We want random access to your thoughts, not to listen to a single path through them, prepared in advance.
OK, obviously this is a continuum, and not a binary thing. But why "random?" Wouldn't you think there is some value in giving the startup founders the chance to present things in the way that makes sense to them, before starting with the grilling? Or is the case that you actually do that ( that point wasn't entirely clear from TFA)?
I believe Paul is using "random" (in particular, "random access") in the programming sense of "arbitrary" – the interview is there to dig deeper into the bits they are interested in/concerned about.
Suppose you are using very slow media and it takes a tenth of a second to retrieve the information. That's an acceptable performance penalty ;-)
On the other hand if you are using external storage, with indirect lookups in order to find the information, that might not be so great.
At the same time, networked responses are great. There isn't a problem with having a different node respond to the request, since this approach uses a sort of strange architecture and session state is shared between all nodes.
I'm trying to picture the RAM chip designed by a literal-minded electrical engineer (even more so than most engineers!). It stops responding if it detects any predictable pattern in the memory addresses you're accessing.
Sure, I get that. What I was saying is, would it not potentially be better go give the founders some time to give their thoughts in the order they have them in... before starting to go off on tangents and probe specific areas? Especially considering that the question you're about to ask, may be something they're going to be answering anyway.
I guess that if pg and crew believe that they learn something by intentionally knocking people out of their planned routine, then that's one argument for it. That is, if the idea is "we want to evaluate your ability to think on our feet, react to unplanned questions, etc." then OK.
They've already explained their ideas in the form they chose in the application. So the interview starts with questions we had after reading it.
We don't try intentionally to knock people about to see how they respond. We don't have the time or the energy for such games. We just have a lot of questions and not much time to get them answered, and empirically the questions we have almost never turn out to be the ones answered in prefabricated pitches.
Do I understand it right then that what one might call "idea pitch" has already been passed once you are being interviewed? Because in this case it makes sense to ask "random" questions.
Before I read the discussion I kind of shared the op's point. But seeing it from the other perspective, it almost seems like the basic "fault" (if you might call it that) the op made was preparing for a test he already passed.
We don't try intentionally to knock people about to see how they respond. We don't have the time or the energy for such games.
That's good to know. I think some people have somehow gotten the perception that you guys do just that.
We just have a lot of questions and not much time to get them answered, and empirically the questions we have almost never turn out to be the ones answered in prefabricated pitches.
In my preparation between the accepted application and the interview, I was cautioned that the interviewers will not remember your application when you walk in the room.
As I understand it, random access just means asking you about different facets of your products, probably in the same way you and your teams have been thinking about and solving problems related to your product. You did not sit during Day 1 and outlined every single problem and feature you will support did you? There is randomness involved in your journey towards evolving a products, and that randomness already answered most of the questions PG might have asked.
Remember that YC backs founders much more than ideas. Also remember that they've already vetted both the founders and the idea, with the help of YC alumni, prior to the meeting. I would also imagine they have a strong aversion to BS and a well-honed BS detector.
Ignoring the pre-packaged pitch you have spent months preparing and asking you some unexpected questions would seem to be a great strategy to quickly figure out what they really want to know - which may or may not have any relation to what you want to tell them.
I think that the biggest thing that people on this thread are missing is that the interviews' main point is not to evaluate the idea, but rather to evaluate the team; the value of the idea itself is of secondary importance in that it reflects the ideas and thoughts of the team itself. As pg has said before, one of the things that they look for is confidence. The problem, according to him, is that some people are very good at faking it. Therefore, an objective of pg is to get people away from canned routine and to be original. Therefore, if the people are smart and capable of expressing their ideas, they should do well.
"the startup founders the chance to present things in the way that makes sense to them, before starting with the grilling?"
If you were sitting down in the coffee shop and talking to a friend or someone you met who had an idea you wouldn't allow them to make a presentation. They would give you a sentence or two about the idea and then you might ask a question. They would answer the question and if they knew of something that was related to what you were asking they would offer it. Every now and then they might blurt out a point which hadn't been covered.
(Edit - corrected "you" and "they" which I mixed up..)
Speaking strictly for myself I don't have any patience to hear presentations. I like to poke around and ask questions and engage in conversations. Conversations aren't presentations and aren't one sided. They are back and forth.
Another thing to keep in mind is this. If you are making a presentation it's possible that the person who is listening will be mind stuck at a point you made in the third sentence and tune out something important you said that comes later. So you can be at a disadvantage depending on who you are presenting to and what they know.
Let's say you say "we are using node.js" and the gatekeeper doesn't process that as a chunk of understandable info. He will stop at that point and miss what you say after that until he recovers.
If you were sitting down in the coffee shop and talking to a friend or someone you met who had an idea you wouldn't allow them to make a presentation. They would give you a sentence or two about the idea and then you might ask a question. They would answer the question and if they knew of something that was related to what you were asking they would offer it. Every now and then they might blurt out a point which hadn't been covered.
Maybe I'm just more patient than most people, but I feel like I'd give them more than a sentence or two, if it was clear that they had more than a sentence or two of things to say... I generally try to avoid interrupting other people when they're talking because A. it's rude, B. it's presumptuous, and C. I don't believe it's an effective way to communicate most of the time. Now if they use a term I'm not familiar with, or whatever, and there's no choice but to ask a question or get lost, then sure, asking for clarification makes total sense.
Now if they were rambling without pause for 10 straight minutes, then yeah, sure, at some point you have to interrupt, especially if there are time constraints (which, to be fair, there are in a YC interview). But if I interrupted with a question, I'd expect to get an answer to that question, and then possibly have them pick their initial narrative back up where they left off.
Anyway, it's all very subjective, and I'm sure Paul has figured out a lot about how to do this over the years. I just find myself wondering to what extent the YC interview gives the startup team a chance to get their initial burst of "stuff" out before starting to take them off on tangents.
"I just find myself wondering to what extent the YC interview gives the startup team a chance to get their initial burst of "stuff" out before starting to take them off on tangents."
I believe Paul answered this in a comment on the thread.
I don't see the need for the start up to present anything in what is an interview, not a presentation. As Paul is conducting a ten minute interview based on the application, why would he have to give any time for the start-up team's prepared presentation? Its only ten minutes, and its him asking questions of the start-up to see how well they can answer the questions that arose from their application.
OK, obviously this is a continuum, and not a binary thing. But why "random?" Wouldn't you think there is some value in giving the startup founders the chance to present things in the way that makes sense to them, before starting with the grilling? Or is the case that you actually do that ( that point wasn't entirely clear from TFA)?