After a period of experimentation, I officially started a new project today. The domain name and everything are registered, but the name is being kept confidential for now.
In fact, what I need is a boundless space that can satisfy my ambition for an "editor that accommodates everything," while also making it easy to add components as my ideas change, and allowing for the continuous accumulation of data and a knowledge base.
This might just be a project that is "friendly" to me. After sharing my ideas with friends at major tech companies, I'm almost certain they wouldn't do this—because when you have everything, it ends up being nothing for the user.
That’s exactly why I have the motivation to keep going: I can spend six months completing the first phase and then many years continuously improving it. With the help of cursor.ai and claude, I no longer have to fear "anti-human" front-end code.
Today, I've integrated all the ideas I've experimented with over the past period:
- Local AI dialogue;
- AI rewriting;
- Multi-document support;
- Block-based architecture;
Currently, there are two major challenges:
- Organizing and integrating Server-Side Rendering (SSR) and Client-Side Rendering (CSR) components requires a lot of mental energy;
- The consistency between BlockSuite's documentation and its code is decreasing. I understand the BlockSuite team is aggressively rewriting their code, but the design logic changes significantly even within a month or two;
Regarding the second point, I suppose that's just how it is; at least my own thinking is even more jumpy than theirs.
I'm posting this update during a break when my brain cells need a rest. I also look forward to one day being able to turn this private repository into a public one.