Unfortunately a lot of what are presented as improvements in Rhino are just gimmicks.
Take for example a recent thread about the MoveHole command. Any competent Rhino user can move a hole by themselves by untrimming the surfaces involved and then moving the hole surfaces and then trimming and joining the surfaces. Why is that difficult to program? Looking at what the MoveHole command does it is clear that whoever programed that function has no clue how to move a hole in Rhino. As a result the function MoveHole is just a gimmick and its not the only one.
Getting back to what was being discussed → FilletSrf.
The RhinoWIP FilletSrf command will now create a string of fillets across tangent faces. That is a huge improvement over the user having to tell Rhino where each fillet goes as in Rhino8 and earlier. However, one serious limitation in the current implementation is when the string of fillets that goes across tangent faces runs into another fillet of the same size. This happens all the time. In the video Bobi posted above you can see more than a dozen places where a string of fillets crosses a same size fillet. If the user has to stop and fix each of the places where the fillet stumbles across a same size fillet that’s a huge hit to productivity.
Here is an capture from the video above where the spherical fillets are marked with red dots.
What is particularly annoying about this bug is that (as Tom and Bobi pointed out) the code to create the needed spherical surface has already been written long ago. Apparently The user is expected to call this code when its needed but the FilletSrf command can’t. Why?
Some of the other bugs can be found in this thread