There are many ways you can learn new tricks to improve your use of AI. From podcasts with guys like Boris Cherny, Andrej Karpathy or other AI figures, videos on YT from respectable channels (think guys like Jeff Su or Nate Herk for beginner/intermediate level, or IBM Technologies for advanced users), to countless reels on Instagram. All of them are good, all can give you inspiration and some validated ideas you can apply in your daily use or at work. But have you tried going for a personalized advice from your all-purpose assistant?
Change the way you think about working with your chatbot
You have to be very flexible in your approach to AI. Sometimes you have to treat it like an intern who requires support, proper handover and constant review of their work. Other times it will be your oracle, having access to knowledge from any field you can imagine. The key is to understand when and to what extent you can trust and what approach to apply for any given task. This way you realize you should no longer say “I don’t know how to do X”, or “I don’t have the skills to do Y”, because somewhere in its training, your AI model was probably already taught whatever you may want to ask it.
The very same thing applies to the use of AI itself. Do you not know how to use AI effectively? How to write requests in an optimal way or design the flow of your work to maximize the quality of the output? Instead of searching for the right YT video to learn that, why don’t you just spend some time with your chatbot and learn it yourself? And if you don’t have a specific skill in mind, but you feel you are missing something and want to level up, why don’t you do exactly that?
Here’s how to do it
I came up with this exercise relatively early in my AI journey. I remember realizing, if this can teach me how the body generates energy through Citric Acid Cycle, it can for sure teach me what are the elements of an effective prompt. It can tell me what types of tasks I can or can’t ask it to do. But what about the elements I don’t even know I should ask about because my knowledge is too narrow? Well, let’s investigate that too!
I’ve performed this couple times already, and each time I was very happy with the outcome. I have also improved it with each iteration and now I standardized it for anyone to use. You are free to try it as it is, adjust it, or use the concept to build something of your own. The idea is very simple: if I currently use AI this way, and this is my actual level of knowledge, what are the next immediate elements I can learn to be better at it – aligned to my goals and things I use AI for.