Training a Personal Writing Assistant with Pre-2024 Data

unixterminal · x · 2026-07-12

The author shares their experience training a personal writing assistant. To ensure data "purity," they specifically collected text from before 2024 (comparing it to low-background radiation steel used for medical devices). They also humorously thank whoever pirated their published WSL book and uploaded it to a shadow library, as it conveniently provided useful training corpus for the model.

Related event: Developer Trains Personal AI Writing Assistant with Historical Data(2 posts)→

Original post →

More from coding & agent

coding & agent channel →