They don’t, this number is extremely misleading. Roughly 97% (I think is the current estimate) of their protein-coding DNA is equivalent, but not 1:1 identical.
Not the same kind of 99%..
For (fraternal) siblings you'll get around 50% or thereabouts.
No, that’s also incorrect. The real number is way higher. They share 50% of alleles, and hence 50% of variants. But a lot of the DNA within two distinct alleles will be identical, except for (say) a single nucleotide difference.
Yes, that's actually what I meant to say.. couldn't remember the right terminology at the time. The point stands though, I think.. the number "99.6%" for identical twins isn't measuring the same as "98.5%" (or so) for humans vs chimps.