Can an annoying bit of game research turn into a useful walkthrough? In this case, it did. After letting AI help inspect links and buttons in a Japanese browser horror game, I published the findings. One day later, that guide logged 301 page views. The twist is what I had already made on the side.
A horror game was getting attention. I decided to make one.
The spark was Migiwa Hospital, a Japanese browser horror experience disguised as a friendly personality test at a fictional hospital. I saw people talking about it and thought it looked interesting.
My next thought was not especially restrained: “I want to make a horror game too.”
So I made one. I saw something fun, wanted to try making my own, and started creating. No grand market strategy was required to make that decision.
My game and Migiwa Hospital were separate projects. The next part of the story concerns playing and researching the latter.
The scary part wasn't the ghost. It was checking every button.
In a browser puzzle, you follow links, test buttons and keep track of clues. The mystery can be delightful. Repeating the same little inspection steps? Less delightful.
I had AI help investigate the links and buttons while I explored Migiwa Hospital. That division matters. I wanted to experience the game and work out what was happening, not personally repeat every mechanical check.
Once the observations were organized, another possibility appeared: maybe the confusing part I had investigated was exactly what another player wanted explained.
Congratulations. An inconvenience had become an article idea.
Publish the answer, then look at what people actually read
I turned the findings into a walkthrough and published it quickly. The Migiwa Hospital Question 24 guide explains a late-game input, an earlier clue and the observed result pages. It distinguishes an answer that worked in one playthrough from a guaranteed answer for every player.
On October 14, 2026, measured in Japan time, that article received 301 page views. Of those, 189 had a Google domain as the referring host. The entire site recorded 847 page views that day.
Those numbers count page loads, not unique people or game players. They also don't prove that AI caused the traffic. The game's popularity, timing and the usefulness of a very specific answer could all have contributed.
Still, the result was real: a problem I found annoying enough to investigate was interesting enough for other people to open the article.
And then I remembered the horror game I'd made
Now there was a page attracting people interested in horror. I had also made a horror game.
So I arranged to add a route from the walkthrough to my own game. It was a simple thought: if someone enjoyed this sort of browser horror, they might also want to try another one.
Arranging a link is not the same as recording a successful visit or a completed playthrough. The next useful numbers would be link clicks, game opens and actual plays. At this point, the page views and the plan for the link are what I can account for.
The original impulse was “I want to make a game.” Somehow it had grown into game development, puzzle research, writing, publishing and promotion.
My hobby had quietly hired five imaginary departments.
What to copy if you're writing a game guide with AI
Start with a point where you actually got stuck. Ask AI to help organize the links, buttons and choices involved. Check its findings against the game or a reliable record of the playthrough; don't present an untested path as a universal solution.
Lead the guide with the answer the player needs. Explain the route and its limits. If you later promote a related game, make the link useful and check that it works. Measure the clicks separately from the guide's page views.
AI can cut the repetitive checking. The useful question still comes from playing the game.
So which part of this person is supposedly lazy?
Look at the actions: spotted an interesting horror game, made another one, played the original, investigated it, published a guide, checked the traffic and arranged a place to show the new game.
That is a remarkable amount of activity for someone accused of not wanting to bother.
The thing I didn't want to do was open the same links and check the same buttons over and over.
Apparently the plan was to save time on boring tasks so I could rest.
Instead, I used the saved time to make more things. This is the world's least convincing lazy-person defence.
