[algoth1]

I think you could probably feed a copy of a toki pona grammar book to a big model, and have it produce ‘infinite’ training data

[MarkusQ]

This is essentially a distillation on the bigger model; you'd wind up surfacing a lot of artifacts from the host model, amplifying them in the same way repeated photocopying introduces errors.

https://dailyai.com/2025/05/create-a-replica-of-this-image-d...

[eden-u4]

There are not enough samples in that book to generate new "infinite" data.