I work in finance. When I read about synergies, I do not need help.
A few weeks ago I was reading about RAG. It is a way of giving an AI model access to documents it was never trained on. I read the proper explanation twice. I could parse every word and still not hold the idea.
Then I read the casual version. It said the model is not remembering the document, it is being handed the document. I understood it, and then the proper explanation made sense too.
We treat casual language as a concession, the version you hand someone who cannot manage the real thing. That is not what happened. It was not simpler and it was not a comparison. It stated the same mechanism in the words a person would use out loud, and that was enough.
Why the first one did not work for me
Nothing was wrong with the first explanation. It was accurate and well written. It was written for someone who already knew a few things I did not.
That is not a flaw. The writer could not have known what I knew about AI, on that day, about that one idea. So they guessed. Everyone writing an explanation has to guess.
A dictionary definition guesses you want speed. A long article guesses you have time and want to spend it here. Explain it simply guesses you want less. Each guess is right for somebody and wrong for everybody else.
Two things are being chosen for you
The obvious one is depth. How much detail you get.
The one nobody talks about is voice. Formal or casual.
Those two get welded together everywhere. Ask for a simpler explanation and you get a friendlier tone with less information, as though you asked for both. You might want the full mechanism in plain, casual language.
Readers should be able to choose the depth and the voice. That is what we built with Clicked.
How it works
Every term gets explained at three depths. Overview is a few sentences. Detail is the full mechanism. Analogy is a familiar situation with the same shape.
You get one card at a time. The other two are there if you want them, and you move between them however you like. You are not scrolling past two explanations you did not want to reach the one you did.
You can also set which depth opens first. Explanations start at Overview unless you tell them otherwise. While reading something, if you find yourself tapping through to Detail every time, pin Detail with the pills at the top. You skip the Overview, and every explanation after that opens at Detail. Same for Analogy. Change or remove the pin with the same pills, and you can still open any card whenever you want.
Every term also gets a slang version of all three. Same information, different voice. That part is harder than it sounds. If the casual version just says the same thing in casual words, it adds nothing. It has to anchor the idea in the way people actually talk.
Take synergies. The overview says it is the extra value two companies expect to create by merging. They sell more together than they could apart, and cut costs they were each paying separately.
The slang version of the same card says synergy is corporate for two things mashed together beating the sum of their parts. Basically one plus one equals three. In practice it is the word executives use to justify a merger. Then everyone discovers one plus one equals one and half the team gets restructured.
Same depth. Same information. One of them will be the one that works for you, and I cannot tell you which. Nor can I tell you next week, on a different term. That is why the choice sits with you. We call it the Clicked way, mostly so we have something short to say.
Where it lives
Clicked is a browser extension for Chrome and Edge. Highlight anything on a page or a PDF and it explains it there, at whichever depth you want.
The gallery at tryclicked.app has 121 terms already written this way, in four languages, free to read.
If you write explanations for anyone, the useful question is not whether yours is good. It is whether you are guessing at a reader you cannot see.
Top comments (2)
separating depth from voice is a useful design choice. i would test each card with a short recall task or one transfer question, then compare errors across overview, detail, and analogy. that can show whether a casual version keeps the same meaning instead of only feeling easier. it may also help to show the source term and a way to report a mismatch between versions.
Worth clarifying: the extension generates each explanation live from whatever you highlight. So, there is no stored version to compare against. What could help here is user feedback and our prompt quality. The gallery pages are reviewed, which is where version-to-version consistency is actually checkable.
The recall or transfer question point is the interesting one, and it applies to both. "Feels easier" and "conveys the same thing" are different, and we do not currently measure the gap. Anecdotally, what I have noticed is that for topics that are really unfamiliar, reading the slang card first helps with the understanding. I think it just helps develop a mental model faster.
On showing the source term, can you say more about what you have in mind? Not sure I follow.