Whenever someone says "ops doesn't matter" I'm reminded of Friendster's huge early-advantage lead that was squandered because they couldn't keep the site up for more than six hours a day.
The next year, MySpace started getting popular. MySpace didn't do all of this complicated social graph calculation at all. It would just tell you if you were friends or not. Naturally it was way faster, and despite other advantages of Friendster, it didn't take 10 minutes to load a page, and it took over completely until Facebook.
Whenever you have a double negative, flip both:
"I can imagine it would be"
If that changes the meaning too much, you can say:
"I can imagine it might be"
> I can imagine it would be
No. That means something completely different. They're saying (probably hyperbolically) that they cannot imagine the first of two possible situations, not that they can imagine the second.
It's the speaker's or writer's task to iterate for a half second to find whatever non double negative sentence fits what they want to say.
English isn't code. You can't say double negatives in conversation or writing, and you have to find the closest thing without using a double negative.
It is in fact not the case that you can't say that. Double negatives, like sentences with a excessive number of conjuctions, are perfectly grammatical (when they don't violate some actual grammatical rule like the one "You do can say that." does); rather, they are in most cases stylistically bad, for much the same reasons that:
stop = 0
while(a() && !stop):
b()
if(!(c==d || !e())) stop = 1
// should be `if(c!=d && e()) break`
if(!stop) e()
Use double negatives less.
The end.
We just have additional options now, which are actually cheaper to start, as you can get a lot of usage out of AWS' free tier.
Sure, once you get to "scale up" it's probably cheaper to do everything on-prem still, but that story is much better too with things like Golang, Docker, Elixir, k8s, etc, etc to make managing/building web-scale infrastructure much easier.
There were no smartphones in 2005 and everyone used the web to access "social media". Now any social media is nonviable unless it has an app for both of the smartphone platforms.
Youtube and streaming video didn't exist in 2005. Now it's table stakes for social media to allow people to share videos.
The bar has been raised by at least an order of magnitude. Sure, it might be easier and/or cheaper now to do what Facebook did...in 2005. But that won't get you anywhere in 2023.
The network? Not really. As a matter of fact, I don't think a new perfect FB would even be used today; people are only there because the people they want to reach already are. People today want a faster social experience (a la TikTok)
Good example for this is Threads. Everybody I know and care to follow are still on Twitter. Threads was cool the first week or so but after that Twitter was too strong of a magnet.
I don't know how a new network gets built these days simply because the major ones that exist are already so engrained.
Politics and whatnot of Twitter aside, of course.