1 article
Why 0.1 + 0.2 lands on 0.30000000000000004, how IEEE-754 packs a sign, exponent, and mantissa into 64 bits, and why float errors are deterministic, not glitches.