Hacker News Reader

Comment by seanhunter | Hacker News Reader

The premise of this article that you assess people's honesty and reliability by ambushing them with a trick (heuristic which is opaque to them) expecting only certain responses to be "correct", is intrinsically dishonest and low integrity. Although I would like to think that I would often "pass the test" here, I would be extremely reluctant to work with anyone who had pulled any bush league BS like this.

In my mind this is very similar to interviewing someone with a trick question/riddle/conundrum and failing them if they don't answer one of a specific set of "right answers" that only you know. It's a great way to feel smart by rejecting people but not a great way to discriminate between good and poor candidates.

View on HN

Totally agree; starting any relationship with a lie is a terrible idea. I think the "Reverse Orange Juice Test" is that anyone who uses the orange juice test on you is being dishonest, and therefore disqualifies themselves from being a client.

Same with the "difficult but do-able" side of things. It might be difficult but do-able for most people, but the person you are talking to has already done it and is going to package something up for you based on their expertise. Like every situation, honest, realistic discussion of the process is the real winner. Tricks don't work; communication does.

What lie?

Stating that you need all that orange juice (which you don't actually need).

I'd focus less on the name of the test. It's a story that's being passed down from someone else to someone else.

None of these people are booking banquet halls and asking about orange juice.

The setup is kind of ridiculous and I suspect just a little bit made up just to give it a pithy name that makes people say "What's the orange juice test?" so you can go into this whole spiel about banquet halls.

The core idea is sound: Don't work with people who either don't know what they're doing or try to make your decisions for you. But I wouldn't work with anybody calling it the orange juice test either, because it seems like something someone made up to feel smart by knowing arbitrary jargon.

Yes, the core idea of looking for someone who neither says they can do something difficult really simply nor says they can't do it is sound.

It's the idea of requesting the contractor to give feedback on a task that you have no intention of having them complete that people (including me) are taking issue with.

The title of the test is irrelevant - the test could easily be asking a contractor if they could develop an application that does abc in such-and-such a time-frame then turning round and telling them "I don't actually want an application that does that, I just wanted to see what you'd say".

Here's where I think you and others are missing something. Orange juice is in the hypothetical that they are discussing. There's nothing that says you have to make up a request. It can be a legitimate request, it just wasn't in the hypothetical.

The key thing is whether the response is "that's impossible" or "you can have that for this price." In the anecdote, the businessman told Gerald that he'd already passed the test.

I don't think so.. it's definitely implied that you ask for something more than you need and then walk back from it. Like he says "“Oh you don’t really intend to have the breakfast. It’s just part of the test."

> failing them if they don't answer one of a specific set of "right answers"

A few years ago I did a round of interviews at Google. Despite what they say about not doing those kinds of questions anymore, they gave me a whole slew of them, in programming form. Not "how much would it cost to wash all the windows in NYC", but questions like "given a grid of land, each block having a price, pick the largest rectangle under $X".

If I'd heard of it before and knew the solution already, I did great. If I hadn't, I did poorly- but generally figured out the algorithm they wanted within a day- just not in the 45 minutes given. Alas, I guess I'm not "Googly" enough.

I don't know, the two questions are nothing alike:

One is asking for a guesstimate, while the other is a proper optimisation problem. To me it's a valid interview question, albeit a bit difficult for a short session.

And neither are really great predictors of coding ability.

But when you have the #1 name in the world for a brand, it's OK to throw out 99 good people for every 100 that apply.

There's a specific best solution involving dynamic programming. If you've heard the answer, like say if you spent a month reading a bunch of common programming interview questions, then you can 'solve' it in 15 minutes. If you haven't, then you need to be ready to invent the same solution in 45 minutes (less the 5 minutes spent explaining the problem).

Maybe Google (and the other big tech companies using the same style of interview) is looking for the types who can brilliantly solve problems like that in 40 minutes, but they're instead going to hire a lot of guys who read the book of problems ahead of time.

I've solved this very problem without reading the answer. If you understand dynamic programming, this is simply testing your skills in it (although I'd agree it may be too much for a short session).

Sure, but usually the problem is decorated in a few layers of bullshit to hide what the true problem is. I remember the core, actual problem, but that wasn't what got presented. So take off another 5 to 10 minutes cutting through that.

They don’t read books, they solve leetcode problems.

A friend was obsessed with getting into FAANG and relentlessly drilled leetcode problems. There’s a whole community around it, basically memorizing enough toy problems that they can pass these interviews.

The companies should know by now that these clever problems aren’t really that clever. I have not interviewed in a long time but I think I’d get kind of depressed if an interviewer asked me 2sum. I’d just lose respect for the company.

Are you allowed to Google in these interview questions?

But this situation will come up organically when you are getting started with a client. Everything you do costs money and it is ethical to be transparent about that. Because the client may not want you to spend thousands on a particular part of the project that they don't actually care about. And there could be an oppurtunity cost in terms of other objectives.

Of course there are exceptions to this. Sometimes it makes sense to overprice work because you know the client will uneccessarily nitpick minor details. It makes sense to accomodate this up to a point, but you can't be completely transparent about what it costs. You don't always know how fussy the client will be ahead of time.

Responding to client requests is a core competency. It's absolutely something you want to test in a potential supplier as part of your due diligence.

interpreted charitably, the test is not to make up a false complex problem (if you're looking for a contractor for a trivial task you wouldn't even need the orange juice test). rather its just to be honest about the problem/challenge and listen carefully to the response?

Reply on news.ycombinator.com