The linked answer on SE basically says statistically there's a near-100% chance that there's a string where md5(string) is 0xfffffff... but there's no concrete

The linked answer on SE basically says statistically there's a near-100% chance that there's a string where md5(string) is 0xfffffff... but there's no concrete proof. Which is basically the same thing the GP says.