Not wanting to be the 'Old Man Yells At Cloud' figure, but this is a perfect illustration of how not to do it, particularly multithreading in Go before using a reasonable algorithm.

Searching for a long string like this is preposterously easy. You can use a classical algorithm, or, given you are looking for a 64 byte long string, you could do a strided inspection - every 32 bytes, look to see if you have a 4-gram from the first 35 bytes of the string. More or less anything works as almost any reasonable string searching mechanism can outrun memory bandwidth.

Or you could go use my old project (https://github.com/intel/hyperscan) and then you could look for regexes too. :-)