It's unclear to me how P6 was a harsher environment for core contributors than P5 could have

In some ways it was and in some ways it wasn't.

P6 had the permission to break everything because it had no users and it had a hand-waved "compatibility" plan that still doesn't work. There was rarely any pressure to deliver anything useful and usable because there were no users and no schedules, and it suffered for a long time because of that.

That lack of market pressure produced a lack of technical leadership, and you can see it in the fights between Dan and Leo, for example. It's difficult to build the right thing when you don't know what to build and it's very difficult to say no to the wrong things when you don't know what to build.

At the same time, Perl had no permission to break anything because P6 held it back for so long. Not only was P6 the future version destined to replace Perl, it had Larry's full attention and was subject to Rule 1 and Rule 2 design.

If I take solace in much from that debacle, it's that Perl has adopted the monthly/yearly release cadence that worked for us in Parrot as well as a deprecation policy (even if it didn't work so well for Parrot).