Continuing the discussion from News Story Dump Thread - Stories Only, because discussion is not permitted there.
Some researchers/engineers from the Dessa company made a Text To Speech (TTS) engine which uses the voice of Joe Rogan. Articles about this mainly refer to this by the quality of the voice, but systems like TacoTron2 were already capable of very convincing TTS voices before this.
Is the novelty here that there was no additional recording done by the voice model specifically to make the TTS engine?
I haven’t seen any proof of this, but his initial reactions definitely sound like he was not involved with the team who did this.
Site with real/synthesizes comparison tests: http://fakejoerogan.com/