I spent days vibe coding the wrong game and ended up making this

For the last few weeks I’ve been doing what I can only describe as low-conviction vibe coding.

First I made 3 npm packages.
Not terrible, not revolutionary, not needed.

Then I decided I was obviously ready for something bigger, so I tried making a procedural backrooms game in Three.js.

I restarted that project 3 times.

Each time I thought ok, this time I have the architecture, this time I know the generator, this time it won’t turn into a weird pile of rendering hacks and broken gameplay ideas.

It did. Every time.

So I gave up and made something smaller and dumber:
a browser game about surviving LinkedIn, recruiters, ghosting, and other white-collar psychic damage.

It’s basically a survivor-like, except instead of monsters you deal with things like recruiter spam, “seen”, other candidates, and general corporate nonsense.

Honestly this was supposed to be a joke project, but it ended up being way more doable than the backrooms thing because the scope is smaller and the joke carries half the weight.

Built it with Pixi + TypeScript.

The funniest part is that making the game itself was easier than getting the tone right.
The early versions looked too serious, then too abstract, then just bad in a boring way.
Only after leaning into a more goofy fake-pixel style did it start to feel like the actual idea.

Anyway, if you want to see the result of abandoning one bad idea and replacing it with another, it’s here: https://career-survivors.protohiro.com/

No roadmap, no deep meaning, no market validation.
Just a small game I managed to finish before restarting it 3 times.