Sometimes ChatGPT and similar tools provide incorrect answers to simple logic or maths problems. However, this is easily remedied by guiding the AI to solve things step by step.
The simplest way to do this is to add a phrase such as 'show the solution step by step' to your prompt. Another approach is to break the problem down into smaller parts and ask the AI to tackle them one at a time.
I learnt this the hard way when I tried to write a small program using ChatGPT. As I had no coding experience, I naturally worked step by step, starting with a basic prompt and then asking the AI to add, fix or adjust things one piece at a time. And it worked!
Later, I decided to combine all the prompts into one big request to save for the future. However, when I gave this to ChatGPT for testing, the result was nonsense.
Here’s what I learned:
Don't give employees big, complex tasks straight away — most people aren't much smarter than ChatGPT.
Don't try to tackle a huge problem all at once by yourself — you probably aren't capable of that either.
And don't try to build a massive start-up straight away. Because your flaws, combined with those of your team, will only multiply.
The first version of your start-up should be as simple as possible.
Take it step by step — that’s how progress is made.