I have not built this schematic, but it is missing a grid leak resistor on the KT88. KT88s have a maximum grid leak resistance of about 270k. On power tubes the grid leak resistor works differently than on small input tubes. On power tubes if your grid leak resistance is too high it will drive your grid more positive, biasing your tube hotter. This is most likely why you are burning out your cathode bias pot. If your volume pot is all the way up your grid resistance will be over 1M. It would probably help to add a grid leak resistor after the grid stopper resistor. I would suggest between 270k and 470k since it will be in parallel with your 1M volume pot. This should fix your problems.