Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 8 Next »

So kind of crazy seeing the implosion of OpenAI over the weekend - what a meteoric rise and fall. I feel for the employees there.

Never mind - you can still do some useful things with ChatGPT and Iguana. I wrote these AI programs as tools which can be helpful in addressing certain types of bottlenecks in starting to write code.

And you can do it without compromising patient health information (PHI). This is what I have found to be successful.

 Use an AI which can read the web

I have been using a paid ChatGPT account so I can access to ChatGPT 4.0 which with bing can apparently browse the web.

It means I can give it prompts which can take it’s generic knowledge of APIs and Lua and then give it a primer to correct some of mistakes it makes with our apis.

It’s bit like programming - except that the language is in English!

 Tightly control what information I send

By using prompts I can tightly control what information I send the AI and just control what data I send to it.

It just like IguanaX - very transparent and clear about what it is doing. You are in control and in the drivers seat.

 Example ask to build code to get a contact in pipedrive

Pipedrive is a nice customer relationship management platform which is really handy for managing communications with lots of people.

I used this prompt with ChatGPT:

Read interfaceware.atlassian.net/wiki/spaces/IXB/pages/2680422530 and then write lua code to extract a contact record from Pipedrive given an contact ID using Lua in Iguana.

  • No labels