Adopting AI

When ChatGPT came out, I was very reluctant to use the technology. Skepticism about its capability, concern over its capacity to replace human intellect, as well as general confusion on how to use it impaired my interest to experiment with it.

However, a couple months ago I finally found my entry point – having it code for me. After switching around from C#, Java, Python, and NodeJS over the years, I have found my ability to quickly know clean and concise ways to program has fallen well below market rate.

Enter ChatGPT and “prompt engineering” - after some trial and error, I finally gained a basic understanding on how to properly communicate coding tasks to AI. The following communication guidelines were essential to me –

·???????Give context about what you are trying to code

·???????Be very explicit about the input and output that you are going for

·???????Articulate the problems as you would to a fellow software engineer

·???????Break down your problems into pieces if you hit a limit on the length of input / output it can do

·???????Always push the AI to do more and challenge it, while also making mental notes on its limitations

Experimenting with and then committing to this process has helped me significantly in my coding speed and quality.

But it really has not stopped there – coding is just a subset of problem solving. To be more efficient, I find myself constantly evaluating how I spend my time professionally and aggressively analyzing things that AI can do for me.

I think after adopting AI a bit more in my work, I am getting more comfortable with it. The true power of the human mind is based heavily in creativity and novelty, not in repetition and sweat. Once AI starts to create novel things without prompting, that may become a very profound reckoning for humankind. But until then, there is a cutting reality that if everyone else is going to be using this tech, we need to as well to keep up.