Building it yourself still costs something
Whenever someone says you can just build the tool yourself now, I have two reactions. First, excitement, because a lot more of those ideas are doable now. Second, a small accounting question: what are we counting as the cost?
In a reply I drafted, I listed tokens, time, and running costs. Pretty boring objection to a very exciting idea. Being able to make the thing doesn't mean I want to be on the hook for it.I always forget the upkeep
The answer keeps moving. Something too expensive or annoying to build today could be an easy afternoon project next year. Something that looks cheap in a demo can quietly eat a chunk of your week.
The first working version has a very good sales pitch
A small tool that does exactly what you asked for is hard to argue with. No screens you never open. You can reword it, drop a step, lay it out the way you work. And you don't have to wait for someone else's roadmap.
But the first version is when you know the least about living with it. Say it pulls data from some other service. Getting that onto a page is easy. Then the API changes, or the data shows up late, or a number looks wrong and you have to figure out why.
That doesn't make it a bad idea. It just belongs in the math. If I only compare the subscription to what the first version cost me in tokens, I'm leaving out most of the work I'm signing up for.
Your attention belongs in the calculation
Time is easy to ignore when building is fun. An evening messing with a tool can be worth it on its own. Not every experiment has to save money. Learning something and having fun are good enough reasons.
Replacing something I rely on is different. If it's part of my daily work, fixing it eats into that work. Even a quick fix means noticing it broke, remembering how I built it, and checking the fix holds.
When I pay for a product, part of what I'm paying for is that it's someone else's problem when it breaks. They don't always handle it well, but that's what the price is for. The comparison only works if I count that.
Make the decision small enough to revisit
For something small and just for me, I'll happily build it and see if I keep using it. I'm not rebuilding a whole product, just trying an idea. If it needs constant babysitting, now I know.
I ask what happens if I stop maintaining it. Can I get my data out? Can I go back to what I used before? If it breaks, does something important stop, or did an experiment just end? That decides how much risk I'm fine with.First question now: can I get my data out
The math will keep moving as the tools get better. I want to stay open to it without feeling like I have to own everything I could build now. Sometimes building it is a good afternoon. Sometimes paying is how I get the afternoon back.
Acknowledgements
From a reply I drafted about tokens, time, and running costs. The math is mine, and it changes every time the tools do.