3 ms·

Oh, it is causing pain already. Recently I pitched to a friend fixing a component as a great project for a very junior engineer. Only to find out that neither h

by lrem 5y ago

Oh, it is causing pain already. Recently I pitched to a friend fixing a component as a great project for a very junior engineer. Only to find out that neither his nor my team has any sufficiently junior engineers on hand. So the oceans will keep getting ever so slightly boiled until we hire some, or one of the too senior engineers runs out of more important things to do. Sure it won't bring Google down right away, but having cruft like that accumulate kills efficiency in the long run. And Google thrives in cutting those epsilons.

PostThisTooFast 5y ago
Google also excels in turning off qualified hiring targets by acting like douchebags in the interview process. They admitted as much years ago (and two of my coworkers confirmed from personal experience) and said they were going to try to be less douchey, if I recall correctly. But the culture there is still douchey, as reported by people who DID tolerate the interview process and worked there.
vlovich123 5y ago
Google cuts epsilon’s in engineering cost that operate on efficiencies of 10s of thousands engineers or engineering components that underpin multiple high-value products. They have entire teams of very senior engineers dedicated to this problem. Conversely, a smaller project wouldn’t demand this kind of attention because solutions don’t scale well and so are managed as local costs at the most local level possible. What kind of scale project are you and your friend working for and which component is the problem in? Note that I’m not disagreeing with the consequences being that there are ever growing tech debt that will live on forever and you just try to manage that and fit that into “fixit week” or “20% time”. That’s kind of the discount for why they still pay the crazy salaries for all work, not just in the teams closest to the revenue generation.
- lrem 5y ago
  We're SREs working on the rollout system used for vast majority of internal software. The component in quesiton collected enough tech debt to start showing "unfixable" defects, but still not enough to justify anything more than "make it retry harder". Any year now it will become worthwile to quantify the actual impact.
  - fragmede 5y ago
    Sounds like a Sisyphyian effort to have the software retry harder (harder meaning in a different, more effortful way, rather than increment the retry loop counter).
williesleg 5y ago
Just do the needful like you always do.

by lrem 5y ago

PostThisTooFast 5y ago

Google also excels in turning off qualified hiring targets by acting like douchebags in the interview process. They admitted as much years ago (and two of my coworkers confirmed from personal experience) and said they were going to try to be less douchey, if I recall correctly. But the culture there is still douchey, as reported by people who DID tolerate the interview process and worked there.

vlovich123 5y ago

Google cuts epsilon’s in engineering cost that operate on efficiencies of 10s of thousands engineers or engineering components that underpin multiple high-value products. They have entire teams of very senior engineers dedicated to this problem. Conversely, a smaller project wouldn’t demand this kind of attention because solutions don’t scale well and so are managed as local costs at the most local level possible. What kind of scale project are you and your friend working for and which component is the problem in? Note that I’m not disagreeing with the consequences being that there are ever growing tech debt that will live on forever and you just try to manage that and fit that into “fixit week” or “20% time”. That’s kind of the discount for why they still pay the crazy salaries for all work, not just in the teams closest to the revenue generation.

lrem 5y ago
We're SREs working on the rollout system used for vast majority of internal software. The component in quesiton collected enough tech debt to start showing "unfixable" defects, but still not enough to justify anything more than "make it retry harder". Any year now it will become worthwile to quantify the actual impact.
- fragmede 5y ago
  Sounds like a Sisyphyian effort to have the software retry harder (harder meaning in a different, more effortful way, rather than increment the retry loop counter).

williesleg 5y ago

Just do the needful like you always do.