Add the fifth powers of 27, 84, 110 and 133 and the screen prints 6.1917364224⋅10+10. Raise 144 to the fifth and the screen prints the same digits. Subtract one from the other and the screen prints 0 — which is the whole counterexample, stated three times. The number is 61 917 364 224, and the largest term, 133⁵, is 41 615 795 893 of it, so the four addends are genuinely unequal and none of them is doing the work alone.
Euler had written in 1769 that whereas a square can be a sum of two squares and a cube cannot be a sum of two cubes, a k-th power needs at least k k-th powers to make one. For cubes he was on safe ground — two cubes never make a cube, which is the exponent-3 case of Fermat's last theorem, and three can, as 3³ + 4³ + 5³ = 6³ shows. For k = 5 the claim stood for two centuries until Lander and Parkin tried the obvious thing and searched. Four years later Elkies produced a counterexample for k = 4 by an elliptic-curve argument rather than a search, with numbers in the tens of millions, which is the usual shape of these stories: the conjecture fails, and the smallest failure is far larger than anyone who only writes down small cases will find.
Answer
6.1917364224⋅10+10
How to read it
The sum of the four fifth powers on the workbench. The second link is 144⁵, and the third subtracts one from the other and prints 0.
Source
Leonhard Euler's sum of powers conjecture of 1769; the counterexample for fifth powers was found by L. J. Lander and T. R. Parkin in 1966 by direct search, and Noam Elkies found one for fourth powers in 1988
Work it in the calculator
These links open the right calculator with this problem's numbers already typed in. The line above is what the screen should show — check it yourself.