I used to rely heavily on Google to search for answers to technical problems. However, since the advent of ChatGPT, it has significantly boosted my efficiency in solving problems and learning new technical skills: ChatGPT. In this post, I will explain how I’ve integrated ChatGPT into my workflow and the steps I follow to approach solutions using this AI assistant.
Step 1: Question Definition
The first step is to determine whether or not to use ChatGPT to solve a particular problem. I use the following matrix to make this decision:
-
Syntax in Programming Languages
While I can use Google to look up syntax, I often find it more convenient to use ChatGPT. The ChatGPT tab is always open, so I simply type in the keyword to search for the syntax. ChatGPT is also useful for refreshing my memory; I can ask follow-up questions, request summaries, or ask for comparisons to similar syntax.
-
Confidence
I turn to ChatGPT for technical skills that I’m not very confident in, such as frontend techniques or DevOps orchestration tools. ChatGPT provides explanations and guidance that help bridge my knowledge gaps.
-
Documentation Writing
As an ESL (English as a Second Language) software engineer, I use ChatGPT extensively for writing emails and documentation. The refined output is often more polished and professional.
Step 2: Evaluate the Answers
The answers from ChatGPT are not always perfect. Sometimes this is due to the context provided in the question being either too much or too little. Since the answers are generated probabilistically, it’s essential to fact-check them based on my experience.
-
Assessing Credibility
For questions with straightforward context, I tend to trust ChatGPT’s answers more. For complex queries, I rely on the answers for the big picture but verify the details, especially code snippets, by cross-referencing with Google.
-
Follow-Up Questions
If I’m in doubt about an answer, I usually ask more specific follow-up questions. If ChatGPT acknowledges a mistake, that’s when I decide to look up the solution on Google.
Step 3: Learning from the Answers
After finding a solution, I summarize what I’ve learned from the problem. I take notes on what I should study further and often consider writing a blog post about it.
You know what? This blog post was initially drafted by me and then refined by ChatGPT. Check the difference here. 😎
Bao-Jhih Shao
A software engineer writing something to keep the memory.
Comments