pull down to refresh

Whilst there is a lot of hype around ChatGPT and other AI chat interfaces - the progress over the last year is also incredibly beneficial for more 'hidden' AI based features like the transcripts in Fountain.

6 months ago - producing transcripts in Fountain was:

  • 2x more expensive
  • 8x slower
  • much less accurate

this trend is only accelerating and cheap AI based features will be easier to offer

awesome, what do you think about allowing users to generate podcasts directly in the app by inputting a blog post or written document?

is AI good enough today to generate realistic, expressive voices for long-form content that sound better than Siri or Alexa?

reply

Yes this is definitely possible - the benefit of the audio medium is that you can listen in your dead time - commuting, chores, etc and this benefit could apply to audio sourced from anywhere

reply

What ASR engine / service do you use?

reply

We use Deepgram - https://deepgram.com

We'll be releasing a blog post today on how transcripts on Fountain work in detail - will share when it's out!