Hacker News Reader

Comment by myth_drannon | Hacker News Reader

"...and the engineers doing the due diligence were very young and were confidently belittling what we had built. Our average engineer was 40 years old,..." - That's so Meta. I wouldn't be surprised if they tried to ask them leetcode hard questions during due diligence.

View on HN

That explains why Waze works so much better than facebook or google products. Because people were hired based on actual competence not high school level "leetcode" questions that serve no purpose.

Do we know those 40yos didn't go through a programming puzzle hiring gauntlet?

(My own experience is that those I've worked with who are the best veterans also enjoy algorithms etc too. As do I. Shrug?)

In the last several years I've done work for some three letter agencies and fortune 500s. None of them required a technical interview really, only about half required a social fit interview. I'm starting to think my experience is really unusual.

Edit: I just remembered I did a piece of work for a power company. The technical lead talked to me before hand and told me the one technical question he was going to ask me during the otherwise social fit interview.

I think there is a lot of the tech world that isn’t quite the stereotype / big tech type world we think of… it just doesn’t get much attention.

FAANG / BigTech is the exception in almost all areas. Scale, pay, interview, just about everything. Some for the better - anyone reading HN can pass just about any non-BigTech (or companies trying to pretend they're BigTech) technical interview without a ton of studying beforehand, especially once they have breadth of experience. And some for the worse - BigTech pay is also the exception, the median programmer salary in the US is something like $110k. 99% of companies give you zero equity, and in my anecdotal experience only 10-20% of companies give any sort of bonus with any sort of regularity.

Work life balance too.

I hear folks talking about insane hours and feeling pressure to work them.

I uh… work when I want… and that’s not uncommon from other folks I talk to.

I think big tech and squeezing out every last dollar in compensation gets folks into a bit of a skewed world where their employer wants to do the same and they don’t always know it.

I’ve never been subject to those crazy programming questions either and I’ve done work for some big companies including Discovery Channel. They were all mainly culture fit. Granted it’s been a long time since I’ve interviewed anywhere. I do however interview people and make a point to not try gotcha questions or leetcode. It’s a how would you do this. Make it clear as long as there’s a solution there’s no wrong answer. I just want to know how they think and approach problems. And not made up hypothetical will never happen questions. Day to day what they’d be doing kinda things. 90% of our interviews is conversation.

The only time I was asked such "leetcode" interview questions was for a contract at a company that coincidentally went under a few months later. Asked something about css colours in HEX and doing some magic stuff with them. When I asked what is the relevance to the task at hand they told me this was a question a friend of theirs was given at a google interview. Hilarious.

I've done more than a few software/"DevOps" engineering interviews at DoD contractors and the only three practical technical questions that I was asked across _all_ of them were 1) what is a CSS selector, 2) how would you join an array of strings into a comma-separated list in C#, and 3) what Linux command one would use to determine disk usage... not exactly LC medium-level problems. These types of places don't seem to be big on even attempting to rigorously test for technical acumen - as long as you had a pulse and sounded like you knew what you were talking about they were happy to extend an offer.

Waze almost certainly hired/s using the same algo based interviews as every other FANG+/FANG adjacent company.

In my experience both Meta and Google's products work great from an engineering perspective. Low downtime, reasonably performant (with an adblock), etc.

It's just that they are designed to maximize ad revenue rather than to provide a good service for the user.

Facebook works great from an engineering perspective? We have entirely different experiences then.

I can't even count the number of times a photo didn't load, a timeline didn't refresh, a like didn't register (or worse, got clicked on accidentally and then didn't unregister).

Google's products have their bugs but the main flows work. That's not the case with Facebook.

Facebook moved name updates to their account center and there are still incredibly basic bugs.

My girlfriend can’t update her name because:

You can't change your name right now because you've changed it in the last 60 days. You can still change the order. Your last name change was on December 31, 1969.

Just compare Facebook to WhatsApp.

WhatsApp is crazy good, and doesn’t make money. Facebook is… what you describe, but very very profitable.

I think you can see the difference in priority in the product.

WhatsApp made hundreds of billions for facebook by preventing rise of major competitor.

It probably helps that WhatsApp's backend is written in Erlang, which excels in fault-tolerance and reliability.

https://www.erlang-solutions.com/blog/20-years-of-open-sourc...

WhatsApp does make money. About $900 million in 2022. It's a small part of Meta's total revenues, but the margins are really good because of how well built and efficient it is.

These are low priority issues. Real issue is how many times did an ad not load up. And a click to an ad not register.

Prioritie$.

And just today I learned that the CEO of the failed sub exploring the Titanic said that he didn't want to hire 50 year old white guys. What a tool. He said they weren't "inspirational."

I would be highly skeptical of any engineer making grand generalizations about a product without actually working with it for a fairly long time. It’s SO easy to belittle code without really knowing the whys and history.

I get that is what their job is when doing due diligence… but I think it’s also a little bit impossible.

I know I have thought of some code as poor, rewritten it “better” only to find my code is insufficient and quickly evolving into the same code I thought was bad…. for all the same reasons that influenced the original code… and thinking much more of the old / bad—- proven code.

There are still some other metrics and heuristics you can use, such as number (and coverage) of tests, overall code metrics, number of compiler warnings, that tell you a little bit about the care put in maintaining the codebase. You can still have quality codebases without those (do I know...) but then you'd have to look at critical business metrics (uptime, latency, satisfaction surveys) or the number of tickets (even there, beware) or the overall 'cost and duration to deliver a fix or a new feature'.

As for architecture, it's harder to find generic metrics and guidelines for 'quality'... I won't dive in there but external evaluation by people with no or little idea of the actual domain is not very interesting. Moreover people tend to be biased by either 'we at x have done it like this, and we're successful, so...' and 'we at x have built shit' and it's very hard to get in a sort of clean-slate mindset.

There certainly are things they can evaluate.

Just that I think big picture wise I would be hard to draw too much of a conclusion from the testable metrics.

Personally I’d want to test the app itself and if it works / customers are happy, I might worry a lot less about metrics and other decisions.

> I wouldn't be surprised if they tried to ask them leetcode hard questions during due diligence.

Based on my experiences interviewing at Google, I wouldn't be surprised if they did that too.

I mean, this is a valid point though? I would not take any stock in a code review or due diligence performed by a 20 year old.

Personally I would hope for a mix of experience when it comes to a due diligence team… but I would lean HEAVILY towards more experience.

When I interviewed at twitter, ~2007 they asked me a bunch of questions which they praised my answers for NOT being "...Well, they way we did it at facebook..."

They were happy that I wasnt a young engineer who thought that the only way to do it was the single exposure of just having done it at FB...

Was your previous employer FB at the time?

NO... MY FUTURE EMPLOYER (sorry for caps)

>That's so Meta.

That's so meta.

Reply on news.ycombinator.com