😭vas a tener stand en algun otro evento a parte de la indie dev day? Lo quieroooo

I haven't been able to open a single user profile all day, it keeps saying they don't exist

i see what you mean, you're right.
Most compression libraries allow for a level compression setting. The higher, the more compressed it ends up, but the slower it runs; with the default setting usually being a middle ground. Perhaps a higher compression level might have done what you're pointing at.

We could achieve the same result by changing the last two operations to
x6 <-31; x6 <-6
or
x1 <- 27; x11 <-1
too, btw. I'm too lazy to calc it, but since numbers are encoded in a way that uses less bits the smaller the number is; these options might be less efficient [4/4]

So what we do is
x6 <-31 - go back thirty-one chars and copy six chars in a row
And now that we have 'p' at the end of the buffer, we can use "the trick"
x6 <- 1 - go back one character and copy six
[3/4]

so, even though it has only decoded one 'p' so far, it's cloned 5 times.
o - add a literal 'o'
x24 <- - go back one char and copy 24 chars
it's the same as before

Now, we want to encode 12 'p' chars. The longest sequence is 6 'p' long, and we can't use the trick because the last char is a 'o' [2/4]

zip works by saying "go back N characters and copy M characters in a row", but it has a little quirk where an instruction can copy from the characters that have just been created by that same instruction.
So, in this case:
p - add a literal 'p'
x5 <-1 - go back one char and copy 5 chars [1/4]