LLM caches only work for 100% input matches for all previous text, which will never happen outside of a single session.

Things such as system prompts are already cached.

Snow Leopard was one of the most celebrated Mac OS X updates for a reason.

Safari does in fact feel snappier.

Embeddings are very very much used for the use cases mentioned, they're just not sexy to blog about.

Trueee,,,the use cases normally are very engineering oriented and not useful for every end user.

You could put in your AGENTS.md/CLAUDE.md for it to use polars.

Modern LLMs use it just fine.

What makes a man turn neutral? Lust for gold? Power? Or were they just born with a heart full of neutrality?

With the exception of data.table (which has a painful API), anything else in R will be much slower. https://duckdblabs.github.io/db-benchmark/

You can do to_pandas() and from_pandas() which doesn't have much overhead, or to_numpy() directly which is typically zero-copy and thus no overhead: https://docs.pola.rs/api/python/stable/reference/dataframe/a...

It's subjective, but IMO polars's APIs are better and than pandas's for ETL. Notably, pandas ported over polars's use of .col() for the 3.0.0 release: https://pandas.pydata.org/docs/whatsnew/v3.0.0.html#initial-...

Also, pandas's group_by() and window function pipelines made me want to tear my hair out while both are easy in polars.

At my work I had convinced the ML pipeline engineers to switch from pandas to polars for even small ETL pipelines and there were notable performance gain with better CPU/memory utilization.

If a library is performant at large datasets, it is likely performant at small ones too.

I’m not disagreeing with that statement at all. You missed my point that there are thousands of people making small Python scripts for education and personal projects everyday. In those circumstances the performance concerns are irrelevant and the ergonomics of good pandas documentation and community knowledge make it a better choice.

That inertia is not a good thing, and it's partially why there's stagnation in data science. Polars is more than mature enough in both documentation and resources for it to be a daily driver.