From time to time, when Claude’s explanations got too complex, I found myself asking it to simplify. The first prompts I tried were the ones that go around among engineers, “explain it to me like I’m an idiot” or “explain it to me like I’m 5 years old”, and both of them overshoot: they do skip the technical details, but they throw away so much of the meaning along with them that what is left is too naive to be of any use. Besides, they made me feel like an idiot or a child – a proof, I guess, that they were successful at delivering what I asked, but not what I really wanted.

Finally, the one prompt that consistently worked was “explain it to me like I’m a smart teenager”, which drops the jargon while still treating the problem seriously, without underestimating your mental capacity.

Stripping out all the technical details and keeping the core meaning turned out to be useful on more levels than I expected. It lets me take a step back and see the broader picture of what I’m actually trying to achieve, it puts the problem in a different perspective, it often helps me identify the right technical questions to ask as a follow-up and, on top of that, it makes it much easier to explain the thing to someone with a non-technical background.

I didn’t use it very often before, but, eventually, I thought I’d turn it into a ridiculously simple /layman skill, which in Claude Code is nothing more than a markdown file with a few instructions in it, that you invoke by name (a name inspired by the phrase “in layman’s terms“).

I already had a /tldr skill in the same repo, and the two sound like they do the same job, but they don’t. /tldr keeps the conclusion and drops everything under it, in the same technical language, just with far fewer words; /layman keeps the whole thing and changes the language instead. One shortens, the other translates, and I reach for them on completely different occasions: /tldr when an answer is long and I only need its point, /layman when an answer is dense and I need to actually understand it.

Ever since, I use it all the time, even on tasks I fully understand, because of the fresh perspective it often gives.