C-SPANN, an Alternative Vector Index for pgVector

I found the following argument on some of the downsides of pgVector interesting:

It's worth reading Alex' whole article - it's mostly bullet points and summary paragraphs. It reads like the opposite of our internal engineering design document for C-SPANN, which has been built to address exactly the weaknesses in pgVector that Alex highlights.

Another article, A Practical Guide for Choosing a Vector Database: Considerations and Trade-Offs, seems to have been written by someone with some real-world experience. Again, their cautions line up with what we're trying to accomplish with CockroachDB and C-SPANN.