Did you pick up that plastic bottle after you were done? 🙃

very cool! nice to see an authentic take on software development

I really enjoy using duckdb, especially when you structure data to take advantage of the built in column compression such as run length encoding. Many of the data assets at Karius at built using duckdb and dbt.

What are your thoughts on COM Poisson since this allows under dispersion? I’ve found for some data this can handle this type of heteroscedasticity

There’s some famous quote along the lines of “Show me the incentives and I’ll show you the outcomes”

Academic software will likely continue this way until there’s better incentives for writing better software

Also I think UML is rarely used now 🙂