


When it comes to material possessions, there’s a basic imbalance between how hard it is to acquire versus get rid of things. We can buy things in a single click or acquire them effortlessly as gifts or trinkets. Getting rid of things takes at least a guilt-ridden trip to the disposal bin. A sale takes at least several clicks, a few texts or emails, and a decent chunk of a day.

All the advice I’ve seen for rebalancing the situation depends on sheer force of will. Retail commerce, especially online, is optimized to enable and encourage impulse purchases. Since that game is rigged to making buying easier and easier, I do my best to tie purchases to other benefits. For example, a thing I buy should replace at least one thing I own, and that replaced thing should have a real way out of the apartment. Or a thing I buy should save me a significant amount of time or money, preferably more than it will cost to maintain the thing. So far, this trade-off plan is working.


Habits are a big area where I’ve been questioning a lot lately, especially in my work. I’m a computer person by trade, and computer people, like anyone else these days, are conditioned to like flashy new things. With computer technology, the situation is so bad that we’re sometimes told to try a product before the person or company building it has found a proper use for it! The obsession with the flashy and new extends past the recent trend with large language models into all types of software and hardware that are supposed to make things simpler. The question is, simpler how? Too often, I find that a tool I use is optimizing for one-click convenience when what I actually need is a solution whose machinations I can understand and customize.

For example, for several months, I set up an AI agent to automate as much of my research process as possible. Just set up the steps and instructions and let AI cook, right? Wrong. Except for monitoring my experiments and analyzing my data and academic paper authors’ logical arguments, the AI agent spun its wheels and didn’t make significant progress unless I intervened. This included models like OpenAI’s latest available Sol and Anthropic’s latest available Opus. I could try Astra and Fable, but in the two days I spent uprooting the agents from the center of my research process, I felt in my brain the equivalent of a good stretch after a long time sitting. I’ve decided that LLMs will not be allowed roles in my personal work beyond spreadsheet and experiment monitor, both of which are probably optimizable to work using local AI models.

I’m also reconsidering the role of big, complex software that mostly supports features I don’t use or need. Why Obsidian (which is not even free software) if I can coax Emacs into doing what I need? Why Github or any Git forge for that matter if a bare Git server will do? The simplicity of directly doing what I mean to do and not doing what I don’t mean to do just feels good.

·