one of the strangest things about working in tech of the concept of doing things "at scale"
doing things "at scale" is hard to convincingly simulate without working for a company that is very successful
once you do things "at scale" you can talk big talk at conferences about really simple problems that probably anyone who knows things about computers can solve
however, once you've done stuff "at scale" people suddenly think you must know all sorts of incredible secrets about computers...
this is all exacerbated by the way that hiring works in tech -- unless you're an ivy-league comp sci grad, you end up only doing things "at scale" by accident because you lucked out in a startup, and once you've done "at scale" people give you more "at scale" work...
but it's all just computers... it's all just the most basic shit, except bigger and more pedantic.
this world is silly.
@terra you must be quite smart. Some stuff was quite surprising for me, like figuring out that key load factor may not be concurrent users to name one.
What I find ridiculous is people making huge kubernetes clusters to deal with scale and throwing cash at Amazon.
Aside from twitter (rip) and probably mastodon (how does it even work without collapsing _all the time_ being a ruby app), a glorious example was early whatsapp. One FreeBSD server, patched to remove socket limits, erlang, 3M concurrent connections. ❤️
Just computers? Sure! Impressive? Yes. Different from most just-computer kind of activities? You tell me
