There are no stupid questions (with AI)
There are a lot of stupid questions that never get asked. Personally, I've found it's easier to ask the stupid question when the audience knows I'm new to the topic/area. In those instances, I try to ask all my stupid questions because the longer I wait the more I'll fear asking them when I'm no longer as new.
I know I'm not alone in this - I've seen many people over the years struggle with some concept because they never established the basics and now it's too late. And, often, it's more obvious than they realize that they clearly haven't grasped the basics. It's probably just as obvious when I'm struggling on a topic too.
So... how do ask the stupid question when it's too late to ask the stupid question?
I like holding "No Stupid Questions" meetings where people can ask any question (anonymously, if they so choose). Sometimes I, or other experts, would ask the stupid question - not because we didn't know the answer, but because we knew others would benefit from hearing it answered.
The tricky part with No Stupid Question meetings is holding them regularly while keeping them relevant. Teams change. People come and go. And if you always cover the same stupid questions, it becomes less valuable to the people who have other and newer stupid questions.
Also - not everyone runs a No Stupid Question meeting... so what then?
I've been sitting on some stupid questions for some time. My plan was to ask someone very specific - someone I trust who won't think I'm stupid. But the timing was never right, and so I was sitting on the stupid questions for longer and longer, at my own detriment.
Enter Copilot. I know, everything is about AI these days - but I realized one entirely unique aspect of Copilot (and other LLMs) is that they won't judge you (yet). I asked Copilot my stupid questions. It answered and only gave me a surface level understanding - so I asked for it to elaborate. I asked for it to give me an example. I asked for it to compare it to something I did know and understand really well. I then played the devil's advocate with it and challenged the arguments it was providing. Finally, I took everything I'd internalized from our conversation, and I put it in my own words and asked Copilot if I missed anything. It let me know I was about 90% there but missed a couple of important points. Not because I misunderstood but (and this was huge) simply because we hadn't covered it in our conversation.
Of course, this only works for information that's readily available and not stuck deep inside someone's head. There's a lot of knowledge contained within a team that isn't accessible to an LLM (again: yet).
For those times, I think it's worth asking your leadership the not-so-stupid question: "Why don't we have No Stupid Question meetings?"