I always like it when someone is able to take a topic that some people would consider boring or too technical, and find a way to make it applicable to real situations in just about any field.

This guy presents an algorithym designed to create the optimal filter for a photo on the moon, but goes on to relate it to everything from genetic sequencing to his own love life.

I find it ironic that the entire time he's going on about the need for an outside source to evaluate each outcome of the algorithym to determine the fitness of the data sets to serve a given purpose. Without this outside critical analysis, the process would be meaningless - just a program spitting out numbers. The function first needs a specified purpose, as well as a set of standards upon which to judge its suitability for the purpose and give feedback. 

The speaker states that he evaluates data fitness based on "no complaints" - zero complaints means the process is optimally efficient, hence no need to change or further optimize it. However, the entire reason he is there is to recieve constructive cricism. He's practically begging the audience to complain, throw a tomato at him, anything to help him optimize his algorithym. But, he only recieves more questions. 

Hopefully he's able to optimize his marriage. :)