CPUs operate on 0's and 1's. All instructions had a sequences of bits that defined them. Thsi was machine code. That hard for humans to read and write. As a first step we grouped the 0's and 1's into 4 bits sequences and used 0 through F to represents them. This reduced the numbers we had to memorize, but it still was not easy to remember the hex code for instructions.
So we started using assembly which had "words" like MOV and ADD. The assembler would replace the instructions with the correct set of 0' and 1's converting the assembly "listing" to machine code.
Eventually we developed "higher" level languages where "instructions" could represent entire sequences of machine code.
