Hacker News Reader

Comment by warp_factor | Hacker News Reader

To make it clear, I absolutely hate coding interviews that makes candidate lose so much time restudying. I think having some critical-thinking is absolutely needed and way too many engineers lack some (especially those swallowed into FANGs)

What I described on my previous post is a credible explanation that my group of engineer friends came up with on why all the FANG companies pursue those heavy memorization algorithmic interviews.

View on HN

To my mind, the more likely explanation is that they would simply get too many false positives if they didn't use the algorithm stuff to filter potential hires. You lose a lot of potentially good hires that way, but the pool you're left with are all of a certain intelligence level. Whereas, if you don't use the algorithm stuff to filter, it's really hard to figure out who is even intelligent enough to do the job.

The article says that making the problems too hard results in a _worse_ selection, and not just more false negatives.

I'm not following, how does memorizing algorithms correlate with intelligence and skill in the engineering discipline of the job?

Not so much intelligence but applied diligence on top of intelligence is my sense. If you want what they’re offering and don’t want to jump through their hoops for it then they don’t want you. For those that ball at the hoop jumping as being beneath them, there’s probably additional characteristics that comes with that that they’re trying to filter out.

Given that many job applicants apparently can't write a Fizz-Buzz implementation, I'd say that being able to implement an algorithm is probably a reasonable way to cull the herd by a hefty margin.

https://blog.codinghorror.com/why-cant-programmers-program/

It's made me angry how many people I've interviewed who can't manage Fizz Buzz with only 2 conditions or at all.

I'm kind of angry that I've never been asked to do it.

I'm curious to find out what people define as fizzbuzz, do you have a good example of something you've asked in the past?

Print numbers from 1 to 100, except print Fizz for numbers evenly divisible by 3, print Buzz for numbers evenly divisible by 5, and print FizzBuzz for numbers divisible by both 3 and 5.

There's a correlation between the belief that memorizing algorithms correlates with intelligence and the application of algorithm questions in interviews.

If you have two great engineers in a kitchen, discussing a relevant problem, you want there to be synergy. That synergy is broken when one of the engineers has to Google how to reverse a binary tree.

It is also a safe place to work for the really exceptional engineers. They can talk freely about complex computer science, without getting blank stares or having to dumb it down. Otherwise it gets frustrating fast.

CS is way too big for any individual to know it all in depth. If it really is a “safe place” because the employees can talk about complex topics without ever losing anyone, that would imply that they’re all specialized in roughly the same topics and severely lack organizational breadth. I kind of doubt that’s really true.

You can have organizational breath with smaller teams. You pair a great programmer with an exceptional programmer and it helps if they are specialized in roughly the same topics.

How does that give you breadth?

The teams specialize.

The other commenter was describing it as a “safe place” where you can talk about complex CS topics without confusing people. Unless your teams never talk to other teams, that doesn’t fit.

One of the most valuable skills for any engineer is to be able to frankly admit the need to look something up.

Admit it to yourself. Then come prepared (to the job interview or inception meeting).

Do you work for google? This seems naive and inaccurate to my work experience, both in Tech and in Medical devices

If they cared about intelligence they would just give IQ tests.

IQ tests are a form of convergent preferential bias and they don’t test for abstraction or creativity which is more than half the definition of intelligence.

They don't test for creativity (though seeing a correlation between IQ and creativity would be interesting) but they almost certainly test the ability to deal with abstractions. That's pretty much what an I test is! What reason do you have for thinking otherwise?

They test for puzzles that appear as visual shape complexities which require use of the visual cortex. That is something a person might refer to as a visual abstraction. That isn't a practical abstraction, such as requiring a person to form an answer from the absent of acceptable criteria (to abstract or form something new to solve a problem).

It's not readily apparent to my that those two abilities should be distinct. It seems plenty likely that they would be at least highly correlated. While you're point about the visual cortex may be relevant, I don't think that's the whole story.

My understanding is that different IQ tests will rely on different kinds of questions (i.e. not all involve visual shapes; some involve word/logic puzzles). The point the scores for all IQ tests correlate very highly (and thus suggest that there exists some common factor).

What is convergent preferential bias?

It means a bias of coming together toward a tester's preferences. Do people excel on the few narrow performance meausures you find most correlated with intelligence?

There are divergent aptitude tests measure quantity and diversity of answers to a given question opposed to the one desired answer. Divergent tests are rarely performed, but at a stronger measure of creativity and problem solving which is typically what people actually want when they say intelligence. The reason why convergent testing is preferred over divergent test is because it is easier to measure and those simplified measures are easier to compare.

Can you provide a link for divergent aptitude tests? I'm afraid my search results all have to do with some young-adult fiction novel (called Divergent).

How much do such tests correlate with things like (for example) creative achievement?

Which is illegal to do, hence proxy IQ tests.

It is not illegal. Here’s Griggs v. Duke Power Co., 401 U.S. 424 (1971) https://supreme.justia.com/cases/federal/us/401/424/#tab-opi.... Go read it and see for yourself what it says.

Right at the top of that link:

> Testing or measuring procedures cannot be determinative in employment decisions unless they have some connection to the job.

IQ tests are not directly related to the job, and so are illegal according to that ruling. Coding tests are directly related, which is why they get a pass.

You didn’t read the opinion and you added the word “directly” to the summary.

Do you scan source code and draw firm conclusions about what it does based on skim reading the first comment you see?

Perhaps my old contracts prof could have a second career as a google interviewer. (He was notorious for cold calling people that hadn’t briefed their cases and eating them alive.)

The purpose of the "Primary Holding" section is to summarize the result, so that you don't have to dig through the whole thing.

> my group of engineer friends came up with on why all the FANG companies pursue those heavy memorization algorithmic interviews

I think the most likely explanation is that "elite" of CS grads usually do competitions like ACM ICPC and olympiads, which are full of problems like those watered-down interview questions. As there is no real authority on what makes a developer good, but there was a measurable outcome of those programming competitions, leading to high status and pride displayed by winners/participants, it was simply taken from there and dumbed down to fit into interviews. Some top colleges even have special prep courses for those competitions and comparing to them FANG interviews are super trivial.

Reply on news.ycombinator.com