Honestly, duckdb looks fine to me (I haven't played with it yet, though articles like this would put me off of it, and make me question how long its going to be around), it's the the tone of the article, which sets up what appears to me to be a strawman, and then plows on.
Also, my naive and quick look at your docs gives me the impression this isn't actually a storage engine (which didn't come across in the article), but instead an in-memory query engine? The FAQ doesn't really clarify this (e.g. is there an actual storage format, how does this compare to using one of the more traditional dbs, how stable it is (c.f. sqlite being recommended by the library of congress), does this assume all the data can be loaded into memory, or can I stream the data in).
Also, my naive and quick look at your docs gives me the impression this isn't actually a storage engine (which didn't come across in the article), but instead an in-memory query engine? The FAQ doesn't really clarify this (e.g. is there an actual storage format, how does this compare to using one of the more traditional dbs, how stable it is (c.f. sqlite being recommended by the library of congress), does this assume all the data can be loaded into memory, or can I stream the data in).