Luke -- Hardly a dumb question at all!
A basic function of vacuum tube operation is that the negative voltage that appears at the "control grid" (pin #6 in this case), controls the amount of current the tube can pass from its cathode element (pin #5). There are other factors that control this as well, but for the purposes of this discussion, they are not important. All this business of the 330K and 680K resistors and whether they should or should not be connected and what their value ideally should be is all a secondary effort to achieve the primary goal: controlling the current flow though the tube to a target value, and causing all four output tubes to operate at this target value at the same time. The current through the tube flows from ground (the negative side of the power supply) to the cathode element inside the tube, and then by way of the emission process, to the plate element, and back to the positive side of the power supply, forming a complete circuit. As designed however, there is no way to know how much current is actually flowing through each tube.
At the time the unit was manufactured, Scott ordered tens of thousands of 7591 tubes, and they could order them with selected characteristics -- so that each tube would in fact draw about the same known current flow with a given bias voltage applied to the control grid. Tubes -- like all manufactured devices -- have physical tolerances to their construction, and these tolerances can then vary the actual characteristic from that of the designed bogey characteristic for the tube type in question. Back in the day then, having the control grid voltage set to a fixed value was no problem, because the tubes installed were chosen to operate with a selected characteristic. Today however, Scott is gone, the American tube manufacturers are gone, so that we are left (primarily) with foreign manufactured tubes -- that while pretty good -- do not have the tight tolerances that the original American pieces were built to, let alone ones chosen for a precise characteristic. Therefore, today we need to be able to measure the current that is actually flowing through the tubes -- and be able to adjust it (via the negative control grid voltage), so that the original target current flow can be achieved. So again, without the small (recommended) 10Ω resistors installed to connect your meter to, you have no idea of how much current each tube is passing; sort of like trying to hit a given speed in a car with no working speedometer. Your adjustments with your foot on the accelerator are like changing the pin #6 control grid voltage. But without a speedometer to look at, you have no idea how those changes are effectively bringing you to your target speed -- or in this case, how the changes in control grid bias voltage are bringing you to the target current flow through the tube. That's why installing the current monitoring resistors is job #1.
The optimum target "quiescent" (no signal) current flow can vary from one application to the next, but generally, a current flow of about 32-35 mA is correct for 7591 tubes in classic designs like Scott used. Thanks for the comment on the wiring in my unit. The current monitoring resistors I used are .5 watt devices with a tolerance of 1%. Using highly accurate high quality resistors ensures that the current you measure will be accurate, and that the resistor value will remain stable as the unit heats from cold turn on temps, to full operating temperature.
Dave