Because apartment hunting is shockingly difficult when you realize they lack any camouflage to speak of, and are, like, completely incapable of outrunning you, AL mitigates your misery by recommending digs (from a growing catalog of 1.4mil) using a bizarre-but-apparently-effective proprietary algorithm. Monitoring user interaction (views, likes, saves...), the increasingly intelligent search matches your results to those of other users based on answers to multi-choice questions, which range from your taste in TV, to your ideal superpower, which right about now is probably "apartment having". Although it'll feed off of user interaction to return smarter results, here's how things currently play out for a search in NYC:
Question: Do you use Twitter? Answer: #isthatarealquestion? Result: Increased the number of pads that were westerly exposed or "sun-drenched", as good lighting is essential when snapping Twitpics of you doing stuff no one cares about.
Question: How many books do you read in a month? Answer: One or less Result: Most of the "Good" matches were in traditionally ethnic 'hoods, probably because homeland authors like Rushdie and Tolstoy write books that take a whole calendar year to finish.
Question: What's your favorite '80s television show? Answer: Knight Rider Result: While listings are always designated as either "Good" or "OK" matches, this eliminated all of the "Good" ones, suggesting NY isn't a good fit for people who demand a a KITT-chenette.
To better acquaint you with unfamiliar 'hoods, you can also connect via Facebook to see if friends've checked in near a listing, though they aren't really your friend if the only reason they hang out is because they're, like, completely incapable of outrunning you.