Rendered at 02:56:17 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
ricardobeat 5 hours ago [-]
Don't waste your time - the 'meltdown' was the underlying LLM starting to speak Chinese while being interviewed.
Their character has slow reactions, seems quite dumb and has no personality at all. They are probably running an old version of ChatGPT 4.x, plus some standard TTS/STT pipeline. There is nothing state of the art there, and a bad look for AI in general. It's 100% a (poor) marketing play and quite surprising it's been going on for so long.
maplethorpe 51 minutes ago [-]
Yep. It's worth checking out GPT-6 to see what state of the art actually looks like.
It won't get the same level of attention, but the question prior to the lapse into Chinese was pretty damning as well. It wasn't a difficult question at all but she didn't understand or answer it correctly until it was asked a second time.
Their character has slow reactions, seems quite dumb and has no personality at all. They are probably running an old version of ChatGPT 4.x, plus some standard TTS/STT pipeline. There is nothing state of the art there, and a bad look for AI in general. It's 100% a (poor) marketing play and quite surprising it's been going on for so long.
https://m.youtube.com/shorts/sJl1KoSL_o8