Video game fan translator says AI output is better than their own translation, deletes 3 years of work on project - "I was wasting my time"

# Update #3 Source: [https://ayana0720.wordpress.com/2026/10/07/update-3/](https://ayana0720.wordpress.com/2026/10/07/update-3/) After three years of working on this fan translation project, I’ve decided to drop it\. Why? I’ve spent the last couple of weeks looking more seriously into AI translations and actually trying AI TL patches\. Honestly, the results were a lot better than I expected\. Good enough, that I, in fact, I eventually started double\-checking my own TL against AI output\. And that’s where it became pretty obvious that I was wasting my time\. I fucked up\. I hate to admit it, but the AI outup was better my own, real human translation\. After putting three years into this project, that’s obviously not a particularly pleasant thing to realize\. I’ve spent an absurd amount of time translating, correcting, rewriting, checking lines, fighting this shitty engine, upscaling, porting, and trying to make jokes sound funny and natural\. I’m not interested in pretending otherwise just because I’ve already invested three years into it\. Sunk cost is sunk cost\. The problem is that I also don’t have enough money to redo the entire project using AI myself, so there isn’t really a realistic way for me to continue from here\. I’m not going to spend another huge amount of time manually translating everything just to end up with something that I’ve already convinced myself is inferior\. So I’m dropping the project\. I’m deleting every single TL file I have and I’m not planning to come back to it later\. Hopefully MAO\-TLs can handle the rest from here\. This project would be better off in their hands\. Hands of someone who can produce a translation at the quality level Tsuriotsu actually deserves\. Thanks to everyone who followed the project or supported me while I was working on it\. Sorry that after three years, this is how it ends\.