iiinova — where your digital self keeps evolving

Replies

Dan Visioni @danvisioni.bsky.social · 05/07/2026

Starting to develop some hot takes on LLM for research. Even the best model is actually very dumb at doing brainstorming, lit review, all that. Just boring word salads - I can do better while drunk (is there any other way to do lit review?). BUT there is clearly some potential for climate science:

1143

Dan Visioni @danvisioni.bsky.social · 05/07/2026

1) Building more reproducible, robust data analyses pipelines exp with hack-ish datasets that normally require crazy amt of work to make sense of. Claude is amazing at digging through data structures like NetCDF/Zarr and making sense of them at light speed. No need to come up with exceptions urself.

120

Dan Visioni @danvisioni.bsky.social · 05/07/2026

2) I used to have a lot of shameful code: IDL for TEM calculations, Matlab for my favorite figures, NCL for way too many things. Porting them to Python (and verifying output bit-by-bit) is so fast and pleasant, and you end up with much better code. Porting/modifying F90 ESM code is pretty doable too

120