On AI

This project is built with AI assistance. All of it. I set the architecture, make design decisions, define quality standards, and evaluate the result. The AI writes code to a standard I enforce through testing, MVC compliance, logging, and refusing to accept anything that blocks the UI or cuts corners.

I am aware of the costs. AI infrastructure consumes water and electricity at scales that hurt the common folk. Governments subsidize power for data centers while household electricity bills climb. Workers in the Global South are paid next to nothing to produce training data under conditions that are frequently unethical. The flood of generated slop is degrading every corner of the Internet. I do not think these are acceptable tradeoffs and I do not think the companies driving them are acting responsibly. I paid one of them to help me build this.

I also could have spent this effort on something with cleaner ethics. An XMPP client. An IRC client. The performant email client that has been living in my head for years. Instead I built a Discord client, a native frontend for a proprietary platform whose values I do not share, because my communities are on Discord and I got tired of watching Electron freeze every time I switched channels. That is the whole reason. I wanted the thing I use every day to not be miserable to use, and I wanted it before the window closes.

I cannot replace Discord as a service. I am one person. Building a reliable live chat, voice, and streaming platform is orders of magnitude more involved and more risky than building a native client for an existing one. So I built the thing I could actually ship.

You can form your own opinion about whether any of that is worth it by reading the source.