the voice itself, when I first demoed it, had a lot of natural voice / neural voice markers like breathing, stopping, starting, emotional impact.. in today's demo, it had none of this. it sounded like a significantly simpler voice, it was reading cues like "spoken in a quiet voice" out loud, it failed to stop at natural stops and jumped the gun on the next sentence, just generally making last-gen TTS mistakes. It sounded like any TTS from 2020, more or less. For the content itself, it completely rejected any question I had about VAPI and especially the VAPI website, whereas my previous interaction with it gave tons of really in-depth and accurate information about VAPI, the website, how to integrate with it, and basically let me interview it with technical questions. it was a fantastic ambassador for your product. today's experience was akin to asking chatgpt the same questions. it felt completely clueless and generic.