Dev unsettled by dessa's voice samples sounding more "alive" than real people
IndraVahan · x · 2026-10-26
IndraVahan says dessa's voice model samples, trained on recordings of real people, sound more "inhabited" than hours of real human conversation, raising unease about how human synthetic voices have become.
More from Fun
- Using Gemini 4 to generate custom 'mate in N' chess puzzles — fofrAI · 2026-10-26
- Voice models sounding too human called 'the most braindead direction in AI' — yungcontent · 2026-10-26
- '3T-parameter model runs on your 12GB GPU': quantization community parody — g3off_record · 2026-10-26
- Cybertruck license plate "EGOKING" is peak AI-industry self-knowledge — NathanpmYoung · 2026-10-26
- Muting is social media's most underrated superpower, says viral post — signulll · 2026-10-26
- Professor catches student cheating with AI who left the prompts in the paper — venturetwins · 2026-10-26