It’ll be interesting to see if there’s an increase in slop or poisoned data to deliberately throw this tool off. Could be a fun experiment.