On YC Interviews
There is already a lot of good advice on what to expect from a YC interview, including recent additions from S'09 batch mates Michael, Scott, Luke, James, and Slava. Do I really have anything to add? How much more can the collective of YC-alums write about 10 minutes?
Those were my thoughts last week. Since then I've spoken to several founders and it's clear folks are still hungry for more advice (and chances to practice pitching). Because I consider myself both a poor cook and poor giver of advice, I'll stick to the hamburger of formats. Below are my ramblings on the YC interview applied to a List of N Things template. You can increase N without bound by writing me directly.
Drop the conspiracy theories
You are convinced the YC partners have carefully analyzed your application, inner joined it with their deep mental database of startup knowledge, and crafted an intricate line of questioning that will trap and filter out founders who lack the Y Combinator gene. In fact, they are so ruthless they will purposefully scan your team for the weakest link. Prepare to degenerate into 2-3 mini-interviews, as one partner distracts the pitch master founder so the others can grill the shy founders. As fun as that sounds, it's just not true. If they really prepped so much for each interview, they could simply drop the prep and invite more teams to pitch!
I would be less convincing back in 2006, but fortunately there are now plenty of conflicting first-hand accounts of the process. This isn't one of those bad college campus job interviews where the company rep asks the same two brainteasers all day long to every candidate. About the only thing we know for sure is the YC interview will last 10 minutes. Or is it 15 minutes? Or 17 minutes? Is the interview an interrogation, or a group brainstorming session?
YC often feels like a collection of exceptions and outliers, and my own experience seems maximally distant from the supposed norm. I didn't have to talk about my market size (but make sure you know yours). My co-founder is just as shy as I am, but they seemed perfectly content directing all the questions to one of us (sadly me, and solo pitching isn't unique). I don't remember anyone cutting off my sentences. By the time Jessica said "12 minutes are up", we had only shown one frame of our demo (the home page).
So either the YC partners carefully tailor each interview into more categories than there are Myers-Briggs personalities, or there is a better explanation: The YC partners are not plotting all week about how they can stump you. All the questions are just-in-time generated, but may come across as pointed or blunt because they only have 10 minutes to understand your vision. Left unchecked, they could talk for hours, but they will cut you off to maintain the schedule. Be concise and avoid wasting your precious time to shine on long-winded responses. Their gut reactions have more depth and perspective than the average hacker, but in the end, they are asking questions because they are excited. They want to picture your company succeeding, so help them out!
Thankfully, they won't ask you why manhole covers are round, to choose three adjectives that best describe you, to tell them about a time in your past where you acted professionally around a difficult co-worker, or the many other classes of annoying interview questions from other industries. For obvious reasons you might get a question that resembles those awful estimation questions ("How many rental units are in the US?", "How many impressions a day does Facebook receive?"), except the correct answer in a consulting/programming interview is very different than the one you should give as a founder and domain expert! 
It isn't a tournament
Again, they already like your application. Your odds are pretty good, better than completing a four-flush postflop. But the poker tournament analogy breaks down very quickly because acceptance to YC is not a competition over a fixed number of winning seats. In my batch, 27 teams came to the first dinner. The class before us apparently had 16, and the class before that 21. I'm glad these are not even numbers . The seemingly arbitrary class sizes are evidence YC funds companies that pass their bar, not some artificial top N companies from each day. You aren't competing with the other founders in the waiting area. In fact, you are best off being social and befriending the other contestants, if only to get your voice in high gear prior to walking into the interview. It doesn't matter if you are normally shy. YC applicants tend to have a lot in common.
Don't waste time scheming
Founders ask me what really made us stand out at the interview. I have no idea. However, please don't spend all week scheming up an outrageous strategy for how you will be the one standout from among 75 or so companies. If you are doing a fashion startup, don't parade into the room with 20 teenage fashion models to prove you are in touch with the market. If you are doing a music startup, there is no need to convert the room into a rave and deliver the pitch as a rap song. There are times in startup life where going over the top is necessary. If this was American Idol or a winner-takes-all poker tournament, the best strategy might be to do something wild and crazy. This is different. You are already past the most harsh and most random part of the funnel. There is no need for a high volatility strategy . Give a good, solid, well-prepared pitch and answer their questions along the way. Help them help you. Confirm their suspicions that you are good.
Instead of plotting a crazy skit, a better use of your time is to practice pitching friends.
Our personal experience
In our case, we arrived an hour early because of the weekend Caltrain schedule. That was way too early, but sitting in the big orange room is far more enjoyable than waiting for a dentist appointment or an orchestra audition. Unlike the dentist, people are happy both entering and leaving. More importantly, no one can hear the audition room. The waiting area doesn't go completely silent as everyone listens for the audtioner in the hotseat to miserably choke.
I chatted briefly with Kate and inquired about directions to the nearest In and Out Burger (forgive me, I never lived in Cali). Sadly, there was nothing walkable nearby, but YC provided plenty of food/snacks (to entice alums to show up, of course). Alum Dan Gackle was breaking the ice, telling tales of his infamous Demo Day chopstick. Two other teams joined the conversation, and lo and behold all three of us made it! Two other teams nervously sat in the corner doing who-knows-what on their laptops. I never saw them again (highly overfit model, I know). Before you ask, I've confirmed the nervous laptop guys did not include James from Listia trying to reinstall Skype.
I realize that wasn't a strict following of the List of N Things formula, which goes to show even the simplest of plans can get derailed. Hopefully you enjoyed this overcooked burger.
Best of luck,
 In one case, the answer is "40 million according to the US Census." In the other case, the answer is "Gee, there are 300 million in the US, and out of my old college hallway, 1 owns a house, another moved abroad, and the remaining 3 of us rent. The typical household is 4 people, and the top 2% of wealthiest households probably own a vacation home, so a first-order approximation is 40 million."
 This isn't like college admissions where they predict a yield rate, set a target number of slots, and then send all the acceptances at once hoping their conversion numbers work out. They could easily set nice even numbers for the number of companies participating because of the rolling nature of the acceptance phone calls. The trend is also not monotonically increasing, so you can't point simply to YC's growth.
 More formally, imagine a conservative strategy that gives you an outcome between 65 and 85, uniformly distributed. Then consider a risky and outrageous strategy that has a 90% chance of giving you a 0 outcome, but a 10% that you strike the right chord and hit a home run for 100. If YC only accepted the top scoring team, you'd have to swing for the fences and hope for the 100. Fortunately for you, a 70 should be just fine, so the solid strategy is best.
The author of this post is Lee Lin, co-founder of RentHop, a syndicator and marketplace for Apartments for Rent in New York City.