Working with ChatGPT is all about asking the right questions
As a data scientist, I use ChatGPT daily. It’s great for:
1) Translation
I have found ChatGPT to be very helpful at translating languages I am fluent in (English or Python code) into languages that I am less comfortable with (for example, HTML). I recently used ChatGPT to help me write code in Apple’s Script Editor app to automate composing monthly report emails to over 20 client companies. This included looking up a distribution list, adding the cc’s, writing the subject and body of the email, and adding the correct attachment. I was able to do this without learning any of the syntax of Script Editor; rather, I worked with ChatGPT to translate specific requests.
2) Positivity and confidence
As someone who struggles with self-doubt, I’ve found ChatGPT’s “fake it ‘til you make it” confidence refreshing. “Can you write this complicated function in this language”? ChatGPT will always cheerfully output an answer, unbothered by whether it’s correct or not.
3) Rubber-ducking
You know those times when you spend hours working on a problem by yourself, only to find the answer staring at you 30 seconds into explaining the problem to a colleague? I am currently reading a book called the Pragmatic Programmer, which advises talking to a rubber duck if no human is available. I’ve found ChatGPT to be a rather effective rubber duck.
The one thing ChatGPT will not do is tell you the right questions to ask.
I used to work as part of the support team for a financial app, and 80% of my job was knowing how to clarify and rephrase the client’s questions. Knowing how to guide someone from a series of vague generalities down to a specific and narrow issue to be solved is an important skill. Once you have the right question, finding the answer can be trivial.
The one thing ChatGPT will not do is tell you the right questions to ask
So, why do I find ChatGPT helpful?
Without the burden of reading through vast amounts of documentation to find the syntax for a particular command, I am free to focus on defining the right problem and breaking it into small logical chunks to be solved.
Like the computer in The Hitchhiker’s Guide to the Galaxy who gave 42 as the answer to “life, the universe, and everything”, what you really need to know is “what is the question”?