At Poly API we were one of the first Alpha users of ChatGPT plugins and today I personally believe we are pushing the cutting edge R&D on enabling enterprises to rapidly adopt ChatGPT plugins and soon to be many more models with the appropriate level of governance.
In this demo, I show a ChatGPT plugin, backed by PolyAPI, which dynamically learns new APIs through observation (today via Postman), and is then able to execute them in real time (without needing to rebuild the plugin manifest file or OpenAI specification). Rather than enumerating the plugin in a static fashion as recommended by OpenAI, we use backend services to dynamically get the data we need to be able to execute any arbitrary API call or Poly custom function.
I am not 100% sold if chat/voice AI agents will be able to fully replace UIs in all circumstances, but I’m confident it will be many UIs for many use cases.
Lastly I demonstrate the ability to have this ChatGPT plugin generate another ChatGPT plugin… hence the name Skynet.
As I mention at the end of the video, if you are an enterprise eager to learn this technology along with us and take it to market as a first mover, please don’t hesitate to reach out.