Understanding is a feature.
Good explanations and interfaces do not hide complexity; they make the next meaningful step understandable.
My small ELI5 tool grew from a simple observation: when I cannot explain something clearly, I often have not fully understood it myself. Simplifying is therefore not work that begins after the “real” technology is finished. It is part of thinking.
Simplifying exposes gaps
A good explanation does more than remove specialist terms. It finds the core, orders relationships and makes assumptions visible. This reveals contradictions that complex language can easily hide. What exactly happens here? Why is this step necessary? Which exception changes the outcome?
In projects, I often take the role between domain expertise, technology and use. Each side holds important knowledge but speaks a different language. My job is not to make details disappear. I have to translate them so that decisions can be made together. This does not only help the recipients. It also sharpens my own solution.
Interfaces are explanations
An interface explains a system through sequence, terminology, feedback and available actions. If users first have to understand the internal architecture to complete a task, the product has delegated part of its thinking to them. A technical error message may describe the failure accurately and still offer no useful way forward.
With somnara in particular, a plausible-sounding text is not enough. People need to understand that an AI interpretation offers a perspective, not a truth. Limits, uncertainty and ways to correct therefore belong in the interface.
Understanding is a feature. Good products do not expose every piece of existing complexity. They show exactly the information needed for the next meaningful step. This is not simplifying the matter at the expense of truth. It is the work of making truth usable.