Darkslide: Struggling to proceed

I have been playing with idea of keyboard-driven photo editor for some time and even published some version. https://darkslide.app/ 99% that this is yet another idea that will be just my. And I have been thinking why it was so hard for me to develop it and probably I put too much into one project: testing new product idea, learning new desktop framework (Tauri), completely new language that is famous to slap devs hard (Rust), plus even is I did UI development - I did not use React that much. The only thing that was not new to me is image editing as domain.
I also made the prototype version with LLM and then spent time unraveling all the shit it did. Now I see that rewriting it after PoC from scratch manually would be way faster and efficient. Recently almost gave up on the idea itself. Well, that it will be interesting to others, so I just started to use it as learning ground for Rust and wgpu framework. And some architectures I wanted to try to build. Mostly manually, something with agent.

Maybe it was too much since the beginning? I kinda struggle between struggling further to keep learning or finish v1 for myself, make it a free app for anyone who needs simple image editor without Photos.app and finish it with help of agent. Then pick something smaller for learning process as I would still like to learn some things about Rust ecosystem just to know when to really apply it. Finish with LLMs since I have very concrete things to finish and probably it will take me less than a week to build it via agent.

But I kinda feel torn between keeping the brain train or give up and use llm. 90% because of lack of energy for almost a year.

But 3rd option: do nothing now - probably will leave the app in current state as really it’s already enough for my current editing needs. As I prefer to make a photo good before editing. As you cannot fix shit photo with sliders.

Still no idea