Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Perhaps I’ve been unlucky but my experience at big tech has been marred by churn and burn culture - high pressure tactics to get the most work out of employees and discarding them when they inevitably burn out. Average tenure was 1-3 years. Amazon was particularly bad at this. The only modicum of power employees have is in quitting and now that they’re exercising it the employers are acting like their throats have been cut.

Edit: a story I like to tell people about what it’s like to deal with Big Co. is that I solo built a $50Mp.a. profit data science artifact in my first year. They offered me the highest once off bonus on offer $30K stretched over 6 years, no promotion (those were used to keep US citizens). I said, give me the same bonus as my manager for for ‘managing’ my project ~$150K in one year or I’ll quit. They said no so I quit. The artifact was extremely difficult to build and written in Scala, it took a lot out of me, no one understood it after I left. They rewrote it in Java which made it even harder to understand. It atrophied over 5 years and now it doesn’t even work at all. I’m in contact with my former colleagues and keep offering to fix it for 7 figures, but no dice. They’ve now convinced themselves it is impossible to do. That’s how committed they are to maintaining the status quo.



Amazon has always had a different culture than other bigtech cos. I have plenty of friends at Google, Microsoft, Snap, and Facebook who work 36 hours at week, don't stress at all about their jobs, and are paid really well.


The story wasn’t from Amazon... but yeah they do have a burn and churn culture.

I think it depends a lot on the group you’re in and where you land in the office politics. Microsoft used to be a nightmare in certain orgs and Intel used to be worse. I hear complaints about Google and Facebook at times, but those companies do seem to be more relaxed about things.

The roles I like tend to be well coveted and so are often the target of undermining. There is a process where one org wants to subsume the responsibility and budget of another org and so they will undermine the other org to ensure its failure. I was once forced to buy thousands of computers that were way over spec in order to blow out the budget and make it harder to justify the investment in the project. This is why we can’t have nice things.


Addendum; you tend to know going into these things that the reason such opportunities still exist in these dysfunctional orgs is precisely because they are dysfunctional. I guess hope springs eternal that maybe you could succeed where others have failed.

A former colleague at the same company optimized a core process saving $10M p.a. in compute cost. They couldn’t find a second person who knew assembly well enough to do a code review so they were never able to deploy it.


Slightly off topic, but could you go into a bit more detail on this artifact? I thought an artifact was something like a trained model, but you said it was "written" in Scala then Java, which makes me think I don't understand the meaning, and the word on Google has very broad definitions.


Data pipeline -> ML Model -> Dataset.

The dataset is what makes the money.

I should have written “the process that produces the artifact was written in Scala.”


an engineering artifact is the tangible by-product of a software development process. It could be a piece of code, dataset, specification, etc. Basically a deliverable.

https://en.wikipedia.org/wiki/Artifact_(software_development...


Thanks for the story, would love to hear more.

Everyone at the company probably thinks the well understood and documented and lovely Java version, which never actually worked, is the better version.


Thanks, well the Scala involved a pretty nice Hierarchical State Machines and an embedded DSL combinator for the custom rules engine. The Java process used Eclipse with State Machine plugins and god knows what for the rules engine. The Java version was an unwieldy monstrosity. It's creators couldn't abandon it fast enough. It was only two people tasked with replacing it, out of a pretty huge company, it really wasn't that big of a deal for the company so few knew about it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: