It isn't even partially human written, it's just an LLM's output.

As much as I’m confident this reasercher is just a skinsuit for a slop machine I don’t exactly trust pangram. I’ve apparently been flagged for possible AI slop comments based on Pangram results before.

Oddly it seems to have happened on longer posts? Don’t know if that may play a part.

▲

esperent 3 hours ago | parent | next [-]

Isn't pangram more accurate with longer texts? Does it claim to be accurate for comment sized text?

	▲	ofjcihen 3 hours ago \| parent [-]
		[dead]

▲

EA-3167 3 hours ago | parent | prev [-]

It certainly isn’t perfect, especially when it gives a probability below 100%z .