new | show | ask | jobs Github

microtonal 11 hours ago

/e/OS speech to text uploads your speech to OpenAI. (I think IA was a typo.)

▲

_ache_ 11 hours ago | parent | next [-]

Yeah, it's a typo (I'm french speaking, AI is IA in french and sometime I type it in french instead of english).

▲

chrisjj 10 hours ago | parent | prev [-]

Ugh. Thanks. Hard pass here.

▲

seanhunter 7 hours ago | parent [-]

Reading the links posted in a sibling thread it only does it if you have text to speech enabled and they use an anonymizing proxy so openai can't associate sessions with any particular user ie it's not perfectly anonymous and private but I don't see how you could have totally anonymous and private until you have a fully offline on-device TTS model, which the fairphone guy said they tried and didn't feel it was up to scratch.

I don't use e/os but it doesnt' seem like a terrible compromise to me personally.

	▲	chrisjj 5 hours ago \| parent [-]
		It is not even imperfectly private. Every word gets heard by a partner of a kakistocratic foriegn regime. > don't see how you could have totally anonymous and private until you have a fully offline on-device TTS model Yes, and? PCs that have have had that for decades - despite orders of magnitude less platform capability.