Hacker News Reader

> The added bonus is all new hires at least know the effects of space/time complexity and how to make space/time trade-offs deeply well.

Do they?

In my experience this was true in the early days of white board interviews where the only people that even thought about this stuff were people deeply interested in algorithms and computation.

But in the post-"Cracking the Coding interview" era (including the transition into the full leetcode era) people seem to be able to pass these interviews without really understanding anything in the same way college students can ace calculus while have literally zero intuition about derivatives.

I don't mind whiteboard interviews, they're a fun game you can train for and the practice is interesting. But the big tragedy now is that I feel almost nobody understands algorithms anymore since everything has been reduced white board coding drills and everyone believes they have an understanding of something they don't.

As a great example of this almost nobody today knows any heuristic optimization techniques (think the old ITA software puzzles), which used to be near and dear to anyone that cared about algorithms. But since you can't test for heuristics they same as you can established algorithmic solutions nobody studies them.

> they are optimizing for the people willing to push themselves that far to learn something

They are optimizing for people in one of two groups. One is people freshly out of college who still have that stuff at the top of their mind. The other is people who have the free time/energy to bring it back to the top of their mind instead of needing a few extra minutes to go back that far in their memories. Pick your privilege, in other words.

> all new hires at least know the effects of space/time complexity and how to make space/time trade-offs deeply well

In my experience, this advantage fades within their first evaluation cycle. And what they certainly don't seem to learn is when to do these kinds of evaluations. Oh, the number of times that I saw one of my coworkers hyper-optimize some part of a system where both N and K were too tiny to matter, then just cargo-cult a crappy higher-level algorithm where such analysis really would have moved the needle.

BTW, since I know someone will launch a "sour grapes" personal attack (and get a free pass for doing so because I'm on someone's target list) if I don't address it, I passed such interviews at 52 with zero preparation and zero trouble. So no, I don't take it personally. I just recognize this approach's flaws. It's why I refused to do interviews at that company, taking a hit for that every half until I left. Some of us stick to what's right instead of rationalizing what's wrong.

My favorite version of this is the driving test you take in California before you get your license.

One of the rules is that you automatically fail the test if you bump a curb with your wheels. That means, when pulling into a spot, you can't inch forward until your front wheels come to rest on the curb. Fail!

At first glance it seems like a trivial thing to fail someone for. I know in normal driving, I use curbs like this all the time. On purpose, and with no hazard to myself or others.

But the fact is, you're told this explicitly up front. "Don't do this, or you fail." Given that you're warned about it—and know this is the one thing you cannot do—if you still contact the curb, then you've demonstrated that you don't know where the boundaries of your vehicle are. (Or maybe that you're not so good at comprehending instructions?)

It works as a simple filter.

(EDIT: I checked with the DMV's test criteria[0], and I don't see this mentioned in there. There's an automatic fail for driving over a curb, but it's not clear if simply contacting it counts. Not sure if my examiner was a sliderule, or if the rules have changed since I first got my license.)

[0, p. 24] https://www.dmv.ca.gov/portal/uploads/2020/04/dl955.pdf

One of the tests is for backing the vehicle up. In that test, you may not touch the curb. It's not otherwise obvious from the document provided that any other contact with the curb is disallowed. So either it's untrue that it's disallowed when pulling forward, or it's untrue that you know about it upfront.

Regardless, not touching the curb, even if arbitrary, can be an indicator of competence. It demonstrates control and awareness of the vehicle.

Memorizing an algorithm is a potential indicator of many things, not all positive, but I don't think it says anything at all about competence.

> ...you can't inch forward until your front wheels come to rest on the curb. ...if you still contact the curb, then you've demonstrated that you don't know where the boundaries of your vehicle are.

This demonstrates one reason why whiteboard interviews aren't great. They don't mean what people think they mean, and people think of them in very black-and-white terms. X always implies Y, or at best they say, "Well, maybe not, but..." and then justify the exact same result with only superficially different reasoning.

In your example, the driver might have demonstrated he doesn't know where the front or rear of the wheel is exactly. But it could also be he was just nervous or driving a vehicle he isn't used to using. Maybe he had to borrow his older brother's car when mom's broke down at the last minute. Saying he doesn't know where the boundaries of his vehicle are goes against all the other evidence: he didn't strike anything with the boundaries of his car. The curb is already inside those boundaries, intentionally and by design of the test.

A lot of that stuff depends on the proctor. When I took the driving test the first time in my state (not CA), he almost failed me because I failed every single left turn. Thankfully at the end he asked a question, and realized that I had done every turn incorrectly because nobody ever bothered to tell me that left turns always should go to the LEFT lane, not right.

You and half the country. The rule is actually you match the lanes left to right. Aka if two lanes are turning left into a three lane road the second from the left goes into the middle.

I suspect that people happened to have just zoned out during this part of the class, although a friend of mine recently said he was never taught that one is suppose to stay as far right as needed to allow people to pass on the left either.

And its not just random drivers, the more esoteric parts of the law (driving on the shoulder in TX, free right turns, uturn laws, center turning lane, etc) aren't even understood by many traffic enforcement officers.

Yeah, these are the gotchas that vary by state.

In California, for a single left turn lane, you're allowed to complete the turn into any of the lanes you choose.

The driver handbook doesn't go into detail about double turn lanes that feed into a 3-lane cross street. My own rule is that only the inner lane (leftmost) must complete the turn in the matching innermost lane. The outer lane can end their turn in the #2 or #3 lane. I tend to choose the #3 lane just to put as much space between me and the car next to me, in case they incorrectly go for the #2 lane. Unless there's an opposing right-turning car that I trust less :)

I failed my driving test as a foreign exchange student in Indiana for this exact reason. I was not amused.

> I used to think this too, but as a filter I think it's having the desired effect.

Is it though? I'm a hiring manager and don't ask such questions and yet have had no problem hiring really talented engineers.

The reason I don't ask those questions is because I know that cranking up the interview pressure and asking arbitrary, revisable, common interview questions doesn't actually prove anything aside how good the candidate is at being interviewed.

I think they are saying not that it’s a filter for talent, but for people willing to “put in the work” to pass that type of interview, which sounds about right. I feel the same way about some colleges.

This is more what I meant.

Defining a filter in order to filter people out is fine, but we should be honest about the fact that it has very little to do with learning, and more to do with the willingness to deal with crap and jump through hoops. More amenable employees are easier to manage, nothing wrong with that.

All exams can be crammed, and having crammed into plenty of gated opportunities in life before, I believe you're mistaken if you believe that it produces more knowledgeable workers as a result.

At the end of the day, it's a market like any other. If there is someone who is more willing to do X for Y reward than you are, then they will get that something instead of you. And if you find that there is a shortage of people willing to play ball, you can always work towards bringing in more people who are willing to play ball.

Agree with your last sentence. If you want something, you have to go get it. If employees aren't motivated enough to organize and close the power gap, then employers will continue to be motivated enough to push for and work towards their own goals.

You have to be careful about your assumptions. Being willing to put in the work could be a bad sign. If you already have a bunch of job offers, you aren't going to spend weeks studying to get one more job offer, unless the place might offer dramatically better benefits.

And, it may be some people don't need to be any extra studying work. There are people who do programming competitions for fun, and there was prior discussion of Google saying those people did worse on the job: https://news.ycombinator.com/item?id=9324209

> If employees aren't motivated enough to organize and close the power gap

"If X aren't motivated enough..." usually ends up being a really horrible take for most values of X. I'm not so sure it's different for X=employees, specially when the type of interview we're talking about is more burdensome to older folks, folks with certain personality types or anxiety issues, arguably most women, etc. Why should the burden be theirs alone, while employers get a complete pass on setting up whatever arbitrary hurdles they want?

I think that is reasonable that a company do some tests on folks that don't have much baggage to show for, because in this case they are doing a wild guess and must count a little bit on luck.

But the parent was on his 30-something when he did the interview, and i bet he had a lot in his curriculum to show for, so in this case the company dont have to guess anymore.

I think it says a little bit that the HR is failing somehow. Failing to adapt to different kinds of human beings.

By the way, there are a lot of psychology into this. Sometimes the right kind of people to do some sort of task is the one that are avoiding the whiteboard tests.

In the end it will lead to an unbalanced company with people very similar and optimized for a couple of areas, but lacking in others.

This is something that has always perplexed me. These interviews may make sense for an entry level candidate with no ink on their resume beyond a GPA, but then any amount of actual work history just gets completely discarded. Experience needs to count for something.

I would think these companies that are built on measuring data would have some idea if the whiteboard interview works. Otherwise, why would they do it?

Q. Why does 90% of human society work the way it does?

A. Because that's the way we've always done it.

I can't speak for SV companies, but my experience of data is that it can just as easily rationalize preconceptions as be used to inform decision-making. Somebody has to interpret that data, after all.

If our way of hiring doesn't work then how come we're so rich!?