Programming monster! Well, so that was what was causing the lag? A question that came to my mind: can DeepSeek help you solve these kind of problems?
Yes, the problem lied with the lack of proper cache implementation, which I knew how to solve circa two years ago, but Ren'Py had certain limitations that made it impossible at the time for me to implement, until one of the recent updates made to the engine.
AI is good at explaining basic programming concepts, generating boilerplate code, and doing repetitive tasks for you. It is not good for anything else at the moment, especially for niche languages like Ren'Py since all Large Language Models, including ChatGPT, Claude, Gemini, Deepseek, Qwen, Mistral will all hallucinate the answers and get lost in Ren'Py syntax, which is a mix of in-house solutions and Python, on top of Cython, GLSL, and SDL2.
The lack of codebase understanding is what kills transformer-based models for apps that are more complex than a calculator. This is not surprising; After all, it's all
just linear algebra.
The only practical use I've found for AI was to employ a snarky and sassy Discord bot to remind me to get off my lazy ass and do some work. You can also poke 'her' to get dunked on.
