Hacker News Reader

Comment by neilv | Hacker News Reader

We in tech reinvented some of it:

1. Recite this problem from the Leetcode corpus as performance art (which FAANGs tell undergrads at top schools to spend months practicing, on top of their coursework, instead of exploring what interests them).

2. Demonstrate this particular method (which happened to be taught at our rich-kids alma mater, but no one else had heard of).

3. Now answer this culture question in STAR format (which we told you to take time ahead of time to practice, and to prepare correct answers, so you could speak them faux off-the-cuff, as a shibboleth).

It's upper classes creating the questionable gatekeeping rituals, and children of those classes who tend to have the spare time and coaching to best pass the filter.

View on HN

Test prep is somewhat helpful for some people but some rich kids can't test their way out of a paper bag.

Standardized tests are a route to social mobility for some. This guy

https://en.wikipedia.org/wiki/Jim_Webb

came from a hardscrabble background in Missouri, standardized test results helped him get into the Naval Academy, he was a marine lieutenant in Vietnam, got a law degree, spent a lot of time in Asia as a journalist, wrote some novels including one about a guy who becomes a US Senator, actually became a US Senator representing Virginia.

There is a lot of concern that attempts to get kids into gifted programs other than through testing will be influenced greatly by family SES

https://ny.chalkbeat.org/2021/2/17/22288448/nyc-gifted-admis...

There was that time I yelled at my son when he was a toddler and carried him under my arm out of a grocery store and somebody called child protective on me. They sent somebody to my house and fortunately my mother-in-law had just been cleaning and the house looked absolutely spotless, full of books, horses grazing in front of the house, at least middle if not upper class. They had no idea that my house doesn't even have central heat!

Had they come to a hovel I could have had a very different outcome. They interviewed people and had a short investigation but based on appearances they saw what looked like "a good home" and might have gathered we'd have resources to fight them in court.

There will always be people impacted positively and negatively by these decisions, so the presence of some people who benefitted from standardized tests doesn't say much about their rightful place in the world. In any case, Jim Webb is an exceptionally weird example to make that case, it's not like he grew up the son of a sharecropper -- his dad commanded a missile squadron for the Air Force and was a full Colonel when Webb joined the Naval Academy.

I think people really overestimate the likelihood of CPS taking their kid after a visit.

There are lots of kids in absolutely horrid conditions that CPS doesn't take, they're not going to take them from your middle class home just because your mother in law didn't happen to clean.

Given that CPS in Texas will attempt to take gender-non-conforming kids who have the support of their parents in being GNC, and that it's in Texas, I am pleasantly surprised that there have been no deaths with that policy.

> surprised that there have been no deaths with that policy

The lunacy of today's synthetic sex identity movement aside, what did you imagine, that the Texas state government would take these kids out back and shoot them?

Pretty sure he was surprised no parents had blown one of the CPS people heads clean off in their front lawn when they came around to take their kids.

That children would commit suicide after being taken from their parents due to their gender politics. Seems pretty plausible to me, transgender folks are already at increased risk of suicide.

Trans people denied gender-affirming treatment have monumentally higher suicide rates than their peers. You know that.

No one knows that, it’s not falsifiable.

If "Demographic A has higher suicide rates than demographic B" is not falsifiable, then I'm not sure what is.

We can’t interview demographic A after they are gone.

Pretty surprising a research scientist doesn't know what falsifiability is.

Well, in general falsifiability and its application to psychology is up for debate because it is generally unethical/impossible to run single variable control experiments on humans :)

https://www.frontiersin.org/articles/10.3389/fpsyg.2015.0062...

You don't need to have single variable control to make inferences about differences between population subgroups, only to infer causality.

We certainly know that people who identify as transgender are at higher risk of suicide than the broader population.

Exactly, I was poking at the confidence expressed by the OP that inferred causality. Having more nuance and expressing less confidence would be closer to reality.

Irrelevant in this case. You don't need a double-blind experiment to determine suicide rates, as I already said.

I...what? If you find someone hanging from a noose with no sign of force or a struggle, you don't need an interview to know what happened.

What are you even trying to do, here?

Trans people given gender-affirming treatment also have a monumentally higher suicide rate than their peers: https://onlinelibrary.wiley.com/doi/pdf/10.1111/acps.13164

Yes, and?

Higher than cis people, perhaps, but much lower than if they had not received treatment.

Which is the whole point. They're better off with proper treatment than without. Anti-trans activists don't care about kids' welfare, they just want them to conform at any cost.

https://www.jahonline.org/article/S1054-139X(21)00568-1/full...

https://jamanetwork.com/journals/jamasurgery/fullarticle/277...

https://www.hcplive.com/view/suicide-risk-reduces-73-transge...

AFAICT this is not taken into account by Texas CPS in any way. Look for articles on it.

Yes, Texas legislators passed a virtue signaling law to do so, but it was blocked immediately. there is also a big gap between what laws are passed and the realistic chance of your child being taken away when the system actually confronts families.

Gender non-conforming children are being helped by this new Texas law, which prevents parents from transing them at the first hint of GNC behavior.

upper classes always win. if now system offer too many social mobility, they will change it.

The alternative is gating on soft skills and job experience. Do you really think it's easier for privileged kids to game our industry's admissions tests?

A poor kid would have a much harder time developing soft skills and getting job experience than a rich kid would have passing our admissions tests, which would be worse for social mobility than the status quo. A poor kid with a blue collar dad can't get an internship at his dad's friend's startup. A poor kid didn't go to a high school with a computer science club. A poor kid didn't grow up developing soft skills at a dozen different after school activities.

However, the poor kid can go to college on a Pell Grant (like me), study hard, and at least have a chance to meet the Leetcode bar and gain entry to the middle class.

truly privileged "kids" ain't going around looking for jobs thorough interviews, they show up as "investors" or are otherwise brought in by their 'connections'...

Granted, but we're talking about upper-middle-class versus poor. Upper-middle-class means you speak the prestige dialect, you know what "business casual" means and you already have the proper outfit, and you have access to the relevant technology at home, not to mention free time to use it.

A servicable laptop at the pawn shop can be had for $200.

but the internet no longer teaches you how to use it

that content has been paywalled.

Chiming in to agree here. Standardized tests + leetcode tests gave me the ability to overcome being a ward of the state and become a software engineer/ then founder of a YC company with zero connections.

Quick aside, in addition to the Pell Grant, there's also for former foster youth the CHAFEE grant (California specific).

> tell undergrads at top schools to spend months practicing

Come on. A friend of mine spent 2 weeks studying the leetcode corpus and then passed Google's coding screening. He complained about having to do the study, and I suggested that Google's salaries were such that it would be the most cost-effective 2 weeks he'd ever spend.

It all depends on how good you are at programming / math. If you have a natural talent for those then you don't need to spend much time preparring (hence why they're used a test). People try to "cheat" the test through preparing for months.

If a person has to study leetcode for months, then either the person:

1. doesn't know much of anything about programming, and should consider taking programming courses

2. is just not cut out to be a high level programmer, and should consider a career path more suited to their abilities

That's just the problem though, isn't it? Because the leetcode format is known, unless the person tells you, how would you know they studied for months to pass it?

Brilliant person A does leetcode for a week to get the hang of the format, then interviews and gets a job offer. Person B who just isn't brilliant but is highly motivated, studies leetcode for months, interviews, and also gets the job.

You might have fantasies of outlawing "studying for months for the test" but be real. Person whos "is just not cut out to be a high level programmer, and should consider a career path more suited to their abilities", it turns out, is doing just fine with their career path having studied for the test.

I also think calendar time is the wrong way to measure it, since there's a huge difference between studied leetcode for 40 hours a week like it's your job or being in school; and studied leetcode for 4 hours on Friday nights instead of going out, for months.

I don't see the problem. If one studies leetcode for months, and passes the test, that's a pretty good sign he learned a lot about programming from it. How could he not learn by doing all this studying?

It's like saying someone cheated on a Calculus exam by studying Calculus.

> not cut out to be a high level programmer

I think that was misunderstood. I'll try again. If one studies for months and yet fails to learn the material, perhaps one is not suited to that field of endeavor. There are certain things others excel at that I would just fruitlessly beat my head into a wall trying. For example, anything involving hand-eye coordination or a musical instrument.

And so what? Sometimes people ask me what kind of career they should pursue, and I don't say "STEM". I advise draw a circle around things you're good at, and another circle around things that pay, and your career is in the intersection.

False.

I've been operating as a high level programmer for over a decade and have a MsC in comp sci from a top school and a statistics/comp sci related Ph.D.

I was bored at my job and thought I'd apply at Meta. I've never really used python professionally (more of an R person, though I've solo dev'ed apps which have been in clinical usage for 3+ years with typescript front ends).

Did hours of leetcode over the course of 2 months to prep, because I was bored at my job and I thought it would be a fun way to improve my python skills.

Failed the coding section of the interview. Feedback was:

"The candidate clearly understands the problem and is an excellent communicator, but seems rusty with python."

No worries and no complaints, the story has a happy ending.

> Feedback was:

> "The candidate clearly understands the problem and is an excellent communicator, but seems rusty with python."

The feedback is always worthless. It may -- or may not[1] -- honestly reflect what the interviewer thought of you, but the interview process is not set up to produce reliable results, which means that what they thought of you isn't really related to your characteristics. There is only a very tenuous connection between their assessment of you and your performance.

[1] The last time I interviewed with Google, the recruiter congratulated me on doing well in the interviews and told me to expect a job offer. What followed was multiple weeks of silence (technically, not silence, they also asked me to make changes to my resume) and then the message that they were not interested in hiring me because my interview performance was poor. Whatever else we may say about this, we cannot avoid the conclusion that some of the feedback you get consists of intentional lies.

Multiple FAANG recruiters have told entire student bodies that.

Interview loops aren't deterministic, nor are they entirely objective. Congratulations to your friend for passing.

I've never understood why FAANGs are so fond of leetcode. Maybe they just get the headcounts and want a dumb and easy filter to hire some not terrible programmers. And that's part of the reason why the massive layoff is ongoing.

But STAR costs you little time to practice and communication skills will definitely pay off for a programmer.

STAR is what you do when you want to hire professional talkers.

Leetcode is not the only way to hire professional builders, but is much better than STAR.

I was surprised by the kind of things people openly admit to during STAR interviews. I don't enjoy doing them but they've surfaced enough red flags from candidates to save a previous employer a few bad hires.

I'd question whether it's really that effective. A dishonest candidate could craft a boring, safe scenario for a STAR interview question that paints them as an empathetic team-player, even if that's not how they've really acted in the past.

Meanwhile, a forthright, but flawed candidate who supplies real examples from their past they regret, but would like to improve on, would look like an iffy hire.

Leetcode is the only thing that scales consistently.

At small companies, a senior engineer can ask about past projects and get a sense of whether the candidate was a meaningful contributor or just along for the ride, but that's subjective and inconsistent when you have thousands of seniors.

Like the SAT, it probably does a reasonable job of providing a repeatable/scalable floor for skills that are a proxy for, if generally not the same as, what's needed for the job. And, like the SAT, the effort to prepare for the test is externalized.

Leetcode is dead, and it just doesn't know it yet. High quality LLMs killed it. Either proctor, in person onsite, or your hire likely cheated.

Leetcode can't die fast enough.

? Leetcode interviews are "proctored"

Obviously not well enough to stop cheaters.

I unironically refuse to go on video for my interviews because of my "lack of webcam" (which is true for my desktop). Literally never had a single interviewer stop due to it, and I've virtual onsite interviewed at most of the FAANGs.

What I mean by proctoring is really intense, invasive shit.

OpenAI chat could probably pass leetcode.

It can pass easy leetcodes. Here's a transcript of it solving an easy leetcode problem, and I also asked it for additional test cases.

https://pastebin.com/BpdcuWEM

There was another problem I tried where it gave an almost correct answer. Fixing the answer from then was probably faster than if I had to do from scratch.

That is O(n^2), when I believe it could have been O(n.log(n)).

Past the edit window; https://leetcode.com/problems/two-sum/ is the problem I fed it.

Of course, the other thing is that there's a ton of effort that goes into writing that question so well. I think if jira tickets were that well written, I'd be able to go much faster too.

Leetcode exercises are pretty effective at weeding out incompetents. If they have a false positive rate, that's fine, you've got plenty of applicants. If they weed out some of the best that's also fine, because the number one concern when hiring is to find someone competent enough that they don't become an embarrassment to anyone involved.

Real world business decisions are dramatically less optimized than people imagine they are.

> I've never understood why FAANGs are so fond of leetcode

Is part of it the recent frenzy to avoid bias?

If you can't take into account how smart or personable the candidate "felt", maybe leetcode is what you're left with?

It’s a poorly implemented method to attempt to introduce some equity or actual measurable outcome.

It has to be poorly implemented because people have successfully asserted in court that assessments that correlate with IQ are illegal. So making an assessment a grind is key to weeding out the lazy or people not crafty enough to figure out what’s up. They handle the intelligence part by targeting select schools.

In government, this has been incredibly impactful. Professional civil service jobs in the US had a mixture of professional expertise and general intelligence component weighted differently depending on the job. Now, if they still exist, they consist of trivia or grammar evaluations. The result is, in professional civil service, professions without gating credentials are hiring lower IQ individuals.

> Demonstrate this particular method (which happened to be taught at our rich-kids alma mater, but no one else had heard of).

This is one of those things that often gets repeated online, but I've never seen in practice. I've never seen a Leetcode interview question that can't be solved with knowledge outside of what you'd learn after the first semester of sophomore year. Helping my wife with Leetcode, I can see why some people would think this way though. Some people simply have better algorithmic intuition than others, so that may come off as memorization.

The only somewhat esoteric questions I've been asked is Indian immigrants asking about the minutiae about the JVM. Even then, I don't think these would be unreasonable if you were explicitly screening for Java developers.

Upper class folks send their kids into law & medicine. Technology is more of a proletariat aspiration.

Reply on news.ycombinator.com