A spreadsheet is a database you haven't yet performed a migration on yet. I think the argument is less around better spreadsheets and more around migrating that data into real datastores once constraints are encountered (whether that's a desire to support large amounts of contributors, not run on Google, the dataset has grown too large, etc).

It's hard to go from Google Sheets to a web frontend to Postgresql and a snappy shared frontend. That is the problem statement to solve for imho (interface, schema, data portability).

No one has really solved the "Spreadsheet —> Database" migration painpoint. We don't have anything that is universally easy-to-use. The equivalent of a Hypercard or a Filemaker for the Web. (And no, I'm not talking about MongoDB or other NoSQL; I'm talking about the kind of thing you can hand over to a typical office worker.)

Agreed. I have a hunch that any solution to this problem will look similar to https://github.com/simonw/datasette if it develops. Tied together with https://supabase.io/ and you've built a versatile, powerful, open data collaboration engine.