A 10x reduction in pre-training means a 10x faster feedback loop. I'm pretty sure any lab would sign up for that. You can start experimenting on different approaches much more aggressively.