Thunderf00t created that video in the following way;
- he wrote a script and then recorded his voice separately as a audio file
- After that, he probably edited that audio file to get rid of any mistakes, and mastered it to make it sound nice.
- He then imported that audio file to a video editor
- In the video editor he added 3rd party images and video to fit his voice over
The video editing process is where the magic happens and is what makes the video so slinky. There are many how-to videos on youtube on the video editing proccess.
That being said, what you talk about in your question, reminds me of what TJ is doing on his drunken peasants channel/podcast where they play a video and pause it at curtain points to discuss whatever was in the video - all of this done in real time. In order to do that, you need to screen capture the video along with the audio of the video accompanied with your voice.