It seems that not only you, but perhaps the video poster you are referencing, are confused on the subject...
"Subtractive EQ" is, as the name suggests, _cutting_ frequencies in a mix, and is generally used to "fix" the sound of something. In a live setting, it is the way an engineer uses EQ to eliminate feedback and "tune" the system to the room it's in. In recording/mixing, it is used to "fix" or optimize a sound in the mix, if you have for instance a squeaky bass drum pedal or a ringing snare that you want to tone down.
"Additive EQ" is used to "enhance" (or "beef up" in some way) a sound, by adding "more" (not really, just accentuating what is already there, often at the expense of other frequencies) of some desirable frequency to achieve more snap, boom, rumble, crispness, or whatever sound characteristic you are trying to achieve. This can also be called "boosting."
Here's where it gets confusing to people. When you are subtracting a frequency, and you have a parametric EQ available (meaning, you have a rotary control for both the frequency and the amount of EQ, respectively), you often use the "sweep, boost and cut" method first. That's right, you have to _boost_ the frequency you want to cut, before you cut it. That's because in order to _find_ the frequency or frequencies that are problematic, you have to sweep thru them first (usually, but not always, mids), in order to identify by ear the freqs that you don't want (or want to de-emphasize). You sweep through the mids, identify where you want to cut, by first boosting it (to hear how bad it really is, and what it is you're trying to eliminate; for example "mud," "honkiness," "box-y" sound, etc.). You find the freqs by sweeping and boosting, then using what's called a narrow Q (on a visual parametric, it will be a cone or triangle shape that you can adjust as thin/tight as possible to really hone-in on the offending freq); _then_ you cut it, by subtracting it or, inverting the Q, so you can visually reference that it has actually been cut... It's the EQ/audio equivalent of a "search and destroy" mission. You have to find it, ID it, check it, then cut it.
THEN, a very important step-- check it _in the mix_. It doesn't help to have an amazing snare drum sound that sounds awesome solo'd, but crappy once you bring all of the other mix elements in.
So, yeah, sweep/boost/cut is a basic technique used by most modern engineers. You should definitely learn and practice it. Depending on the type of EQ you're using, "Subtractive EQ" and "Sweep/Boost/Cut" are the same thing. Or more properly, sweep/boost/cut is a method of subtractive eq (that doesn't work if you're using a graphic eq, because you have less control, and you have to both know your freqs and/or guess more often when notching them out).
GJ