For many years, I used a different approach to memo() than React. I liked it, but I didn't implement it in the most elegant way. I tried many different options, but the results weren't satisfactory. Besides memo(), I was annoyed by the inability to move elements between parents. But what I really like is working with mutable data and the ability to use any source.
I was planning a new attempt at implementing a React extension, but I came across a discussion about AI defaulting to React. And the thought crossed my mind, "Why not write an entire framework?" Given the AI-First trend, why not write libraries in this style? Perhaps the library will be useful for more than just me. Furthermore, the specification and tests can be used to test the capabilities of different models, which is something I'd also like to try out eventually.
The decision to write the library from scratch, rather than building an extension over ReactJS, was purely spontaneous. I didn't think a topic for a post here would come up so soon. This library is kind of a showcase of what's possible with free chats, a subtle jab at React developers, and a benchmark to see which AI writes code better. Oh, and it also features some solid ideas that I’ve been using for a decade already.
I was planning a new attempt at implementing a React extension, but I came across a discussion about AI defaulting to React. And the thought crossed my mind, "Why not write an entire framework?" Given the AI-First trend, why not write libraries in this style? Perhaps the library will be useful for more than just me. Furthermore, the specification and tests can be used to test the capabilities of different models, which is something I'd also like to try out eventually.