When Steve Jobs introduced the first iPhone in 2007, it changed the way we interact with mobile phones. Let's recall the legendary quote from Steve Ballmer, then CEO of Microsoft, who claimed that the iPhone had no keyboard, meaning it was difficult to type emails on it. I remember that I also reacted lukewarmly to the first touchscreen phone. It was only later that I realized that I could surf the internet and consume content much better on a touchscreen display. The iPhone simply changed the rules of the game. And just as users had to learn to work with the first iPhone or search engines, ChatGPT and its future successors represent a new chapter that changes the way we work with data.
Join Martin Smětala for practical tips on working more effectively with ChatGPT - a new, comprehensive tool that looks simple but is quite complex on the inside. If you're interested in a general introduction to ChatGPT, read the article Artificial Intelligence – Humanity's Best Friend or Skynet?
Article Author: Martin Smětala
ChatGPT and Commands (Prompts)
ChatGPT has a really simple interface – a single command line where you can type any text, and ChatGPT will write a response in no time. However, just like with Google, if you type the wrong keywords, you'll get unexpected results.
The input is always your command. And this is where the "catch" lies. For more complex tasks, you may find that the output from ChatGPT is not satisfactory – it can be too general, partially made up, incomplete, not specific enough, written in the wrong style, etc.
In general, the quality of the input is directly proportional to the quality of the output from ChatGPT.
How to Achieve Accurate Results When Working with ChatGPT?
Don't expect results based on just one command. ChatGPT needs to "prepare" similarly to how a developer needs to set up their development environment before they start developing.