I would say if, and only if, there is further improvement in doing so. I think in my implementation the 5998 still isn't biased optimally, just more so than it was before the mod. This I'm guessing was the maintain 6AS7G functionality. Glenn told me it bumps the plate current 20mA. Technically you can run the 6AS7 with the switch in 5998 mode, this makes the amp behave the way john's does but it runs hotter and most American 6AS7 tubes scream uncle. I tried this with a pair of RCAs, they started humming 3 seconds after I flipped the switch. A tap on the glass actually settled it back down and indeed I thought they sounded better this way, more full sounding just like the 5998 does, but the 5998 itself sounds better still.
I never actually asked but I'm guessing Glenn did this for safety sake so a person couldn't accidentally nuke a 6AS7 by having the switch in the wrong position. Since the extra cathode resistors only change the plate current by 20mA it is technically still safe for a 6AS7 tube in either mode.
Glenn posted in here a few days ago that the ideal bias on the 5998 tube was 125mA which is considerably higher than what he told me my amp is running even with the mod, so maybe there is still room for further improvement. A PM to the man himself would clear everything up I'm sure.
One note though, and an important consideration for you in particular: if that 125mA number is actually true as opposed to 80mA which is what my amp is supposed to be running, then it likely rules out the 596 as a rectifier in the amp post mod. You would be at 268mA current draw, uncomfortably close to the 275mA max of that tube.
The guru himself will probably chime in and correct any errors in my logic here.
I don't want to pretend to be an authority when I clearly am not, I'm just re-posting numbers that have been quoted to me in the past.