If you know someone who has worked in tech circles for at least a few years, you may have noticed what looks like toys on their desk. Sometimes they are actually toys (to be clear) but in a lot of cases they have a really important purpose.
Quick history then we’ll bring it back - there is a concept called “rubber duck debugging” that first came up in a book in the late ‘90’s.
The idea is to have some kind of inanimate object (in this case a rubber duck) around so that when your code doesn’t work the way that you expect it to, you can go through it line by line and explain it to the duck.
This helps people slow down and get clarity about what they want to happen vs. what they actually asked for in their code.
The idea is really powerful and has expanded to other areas of tech (and possibly other fields).
I think that there could be tremendous value if people who use modern-day AI tools adopted this idea.
If you don’t get what you want from ChatGPT, Gemini, Claude, Hermes, etc., instead of immediately calling it “AI slop” or blaming the models, what if you paused and went back to what you wanted, what you got, and how you can update your approach (prompt, context, skills, etc.)?
If it helps, have something available (like a rubber duck) that you can use to walk through everything. It’s weird at first and then it can really help you work through things.
Some AI tools have an added bonus of actually being able to “talk” back to you. So instead of using the inanimate object(s), you can actually work through it by going back and forth with voice mode.
Food for thought…