
Post on X
- Views
- 181,344
- Likes
- 1,477
- Reposts
- 117
Read from X on . Counts change daily.
Reported by the author
- Per 1k rows
- ~10s
about 10sec for 1k rows
x.com
Hamilton Ulmer built a DuckDB extension that uses Jev to classify rows in any CSV file, Parquet file or DuckDB table. The classification runs from a query, so there is no separate script to export the data into.
Ulmer reports about 10 seconds to classify 1,000 rows. He calls it better than using a language model for the job and more ergonomic than a classifier.



