The square is computed by code, not by a neural net
Code computes the square: it’s pure arithmetic over the date, impossible to get wrong. The neural net is responsible only for the explanation — it reads the finished table and writes the interpretation in words. It cannot invent a digit that isn’t in a cell.
Why this matters
A language model, simply asked to «compute a Pythagorean square», regularly slips on arithmetic: it doesn’t calculate, it reproduces something similar. That’s exactly why numerology in an ordinary chatbot is unreliable.
In Viki this layer is separated. First the code sorts the digits into cells; then the model receives the finished table and explains it.
What the code computes
- Splitting the date into digits.
- The four additional numbers by fixed rules.
- Distributing all digits across the nine cells.
- The life path number and the personal year.
What the model does
Reads the table, remembers your history and writes the reading: what the cells mean in your combination, how it connects to what you discussed before, and what to do about it.
Where uncertainty remains
Not in the arithmetic — in the interpretation. Schools differ on cell meanings and on the additional-number rules. Viki relies on Victoria Rommer’s methods — she checks and refines them personally.
You can verify
Ask for the calculation step by step and redo it on paper. There’s a separate guide on that.
Ask for the steps — recount on paper.