Hacker News Reader

Comment by Salgat | Hacker News Reader

The leetcode grind is my biggest dread. I can do it and I do well, but I'll be damned if it isn't a lot of mindless rote memorization that immediately leaves my brain the second I land the job, never to be used again (even in my job) until the next time I apply around.

View on HN

Honestly I just stopped interviewing at leetcode places and those that expect you to "prepare" for their interview, especially those that are dumb no-name SaaS companies.

It's less about me being stubborn as it is those places are hiring based on the wrong skills, and those end up not being good places to work.

At least you can do it well, imagine the grind when you can't.

Remember when fizzbuzz was the gold standard in trick interview questions? And Joel was a revolutionary for advocating it? Things have certainly changed.

In some sense, leetcode is fizzbuzz on steriods. As I recall, the "trick" to fizzbuzz is knowing about the modulo operator. Most leetcode problems have a similar trick that you need to know/learn to complete them correctly.

Fizzbuzz wasn't about seeing if a candidate knew the 'tricks', it was to test if you knew how to code at all. If you know about loops and if statements you can do fizzbuzz even you've never heard about modulo operators. Fizzbuzz was designed to catch the complete bullshitters who literally could not code. Apparently that was a real problem people were having at the time.

At the time? It's still happening today!

The key is to jump jobs before they catch on

    > Fizzbuzz was designed to catch the complete bullshitters who literally could not code.

I'm unsure if we are in agreement or not. To me, the expansion of leetcode and friends is driven by the desire to filter for higher quality candidates. For example, in year 2000, fizzbuzz was good to filter out people who could barely write a for loop. Then, big tech needed higher and higher gates to keep out people who did not meet their standards. So fizzbuzz was given a healthy dose of anabolic steriods and told to "hit the gym". In a sense, fizzbuzz was transformed into leetcode, but the spirit was unchanged.

I wonder what’s the value of even doing leetcode when AI can solve it in 10 seconds…

It’s not about testing ability to solve those problems, it’s about testing conformity, determination, and IQ (while being job-specific and thus legal).

It's about conformity.

Every place that tested leetcode in my interview (easy/mediums) never impressed me with their work culture. Usually it's a proxy test for obedience and/or accepting bureaucracy.

Also forget about career advancement. (Why would we promote you when we can replace you with another code puzzle solver?)

IQ tests aren't illegal. Several big companies overtly use them.

AFAIK Hasn’t been legal unless the test is closely related to the job since Griggs v. Duke Power Company, a 1971 SCOTUS case.

You can just go look this up with IQ test vendors; you don't have to derive it axiomatically. They proudly list their (large) clients.

You need to ensure that the test doesn't discriminate on the basis of a protected characteristic though.

These are just completely bog-standard cognitive aptitude tests; in fact, they are if anything less rigorous than a modern IQ test. They're generally not specialized to specific jobs. You can read Reddit threads about people taking (and studying for) them.

Don't get me wrong: I think this is an incredibly dumb practice. But there's a mythology that IQ testing is a super-effective tool for recruiting that has been suppressed by anti-discrimination law. That is not the case. Most companies don't use IQ tests, because they're not fit for purpose.

It really really depends. If you're gonna train junior people then IQ tests are great, otherwise work sample tests are better.

It definitely is true that IQ tests are used less in the US because of prior case law though.

You mean they're selecting for lower IQ? (higher candidates will avoid the leetcode jobs)

The highly specific/specialized problems leetcode has presented have never held value, outside of a very small set of companies (i.e. MOFAANG).

What's the O in the acronym?

Oracle

First time hearing that one, no doubt coined and promoted by Oracle themselves

While the actual implementation isn't particularly good having a decent understanding of algorithms is useful in a lot of contexts.

Lots of jobs don't require Leetcode. I haven't done it in years yet I still get jobs consistently.

Too many folks in this industry think any job other than FAANG/startups is not serious.

Look at this way: It's just something you do for money.

Simply don't, lean on your network if you can. It ends up bottlenecking your search anyway.

Except at smaller companies, your network just gets you a referral and interview. You still have to go through the same interview process.

The two exceptions I have had were when a former coworker who is now the director of an f500 non tech company was going to create a strategic position for me. He needed someone he could trust.

The other time was when a CTO and the director of application development were both former coworkers from another company and it was just a matter of me saying yes.

The first job would have been more stress than I was willing to deal with and I don’t do large companies (I was suffering from PTSD from my time at AWS). The second didn’t have the budget to meet my compensation target.

Oh totally! My aim is to avoid leetcode as a 10y+ developer with real track record delivering $XM projects. To me it is a game breaker to have to study up on leetcode problems, as it keeps me from juggling multiple interview in the pipeline.

Every network job I've gotten (two) has been a walk on interview, I'm fine with this as long as I don't have to do circus problems.

Sure you can avoid leetCode if you’re okay with not maximizing your compensation by working at one of the FAANG or adjacent companies that pay at top of market.

But you aren’t going to have a walk on non leetcode interview at one of the companies that pay a quarter million+ for a mid level developer based on a referral. I’m not saying you are a mid level developer just giving an example of comp.

The amount a returning intern I mentored when I was at BigTech had a return offer that was the same as I made two years prior at 60 person startup.

Now that I’m out of BigTech, I had to get a job as a “staff software architect” at a 3rd party consulting company to come close to what I was making in 2020 as mid level consultant at AWS and I’m making over $80K less than I would make as a senior doing the same thing at AWS or GCP.

I’m 50, an empty nester and I’m good with making that trade off. I would rather get an anal probe with a cactus than ever work at any large company again.

But if I were 30 in 2024 instead of in 2004, yeah I would grind leetCode to make a 300K+ a year.

And yeah I got into BigTech without a coding interview and could probably weasel my way into Google/GCP in the consulting department without one. But that needle would be hard to thread for the vast majority of prople

That’s why I always advise anyone in CS to practice for coding interviews

good advice - I do hard disagree with Sure you can avoid leetCode if you’re okay with not maximizing your compensation by working at one of the FAANG or adjacent companies that pay at top of market.

I think the BIGGEST misconception in our industry is that it is only FAANG that pays TOP dollar. No one teaches CS grads coming out of college the hardest truth of it all - you will get paid what you are worth to the company and your career should be geared towards figuring out how to make yourself more valuable to a company that company is to you - this is where REAL money is and this can be had in A LOT of places. there are maybe 10 people at FAANG that have this - 99.76% of people at FAANG are expendable. too many bodies - most of the people are no ones. on the flip side there are 1,000’s of companies that have been in business for decades - there, with proper planning (much better time spent than fucking leetcode), you can become more valuable to the company than company is to you and your salary will reflect that - even higher than highest of salaries at FAANG…

I don’t care how much I’m “worth” that 60 person startup I worked for pre-AWS wasn’t going to pay me $225K and I went in as a mid level consultant. That was fair, I only had 2 years of AWS experience at the time.

The 600 person company I work for now isn’t going to match what I could make at Google. On the other hand, Google would require me to be in an office and I would not have unlimited PTO.

Everyone is expendable. No matter where you work.

I don’t think you know the salary ranges at BigTech and how they compare to the rest of the market. Those thousands of companies aren’t going to pay FAANG salaries.

That grad I mentioned coming out of college is making $160K their first year and that was working in Professional Services not development. An SDE was starting out at $175K - $190K

Everyone is expendable. No matter where you work.

I can tell you from personal and numerous other examples that this is false

I don’t think you know salary ranges at BigTech…

I may not (even though of course I do as this is public information at this point) but I was getting paid $900k at 15th year. And know people in similar-ish range - just coding…

true story, I was on a honeymoon when I got an emergency call at work - no one knew what was going on, jumped on and fixed it in 10 minutes - got a bonus enough to buy a new car…

every junior I encounter in my career nowadays I teach that they should NEVER consider themselves as an “employee” but as a “corporation” and working somewhere not as “employment” but as “partnership” - makes a huge difference through and through when you cement this in your brain

> I can tell you from personal and numerous other examples that this is false

You think if you got hit by a bus tomorrow your employer would go out of business?

You think if you got hit by a bus tomorrow your employer would go out of business?

my current employer - no, I am winding down my career. my former employer obviously not since I am no longer there but they paid me 6 months extra to do nothing but train people when I was leaving and have also done “emergency” work for several years after.

companies should strive of course to not have 10(000)x irreplacable employees but in 1,000’s of places they are always there. and path to become one of such people is easier than slumming it at FAANG I guarantee you

Again, you think 1000s of companies would go out of business if one key person quit? Unless you are talking about something like a private practice one or two man shops

I am giving actionable, repeatable advice, not once in a lifetime lottery tickets.

Do you have actionable repeatable advice that would allow that junior developer to make $175K straight out of school and over $300K 3-5 years in the workforce?

Those FAANG and adjacent companies are paying collectively 10s of thousands of developers that every year.

Again, you think 1000s of companies would go out of business if one key person quit? Unless you are talking about something like a private practice one or two man shops

NOT at all what I am saying… what I am saying is that there is a path in this industry where you are not a slave to “FAANG” where no one knows your name. there are (tens of) thousands of companies where you can set your career such that you are more valuable to them than they are to you. and once you get that money is no object no more… and you get to stay decade (or two) in the same place vs doing the leetcode-change-jobs-like-socks dance.

actionable/repeatable advice: - find a company which is at least two decades in the business they are in - preferrably some niche thing but not required - make sure books look good (the barometer here is whether you would invest part of your 401k in the company) - less than 100 dev/qa combined preferrably

when you start look for patterns, specifically things that no one wants to do. there will always be that. if you hear “don’t touch this code” spend months studying it and understanding it (on your own time). make sure to understand FULLY every aspect of the business your company is doing, suck up all the domain knowledge… every production issue, no matter how big or small you will volunteer to help and go through all post-mortems… on this path things will become clearer and clearer what you need to get done as time passes.

soon enough it will be YOU that is a go-to person for everything, most important customers will know you by name and will call you on your cell instead of creating ServiceDesk ticket…

above all - never look at your career as an employee - you are a corporation and your relationship with your “employer” is partnership between two business entities. no one bats an eye if company X pays 7-figure yearly to company Y for its services… and yet compensation for an “employee” in those figures would be like “huh? no way.” and yet just how Jira might bring enough value to company X to justify 7-figure compensation, so can your services as well if you position yourself appropriately. no one will teach this in America as America needs obedient employees…

> you can set your career such that you are more valuable to them than they are to you. and once you get that money is no object no more… and you get to stay decade (or two) in the same place vs doing the leetcode-change-jobs-like-socks dance.

How many of those companies that know your name will pay you straight out of college $160K-$200K and up to $400K-$500k+ within 5-7 years?

If you want to be at a company where everyone knows your name, it’s going to be a small company.

How many of those companies that know your name will pay you straight out of college $160K-$200K and up to $400K-$500k+ within 5-7 years?

the days of $160-200k out of college are slowly but surely coming to end (try to find some right now and see how many there are compared to say few years ago). Hard to compare but I started in 1999 at $117k which is “there” inflation-adjusted but overall I would say it is not easy starting in that range. $400k-$500k in 5-7 in definitely achievable, little bit of brain and little bit of luck.

All of the BigTech companies are still hiring - just in lower numbers. But so are the non BigTech companies. So if you are looking at a smaller pool either way, you might as well shoot for the moon.

Even if they do have to “settle” for enterprise dev, they still need to practice for DS&A to increase their “luck surface area” to get into those companies that pay top of range.

BigTech in a nutshell :)

https://news.ycombinator.com/item?id=42652056

I feel you dude, I definitely am not comp maxxxing, and it's because I simply don't want to run with the horses at this point in my life. I've had enough banality of the corpo world for now.

Glad to hear there are still some old dogs like us making good money. :)

Reply on news.ycombinator.com