Topic
Where model-program money actually goes: corpus curation, expert labour and prior compute versus the final training run.
No current published clusters are mapped here yet.