Hacker News Reader

Comment by edge17 | Hacker News Reader

Why does this article claim to know that Twitter might go down due to cost cutting? Isn't there a possibility that their infrastructure is just hugely wasteful?

Having never worked at giant web app company, if it were run efficiently then what should we expect their server and infra costs to be? Surely that is the type of Fermi calculation someone on this site can make.

View on HN

Right? The real cost cutting is usually “hey there’s this huge X sitting around that nobody uses 99% of the time” or “This process is 10000x faster after someone did some performance engineering” or “we’re storing this PB dataset a dozen times for no reason”

There are often lots of ways to cut the fat that have nothing to do with reducing (reliability related) redundancy.

There's also reducing replication from 4x to 3x which can save a lot but also reduce redundancy.

It is possible to simultaneously have unnecessary spending on infrastructure and then cut it in a way that makes whole thing unreliable.

Given large layoffs and general chaos, mistakes and bad cuts are much more likely. Especially when you git deadline and goal saving made by clueless executive.

> Why does this article claim to know that Twitter might go down due to cost cutting? Isn't there a possibility that their infrastructure is just hugely wasteful?

I see the authors (Sheila Dang, Paresh Dave and Katie Paul) quote "sources".

Reply on news.ycombinator.com