CRDTs are also an excellent example because of how their supporting infrastructure is impacted by their design, namely that Postgres’ method of dealing with upd

CRDTs are also an excellent example because of how their supporting infrastructure is impacted by their design, namely that Postgres’ method of dealing with updates makes for massive write amplification. I wrote this [0] previously, but it still applies. IMO, as a dev, there are times where you really do need to think somewhat deeply about infrastructure choices. Unfortunately, knowing when you need to care practically requires you to already understand it. [0]: https://news.ycombinator.com/item?id=40834759 https://news.ycombinator.com/item?id=40834759