# Homer's Last Theorem

**boingboing**#1

**LemoUtan**#4

Seems 3987^{12} + 4365^{12} = 4472^{12} + 1211886809373872630985912112862690. Looks quite a long way off to me. Such is the power of powers.

In the Homer Cubed segment of Treehouse of Horror VI, David Cohen slipped in 1782^12 + 1841^12=1922^12 among the equations floating around (which, if true, would disprove Fermat’s last theorem). Cohen has joked about Fermat’s last theorem before.

**Ambiguity**#7

Actually, I have found an error in Wiles proof, but unfortunately the mark-up capabilities of the forums are not advanced enough for me to post it here.

**WearySky**#8

Well, my windows calculator puts the 12th root of 1782^12 + 1841^12 at 1,921.9999999558672254029113283703. So yeah, it’s pretty close, but no cigar. The funny thing about exponentials is that the actual numbers work out to

1782^12 + 1841^12 = 1922^12 - 700,212,234,530,608,691,501,223,040,959

**danchall**#13

I especially enjoyed the part about the near miss. But in trying to follow along using *Mathematica* software which has very powerful numerical capabilities, I found somewhat different numbers. For the 12th root of the sum of squares, you reported 4,472.0000000070576171875. But Mathematica reports 4472.0000000070592907382135292414494, different from your number, starting at the 12th digit to the right of the decimal point. When I take your reported number to the 12th power, it agrees with the sum of squares for the first 14 digits, but not after that. Taking my number to the 12th power results in the precise sum of squares (according to my calculations using Mathematica).

This doesn’t change the validity of a word you’ve written, but it seems that one of us is using a tool that is delivering inaccurate results.