Conversation

To clarify, my point is that the marginal effort in getting to intermediate scales (say 1000 users) is almost the same as getting to maximal scale, since higher orders of scaling are mostly handled by infrastructure that's mostly agnostic to the specifics of the code
Quote Tweet
The annoying thing about computing is that every damn thing has to be live in an environment of scale. It's like if there were only 2 scales of cooking: cooking a meal for your family at home, and cooking for the entire planet. No intermediate scales of intermediate difficulty.
2
8
Sure if you're talking about platform code, at Facebook feed or Google search scale, the effort is more proportionate to scale, but this is a minority of code
1
3
Replying to
I think this is an artifact of older server-bound software. Serverless stuff is much easier to elastically scale. E.g. Gatsby :-) no more work to serve a million vs 10 people
1
2
Replying to
Well, static content is a very special case. Basic question is probably: does an app have different behaviors at different *social* scales. Sending 7 billion people a static message != getting them all to play Fortnite
1
1
Replying to and
Also, I think the breakpoint where scale effects kick-in is not 10, but whatever is the max the programmer can personally support/troubleshoot on a bespoke basis. When complexity forces you to switch from managed to DIY/automated service provisioning.