3 ms·

> The app is simple but doing simple things at scale is hard. I wonder if we'll see more downtime and issues now. So, here's what I've never figured out... On

> The app is simple but doing simple things at scale is hard. I wonder if we'll see more downtime and issues now. So, here's what I've never figured out... Once your app is built, and you've figured out how to scale it, and everything is autoscaling nicely, and you've reached the critical mass Twitter has...why do you still need all the engineers? Why do you need to keep growing? At some point, isn't the scaling solved? Once you have a platform that can handle 400M MAUs, what more is there to build? Does it really take 1,000 engineers to maintain? I'd think automation could take care of most grunt work involved with maintenance.

btgeekboy 4y ago
Once your app gets to that level, even keeping the lights on requires work. Third party dependencies don’t stop updates, security or otherwise. Laws and regulations change. Advertisers demand features that aren’t seen by the general public. Multiple SRE teams are needed to handle oncall rotations, outages, etc. And now that you have thousands of employees, you need to start building automated HR and IT tooling.
dstola 4y ago
I think the issue with these massive scale apps is that people want better and fresher features which are not always well defined and get lost in the maze of product development of a huge company. They have to make decisions, do we go with web3 stuff or do we improve current NLP system etc etc. Making decisions is much easier for a small startup than a behemoth of twitters size. And yes, it takes many people. 1000 engineers? Maybe not, bloat has set in at that point, but a few hundred for sure.