They all go HI as a RESULT of the first clock pulse!The way that his counter is drawn is that he is clocking the 'C' bistable on the rising positive edge of the clock pulse.
If you check the image he has posted of the C,B,A states they are all high on the first clock pulse.!
Look at his sim results. Before the first rising edge all three FFs are in the LO state. So they WERE initialized somehow -- they are NOT in an indeterminant state!
So the counter starts out in State 0. Then on the first rising clock edge the C FF toggles from LO to HI. But since the Q output of this FF is used as the clock input of the B FF, it also sees a rising clock edge and so it, too, toggles from LO to HI. The same thing happens when the Q output of the B FF clocks the A FF. So the counter goes from State 0 to State 7, which is a count backwards. This behavior has absolutely nothing to do with reset it issues, it is a direct result of how the counter is wired.