LaserToy

I know. I worked on Presto :)

The problem with Trino is that it is not that easy to scale to possible RPS of Clickhouse, it introduces tons a of latency and push downs are faaaar from perfect. Uber has a smart solution for Pinot, when they run it as a single node proxies

I’m not the author, but yes it might.

The regime is getting crazier and they eventually might start cleansing folks who said something online. Or their relatives. You just never know.

I kind of agree. The only thing that matters for me is a cost of ownership in $$$ over some period of time.

You didn’t have to do it, you know it right?

I’m really curious how it is going to play out. I’m a bit concerned that our salaries will be slashed, especially if engs from other countries enter the wfh market

It is matter of perception. All tools suck in some ways, all codebases are have bad stuff (except toy projects of obsessed perfectionists). We just deal with it :)

So, you are stating that Java language is not applicable to any situation.

Here is one: there is a database written in Java and you need to implement more features. There is 10000000 lines of code in the companies source and new code should be the same. I can go on and on.

Personal advice - if you want growth, drop preferences. When FAANG interviews nobody even asking what is your preferred language as nobody cares.

Because preferences are based on emotions are not rational.

I’m talking about picking a tool that does the job according to the requirements (speed, memory, debuggability), not based on what some YouTuber said.

We are engineers and paid for results. It help with carrier to embrace it.

Why would any reasonable person limit their choices based on a tool? Rust or Blast - whatever

I really don’t care what language I’m using if it it doing the job. I use Java, Java Script, Go, Python, Scala when appropriate.

I mentioned C as a language of choice. Cool kids are picking node js and esoteric Closure, not C

I’m talking about the language, not frameworks.

AOP sucks because it is paaaain to debug, proxies are magic. There is Dagger that solve some stuff

Verbosity is not a problem in large code bases. Speed of understanding and debugging of what somebody else wrote is.

So, less magic is better. I will take Java codebase over Scala or Python every time, unless it is a toy script which is 100 lines

Java is not hyped because it is a tool. Like a microscope, it just does what it supposed to do.

IMO, if you want to get a descent job in a large company working on backend systems, you want to learn: Java, C/C++ or recently Go. You will be able to build pretty much anything, from a micro service to self driving car.

If you want to stick to startups, you can do well with anything else till it growth, experienced engineers are hired and rewrite it in one of those languages.

I had a similar feelings, even though the circumstances were different. I worked in medium sized companies for 7 years before getting into FAAMG. I thought I’m a decent SWE who knows patterns, processes and can deliver high quality code.

First couple months there were hard. My PRs were teared apart with tons of nits, I had troubles communicating as I was embarrassed.

Dev process was also very slow with a lot of back and forth. I felt I was doing something boring + doing it so slowly I stopped caring.

It was miserable and I quit after a year. Looking back, I glad I did. The company did teach me something, but I also wasn’t compatible with their slow moving culture. Did some very interesting work after I moved on and my carrier skyrocketed.

You will be surprised, but I saw the same behavior in the large tech company.

Department either uses or loses the budget, so, there was a push to make sure nothing is left.

I’m, I just not willing to wait for them. It also tells me that they are not interested in the company, and the best people both good and interested. And yes, it saves money (both hiring and motivated folks produce more) - no surprise here

Fun fact, that when I’m looking for a job I will be arguing like you.

No, the offer might be good, but I’m just not willing to wait a month as there are other candidates in the pipeline.

Had similar marathon 8 years ago and it is draining.

Now, being on the other side (hiring folks for a startup), I would argue that some of flags are a bit misleading. Examples: “Undue pressure to accept an offer letter” - interviews are not free, companies are investing eng time into evaluating each candidate and once they find great one, they really want them to join. They also do not want to be part of a bidding war, so it is natural to impose limits (some folks do go overboard). Recency bias also tells us that if you sit on the offer for a long time and continue interviewing, chances of you accepting the first one are going down. And something I learned only recently - offer stage is also part of the interview, it shows how 2 sides will be working together in the future.

“ Not enough clarity about your role” - normal for growing startups. I tell the candidates - here is what we do, pick something you like and run with it. Of course it is biased towards entrepreneurial types who thrive in uncertainty, but that is the state of many startups.

Mostly agree with other examples. One fun thought, they had 60 interviews. Assume some were 1 hour phone screens and some 1 + 5 hours virtual on sites. If we do 50:50, the industry spent 30 + 30 * 6 = 210 hours of eng time + debriefs, outreaches, negotiations, ... I think statistically, it is not in the interests of companies to participate in such marathons as ROI will be low.

Send it to Texas

I’m the author, I’m not trolling and I gave examples.

CloudSql is a misleading name and a bad product.

Wow, is it google team that downvoting?

Ok, here are some: if you resize your instance you will get a downtime.

There is no support for CDC.

You pay for multi master but failover is painfully slow

If you launch more than one using terraform all except one will fail and will stuck in an unrecoverable state

They broke something last week so our proxies lost connectivity

Google Cloud Sql is a shit product. Ask me why.

I changed my opinion of XML after I had to deal with yaml.

Whoever thought it is a good idea for complex configs (kubernetes) was wrong

I hope you folks know what you are doing. If your screw it up you will have a lot of angry former customers, us including.

Maybe less aggressive rollout strategy?

Why can’t pay taxes to an entity I prefer? Isn’t it a service? So, if some entity can give me healthcare, I will pay them

If everything is free market, why do we need government?

I know the answer, do you?

I did read the book and i didn't find anything too toxic.

Especially is you compare it to other industry giants (hello stack ranking and sports team mentality.)

I’m curious - how much do you earn per hour and how many hours do you spend on cooking/shopping/cleaning per month?

Is anyone interested in working there? I know they are looking for strong SWEs. dm me (y<username> at gmail.com) if interested.

I was in eng leadership at largest gaming network. Our customers (gamers) absolutely didn’t care about reasons they couldn’t play. It was always our fault as we were responsible for tech choices, not them. And I believe it is a way. If I’m selling you a service it is my responsibility to make sure it is reliable and not to blame it on something that is under my control (like cloud providers). Of course I can’t fix local internet going down, but I can make sure I’m not married to 1 vendor

Agree, but it comes across as we don’t care. I worked for a major gaming network, and believe me, gamers don’t care why they can’t play. They want to play. Now. And I can’t imagine anyone on the team relaxing when something like that happened.