# Why can't Bernstein-Vazirani algorithm be classicaly computed in one try?

Instead of using n inputs (000...01, 000...10, ..., 100...00) and multiplying each one with the hidden number x, why can't we just try one input: 111...11 ?

I can't think of an example of a binary number x which we couldn't discover by multiplying it with only 111...11. I assume I don't understand the problem correctly, but it is how I perceived it after watching Qiskit videos on Youtube.

If you input $$1111\ldots 1$$ to the oracle for the Bernstein-Vazirani problem, you get a 1 bit output which just evaluates $$111\ldots 1\cdot s$$ (where you're trying to learn $$s$$). In other words, you only learn whether the number of 1s in $$s$$ is even or odd. That is nowhere near enough to reconstruct the whole of $$s$$.
Indeed, this is the start of the argument for the classical lower bound: to learn $$s$$ I need to find out $$n$$ bits of information. Each use of the oracle can give me at most 1 bit. Thus, I must use the oracle at least $$n$$ times.