RF as in radio frequency. They work by sending radio energy down a coaxial cable to the tip where it's absorbed. At the desired temperature, the tip alloy reaches its Curie point and stops absorbing energy. The instant it cools, it absorbs again. It's a far faster control loop since it's all physics and doesn't require a sensor (with some non-zero level of thermal impedance between it and the tip itself), no separate heater element (the tip is the heater, again with that thermal impedance) and feedback to the controller.

The downside is the only way to change temperature is to change tip to one with a different alloy. In practise, this isn't any kind of issue because the only reason you usually need to ramp the temperature with a conventionally-heated iron is because you are trying to compensate for the heater-tip-sensor feedback loop being unable to keep up when soldering big things, which is a problem the Metcals don't really have in the first place.

The Curie point system as what the FX-100 uses too, so probably they're as good for the same reason, but they don't seem to have a good second-hand supply like the Metcal/OKIs (I scored an OKI power supply for under $100, but even if you don't get lucky they're relatively cheap).

Also if you're into that kind of thing, the MX-500P-11 at least has schematics of the entire power supply that have been reverse-engineered (notably it appears to contain no microcontroller).