We are getting closer to “Her” where conversation is the new interface. Siri couldn’t do it, so I built an e-mail summarizing feature using #GPT3 and life-like #AI generated voice on iOS. (🔈Audio on to be 🤯with voice realism!) How did I do this? 👇

I used the Gmail API to feed in recent unread e-mails into a prompt and send to the @OpenAI #GPT3 Completion API. Calling out details such as not “just reading them out” and other prompt tweaks gave good results

Here are the settings I used, you can see how #GPT3 does a great job of conversationally summarizing. (For the sake of privacy I made up the e-mails shown in the demo)

The audio model was fine-tuned on speech from the movie Her. I got good results with TorToiSe, but have also experimented with ViTS & YourTTS from @coqui_ai and more recently @ElevenLabs. None are fast enough for a snappy response together with da-vinci-003 completions, so...

It’s crazy to think Amazon and apple are two of the largest tech companies in the world yet interacting with Alexa or Siri is 99% trash

1) this is 🔥🔥 and 2) your emails are a lot more fun than my emails! 😅😅

Is there a simpler tutorial than just his tweet storm on how to build this?

Can you make one to help me choose a Netflix title to watch, cuz that shit gets tedious after 20-30mina of browsing?

Good idea to pre generate audio. What sort of latency are you getting from your GPT-3 call? You might be able to afford real time speech generation if you move to local FLAN-T5-XXL which fits in 8-bit on a 24GB card. I’m moving to this strategy for my own assistant stuff.

We created the torment nexus from the classic tale “don’t create the torment nexus”

Amazing! The prompt engineering you did to get this output is just great; something to learn from!