If you dump it as hex, you can see the difference. I just ran it through xxd.
The first sentence:
0000000: 7065 d0be 706c 6520 696e 2074 68d0 b520 pe..ple in th..
0000010: 77d0 be72 6c64 0a w..rld.
The second sentence:
0000000: 7065 6f70 6c65 2069 6e20 7468 6520 776f people in the wo
0000010: 726c 640a rld.
This one appears to be valid ASCII.
Both look very similar, but instead the first sentence uses the unicode character set and uses Cyrillic characters instead of normal ASCII ones.