Raw LLM Responses

It's helping me currently. I've been doing open source coding for years, very slowly. I've applied some Claude just recently to some of those projects and it's already identified a ton of issues I hadn't. It even wrote a text console 3d raycaster after I struggled for a few months. I'm using various AIs to help me write my space simulator game. It's going pretty well, it's like having an enthusiastic partner to work on the code with me. GitHub Copilot has really neat online code suggestions as I'm typing - it often gets the gist of what I'm laying out in HTML or JSON and suggests useful items. Or it sees I'm multiplying numbers by 10 and start suggesting similar attributes later in the file I wanted to update. Coding agents are pretty powerful, but you still have to understand what it's producing and guide it in a way that makes it extensible and doesn't build up technical debt. Like any intern, it sometimes gets it wildly wrong. Knowing what is right code and nonsense code is important. I'm pretty excited with what it's helping me create. I haven't gotten this far on a game since the time I had other people working on a game with me, a very long time ago. For a solo dev or small teams there's already a lot of benefits to it.