I find it's not the frequency of these words, but the volume and inflection that's the problem. There are a series of computer science lectures on youtube, and also some courses I had at university where the lecturer says UMMMMM at a substantially higher volume than the rest of their speech, and drags it out for a second or more.

The resultant sound is like a cow mooing, or someone doing an offensive impersonation of a downs syndrome patient at the top of their voice. The effect is totally unbearable and I have no idea how people sit through hours of this stuff.