How To Make Fire Demos with GPT-3

Making fun and easy to read demos for people to consume across social media and the web at large is an important way to get people excited by your GPT-3 projects.


  • Use Carbon to make readable pictures with different font styles
  • Try voice and video! The SpeechSynthesis APIs can be used, or more involved TTS solutions like Sonantic
  • Decrease Noise in text (remove unnecessary writing) and Increase Font Size

You can read the full post here.


Thanks for sharing Bram, great tips!


Great article! So many helpful practical tips – based on experience. The Open AI interview was awesome!!

I read about the SpeechSynthesis Web API a while ago; it’s really come a long way. For those new to the SpeechSynthesis API, be sure to check the demo from the docs to get a sense of its capabilities. The voices have a robot synthetic feel – or edge if you prefer – that actually fits well with GPT-3 output :).