Hacker News new | past | comments | ask | show | jobs | submit login

I've been collecting examples of the abuse and misunderstanding of the no-free-lunch theorems, and this is a pretty flagrant one:

> "Any simple major enhancement to human intelligence is a net evolutionary disadvantage." The lesson is that Mother Nature know best. Or alternately, TANSTAAFL: there ain’t no such thing as a free lunch.




Why do you think that is wrong?


The NFL theorem (for search and optimisation) states that all algorithms for searching a fixed problem space perform the same, in expectation, when you don't know anything about the fitness (cost) structure over that space. There is a tradeoff in a sense: an algorithm can only do better than random search on some instances of the problem if it does worse on others. There's no sense of a tradeoff between multiple objectives on a single instance of a search problem, which is what gwern is talking about.


I believe it's actually a reference to the general expression "There ain't no such thing as a free lunch" and has little to do with the computer science theorem of a similar name.

(https://en.wikipedia.org/wiki/There_ain't_no_such_thing_as_a...)


Yes, but Gwern links to the wiki page on the NFL theorems.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: