Read the paper. It claims that this is consistent with 2 spillover events, but it's also consistent with 1 spillover and an early mutation. The mutation between the 2 lineages could happen on either side of the spillover.
But a 2 spillover event suggests there was a pool of infected animals with multiple lineages that were all already capable of zoonosis. So why only a single secondary event? This suggests the pool was small, contact was limited, and that the pool wasn't sustained for long. OK, then it's comparably likely that a mutation would happen on either side of the spillover.
We don't know how many spillover events there were. Just just know that it's at least two. Most spillover events probably do not lead to a sustained outbreak.
The problem with positing that the mutations happened after spillover is that the timeline is way too short. Multiple variants were present at the market just weeks after the initial human cases. That points to differentiation before spillover, probably on the farms where the animals were being kept.
But a 2 spillover event suggests there was a pool of infected animals with multiple lineages that were all already capable of zoonosis. So why only a single secondary event? This suggests the pool was small, contact was limited, and that the pool wasn't sustained for long. OK, then it's comparably likely that a mutation would happen on either side of the spillover.