Isn't the beauty and the hype coming (mostly) from DuckDB itself? Which is free, extremely fast, and developer-friendly. I haven't used MotherDuck service yet, but DuckDB things like:

```

CALL load_aws_credentials();

SELECT * FROM "s3://some-bucket/that/requires/authentication.parquet";

```

or

```

-- read all files that match the glob pattern

SELECT * FROM 'test/*.parquet';

```

I used DuckDB here and there, and the speed and efficiency, e.g., as a zero-store layer on top of my data lake (e.g., see Build a poor man’s data lake from scratch https://dagster.io/blog/duckdb-data-lake) is pretty cool to me. Also, that you come pretty far with a single file database is somewhat surprising to me.