The bulk parser only needs to read a single character and only needs to care about two characters (brackets). The sub parser then looks for either {...} or a single character.

Markdown requires all sorts of complexity and it's nearly undecideable. Before CommonMark there wasn't even a spec and many things were largely ambiguous.

Also, tables and other structured data in markdown are almost entirely hacks