[sheepdestroyer]

They could easily list the data used though. These datasets are mostly known and floating around. When they are constructed, instructions for replication could be provided too

[coliveira]

They could, but even if they give this list the detractors will still say it is not open source.

[rvnx]

yes and as a bonus they may get sued, which in the long-term, makes free / offline models to not be viable

It would be so much better if all models were trained with LibGen.