Build Your Own Tool in 15 Minutes, at Your Own Risk

Why I agreed that anyone can build their own tool in 15 minutes, counting from installing 1Password, where I draw the line on password work, and how this…
Someone posted that anyone can build their own tool now, and I replied that this is genuinely true. Even counting from installing 1Password, 15 minutes is enough, so I think people should just do it — with the caveat that this work touches passwords, so research it properly and proceed at your own risk. That was the whole post. This article is the reasoning behind that short reply: what I actually meant by 15 minutes, where I draw the line when credentials are involved, how this plays out inside Japanese companies, and one note about my photo showing up in someone else’s post.
A short piece of context for readers outside Japan: in Japanese companies, the gap between what an individual can do in an evening and what an organization will formally approve is unusually wide. That gap is most of what this article is about.
Why “anyone can build it” is true now
More people are arriving at the feeling that they could probably build the thing themselves. I agreed with that post because the feeling is now correct.
What separates the people who build from the people who don’t is not technical skill. It is whether they take the first step. Only the people who actually touch it get the small moment of “oh, that’s all it was.” The people who don’t keep carrying the assumption that it’s beyond them, and that gap widens the longer it goes unchallenged.
The important part is to avoid framing building as a talent story. If it is now true that anyone can do it, then the way we talk about it should shift from “look what this impressive person made” to “this is a thing you can just finish.” That’s why my reply wasn’t admiration. It was an estimate.
What “15 minutes from installing 1Password” actually measures
When I said 15 minutes, I didn’t mean the time to produce a finished product. I meant the distance from nothing installed to something running. The phrase that mattered was “counting from installing 1Password” — I deliberately started the clock at the earliest possible point.
The most common reason people never start is the belief that setup will eat a whole day. If you picture losing a Saturday to environment configuration, you don’t begin at all. Starting the count from the very first install is a way to kill that assumption. If the whole thing fits in 15 minutes, you don’t need to clear a weekend; you can do it after work. The decision stops being “how do I spend my weekend” and becomes “do I have 15 minutes today,” and that single reframe changes how many people actually start.
An estimate has a second use: it tells you when you’re stuck. If something you expected to take 15 minutes has passed an hour, that’s the moment to separate the two possible causes — something specific to your environment, or a step you misread. Naming where you’re blocked gets you out faster than grinding. And to be clear, 15 minutes is my own sense of it, not a promise that holds for everyone.
Password work is where “at your own risk” stops being a formality
I added a sentence to that post on purpose: because this touches passwords, research it properly and do it at your own risk. That wasn’t boilerplate. If you’re the person telling others they should go do something, that sentence is part of the job.
Credentials are a domain where the cost of undoing a mistake is higher than almost anywhere else. My own minimum conditions are three:
- Be able to explain, in my own words, what I’m touching before I touch it
- Have the way back in place before I start
- Keep it inside a boundary that doesn’t involve anyone else’s assets
“Anyone can build it” and “nothing can go wrong” are completely different claims. As the work gets easier, more people skip the checking step. If you’re going to say people should do it, you should say which part is on them with the same volume.
How this lands inside Japanese companies
For anyone pushing AI inside a Japanese organization, this is the real subject. What takes an individual 15 minutes takes a company weeks. The blocker is almost never technical. It’s that nobody has decided who owns the risk.
Credential handling in particular is the thing IT departments least want touched. A motivated person on the business side builds something genuinely useful, company account information ends up inside it, and now there’s a system nobody can see. Personal risk-taking and company assets shouldn’t sit on the same desk. Learn in your own environment; when you bring it inside, put it through the approval path. That two-stage split lets people move fast without creating something unmanageable.
There’s an implication here for overseas companies trying to reach the Japanese market too. “Anyone can do this in 15 minutes” does not move a Japanese decision maker. They stall at “can we, as a company, permit this?” So the message works better when ease of use and the boundary of responsibility — what permissions are involved, where accountability falls — appear in the same document. That’s my read, but the place things stall is remarkably consistent.
Someone used my photo in their post, and that’s fine
The last thing in that post was unrelated to building: a person named Yu had used an image of me. I’m completely fine with it.
If you publish consistently, your face and your diagrams get clipped and carried around. You can’t stop that. So I decide up front whether I’m publishing something on the assumption it will be reused, or not publishing it at all. Getting angry after the fact about something I chose to publish is, in my view, a design error on my side. Treating it that way is simply easier to live with.
In Japan, attitudes toward quoting and reposting vary a lot from person to person. So the practical move isn’t “always escalate” or “anything goes” — it’s having your own standard and knowing it in advance. Mine is that if the use doesn’t create a misunderstanding, I take it as distribution.
Taking this into your own team
If you want to use this internally, it comes down to three things.
First, pick one piece of learning that genuinely fits in 15 minutes and have everyone do it. Lowering the barrier to starting beats running a large training program. Second, decide as a rule that any learning which touches passwords or credentials happens in personal environments, not on company accounts. Third, prepare the images and materials you publish on the assumption that they will be reused elsewhere.
Now that anyone can build, the difference comes from how fast people start and how clearly the lines are drawn. Deciding both at the same time is the fastest path.
FAQ
Can you really build something in 15 minutes?
I said 15 minutes is enough even counting from installing 1Password. That number measures the distance from nothing installed to something running, not the time to finish a polished product. Some people will get stuck, so it is not a guarantee.
What should I watch out for when the work touches passwords?
In my post I added that you should research it properly and proceed at your own risk. My own minimum conditions are: be able to explain what you are touching before you touch it, have a way to undo it, and keep other people's and your company's assets out of the experiment.
What should I do if someone uses my photo in their post?
Someone used my photo and I am completely fine with it. The practical approach is to decide in advance whether you publish material assuming it will be reused, or not publish it at all. People's standards differ, so there is no single right answer.
Based on a post on XOriginal post ↗