This is an interesting observation. I wonder if it's because, on Paul Graham's disagreement hierarchy [1], inline annotations strongly encourage DH3 responses, make it challenging to write DH4-5 (because of lack of space, and interruption of the flow of the main piece), and outright make it impossible to write a DH6 response.
That doesn't necessarily make them bad - often you can provide useful elaboration or contradict factual assertions - but I think you're right that they do have a petty vibe to them.
Ah, this is great. I had forgotten about this essay. I think this part is particularly relevant:
> Even as high as DH5 we still sometimes see deliberate dishonesty, as when someone picks out minor points of an argument and refutes those. Sometimes the spirit in which this is done makes it more of a sophisticated form of ad hominem than actual refutation. For example, correcting someone's grammar, or harping on minor mistakes in names or numbers. Unless the opposing argument actually depends on such things, the only purpose of correcting them is to discredit one's opponent.
It's really easy to end up doing this when replying inline.
That doesn't necessarily make them bad - often you can provide useful elaboration or contradict factual assertions - but I think you're right that they do have a petty vibe to them.
[1] http://www.paulgraham.com/disagree.html