LeetDezine Everyone can describe a URL shortener. Write a row to the DB, generate a short code, cache it on reads. The base design fits on a napkin. The interesting part is what happens when you push on any one of those steps. Where does it break? Why? And what's the fix that actually holds at scale? Here are four traps I've seen candidates walk straight into — each one looks correct on the surface. 1. The Truncation Trap A Snowflake ID is 64 bits. You only need 36 bits to cover 50 billion URLs (2^36 = 68 billion). Encode 36 bits in base62 and you get exactly 6 characters. Clean short code, no collision check needed. The natural move: drop the rightmost 28 bits, encode what's left. You're keeping the timestamp, which feels like the important part.…