Decision: I’m going to let hobbies justify themselves by curiosity and learning. I’m going to judge responsibilities by whether I can make their total cost legible and accept the tradeoffs that cost creates.
That distinction has taken me longer than it should have. I’m comfortable assigning a price to hardware, software, and a cloud bill. I’m much less consistent about pricing the Saturday afternoon, the available machine capacity, or the attention I keep returning to a system after I’ve technically finished using it.
The result is predictable. A project looks nearly free because the equipment is already here and the software costs nothing. Then it occupies storage I wanted for something else, consumes the quiet block I meant to spend writing, and sits in the back of my mind while I’m doing unrelated work. None of those costs appears on a receipt. They still get paid.
Curiosity is allowed to be the return
A hobby doesn’t need to produce a practical return. That’s part of what makes it a hobby.
I can spend an evening trying a local model because I want to understand how memory pressure changes its behavior. I can rebuild a small tool in an unfamiliar language because the language interests me. I can make a dashboard that nobody needs because turning raw facts into a useful view teaches me something about the system underneath.
If the result is awkward, temporary, or less efficient than the thing it was meant to replace, the time isn’t automatically wasted. Learning was one of the outputs. Enjoyment may have been another. Trying to make every hour prove its commercial value would turn curiosity into a badly managed consulting engagement with myself.
I’ve made the opposite mistake too: defending every technical project as learning when I was mostly avoiding the question of value. Curiosity is a valid return, but it isn't a magic coupon. I still have a finite amount of money, space, energy, and concentration. Saying “it’s a hobby” changes the expected return. It doesn’t make the inputs infinite.
This is where bounded commitment helps me. I can decide that an experiment deserves a modest amount of cash and a few focused sessions because the question is interesting. I don’t need to predict a useful product at the end. I do need to be honest that the same hours can’t also go toward another project, rest, or time with people I care about.
That opportunity cost used to feel too abstract to count. It became clearer when I stopped treating free time as leftover capacity. A free evening is not idle compute waiting for a job. It has competing uses, including doing nothing technical at all. Spending it on a hobby can be a wonderful choice. Pretending I spent nothing because no invoice arrived is just inaccurate accounting.
Responsibility needs the whole price
A responsibility is different because somebody is relying on the outcome, even if that somebody is only me. The commitment has crossed from “this might be interesting” into “this needs to remain usable enough to support another plan.” At that point, purchase price is the least interesting number.
The total cost has three parts for me. The first is money: equipment, paid services, electricity, replacements, and anything else that competes for the same budget. The second is scarce capacity: storage, memory, network headroom, desk space, and the time windows in which disruptive work can happen. The third is attention: the concentration required to understand changes, recognize odd behavior, and make decisions when I’d rather be thinking about something else.
Attention is the line item I’m most likely to falsify. Ten small tasks do not necessarily equal one larger task. Each can require a different mental setup. Reading a short release note, remembering why a setting matters, and checking one result may consume fifteen minutes on the clock and much more of the evening’s usable focus. Context switching sends no bill because it knows I would dispute the charges.
Capacity has similar traps. A machine with unused memory appears to make another workload free. But that margin may be what lets a demanding job run without negotiation, what keeps an upgrade from becoming a storage shuffle, or what allows me to test something new. Using the last available unit has a cost even when the dashboard says utilization improved.
The economic test isn’t whether a responsibility earns revenue. A household system, a personal archive, or an automation that saves repeated effort can be worth real commitment without making a dollar. The test is whether the benefit is substantial enough that I’m willing to give up the other things its full cost would have purchased.
That wording matters. “Can I afford it?” often produces a yes because I examine each resource separately. I can afford the subscription. The server has room. The work probably fits somewhere. The honest question combines them: what money, capacity, and attention will no longer be available if I make this promise?
Sometimes the answer favors the responsibility. A dependable local capability may be worth less room for experiments and fewer spontaneous changes. Sometimes it favors the hobby. A weekend spent learning something useless can be a better use of my attention than squeezing a marginal efficiency from a system that already works.
The tradeoff I’m accepting is that not every promising project gets promoted into a promise. Some things can remain enjoyable precisely because they aren’t required to justify more than the curiosity that started them. Other things deserve a firmer commitment, but I won’t call that commitment cheap merely because the software was free.
I still like building systems enough to underestimate them. I probably always will. The correction isn’t to make hobbies defend themselves with a business case. It’s to protect curiosity from false seriousness and protect responsibilities from false accounting. Before I make the second kind of commitment, I want to see the whole price and be able to name what I’m choosing not to do instead.