It’s a broad generalization, but it’s not really a matter of opinion. We can scan people’s mouths and faces when they talk (and have in order to demonstrate this stuff). I think the last example probably only applies that way in particular circumstances though, since English speakers automatically group, contract, and arrange certain phonemes in certain orders (e.g., I’m not, I ain’t, but never I amn’t–and in real speech “I ain’t” is almost always one syllable). In this example, more frequently my country ass contracts the first syllable of “gonna” away instead of the second, so “I’m 'na head to the store; y’all need anything?”

The hot potato example just stands for the premise that in real speech the t at the end of hot and the p at the beginning of potato slur together, and if you deliberately enunciate both consonants, you sound like you’re reading to a transcriber. Compare the way a normal person says “let’s go” to the way you sound if you force separate the words: you sound like you’re doing a Mario impression.