Any of the two threads can utilize the entire core to process something, there's no main thread and secondary thread. If you assume that threads 0 and 4 belong to core 0, any of the two threads can utilize the entire core 0 to do something as long as the OS doesn't schedule something to be done at the other thread of the same core. If this happens the core will have to balance between the two which will cause a performance loss while you could have avoided this if you threw the other thing at a different logical thread that would belong to a different physical core.
If you still didn't understand then just go watch that Lake's video again or read the Wikipedia article which is surprisingly correct.