Odaily News Ethereum co-founder published an article titled "Binius: Efficient Proofs of Binary Fields", in which he pointed out that SNARK relies on "arithmeticization", a method of converting statements about programs into equations involving polynomials (sometimes vectors and matrices). In order to keep the numbers within a reasonable size, "arithmetic" must be done not on regular integers, but on structures called "finite fields". Modular arithmetic is the simplest example of a finite field, but there are other examples. In actual programs, most numbers are very small: for loop indexes, True/False values, array indexes, counters... If the field is large, the "extra" values generated during the proof calculation will be much larger, which is a key source of inefficiency. Plonky2 and similar protocols reduce the field size from 256 bits to 64 or 31 bits. But it is more efficient to use binary fields directly. The binary field is a fascinating mathematical structure with many unique properties. The tower structure is a fascinating way of production, which adds more advantages.