How exactly does that work?
AWS/GCE are not cheap. They're cost-effective for certain classes of use-case, but beyond a given scale, or with certain base-level requirements, you're way better off colocating a bunch of a machines.
Plus they were a pretty early-adopter of AWS, and they have a ton of infrastructure running on it already.
If you're comparing some servers in a 4th tier colo with a poor SLA to AWS and pretending that devops time is free, then sure. But if you're actually building something with comparable reliability and availability, then no way. You're going to spend a lot more than you'd expect on extra capacity, a team of service engineers, load balancers, multiple locations, etc.
This has nothing to do with twitter's reliability. It just means companies have said twitter is more reliable than they are so look for our update on twitter opposed to some company specific system status page.
Your SLA will likely be better than with public cloud.
There's obviously room here though because AWS definitely takes a margin and that means opportunity!
It is a function of complexity, but complexity is often necessitated by scale. More users means more problems with scale, which leads to increased complexity to address the problems, which leads to a larger team needed to maintain the software.
That being said, 50 engineers seems enough to me for maintenance.
Security, abuse / harassment, fraud, community interaction, developer relations, 24/7 maintenance & ops, APIs, marketing & PR, legal, advertising, physical security / cleaning etc (whether outsourced or not it's a cost), secretarial, accounting, human resources & recruiting, sales, engineering, various management and so on.
The notion that you could run a $2.5 to $3 billion sales business with 100 people is a very bad joke. To deliver on that kind of sales level, you need more than a hundred people working in your sales organization alone. There are only a few types of businesses that can operate that thin at that size of sales, most of which are in the financial world.
Twitter doesn't need to cut down to 100 people to become nicely profitable, they need to cut down to 1,000 - 1,500 (while reducing infrastructure costs). Facebook delivers $1.5 million in sales per employee. Twitter's overall business is less complex and easier to operate than Facebook, they could deliver a higher ratio.
However, there's a Presidential Debate tonight so you know Twitter's traffic is going to be off the charts.
I don't think it's comparable.
It does have less features, but many of Twitter's are of dubious value.
One of the notable things about Twitter is that they're a huge marketing machine. They have hundreds of employees that handle sales, marketing, advertising. Hell, they've acquired startups that handle advertising analytics, and have brought them in-house to build their own platform. They now directly sell contracts to TV channels and businesses that allow them to view how their brands interact on Twitter.
I'd argue that this has helped to create bloat. Anyone that has had to use Twitter's API over the years can see that they've gone from being an incredibly open company to building up a walled garden to protect their brand, in a similar way to how Facebook scaled back their efforts to be an application/gaming platform.
If I were Twitter, I'd give teams the opportunity to split from the company, but to continue to use their platform. If they believe in the product they've built within Twitter, they can be happy to compete against the open market. It allows them to drastically cut their workforce in key areas, while re-opening their product to the world. They can continue to be a sales company, but can fully allow others to continue to build great things on their platform, or to help shape how it works in the next decade.
From my experience, this type of re-org leaves the resultant company unsustainably top-heavy. I recently left a company who downsized but refused to hire anyone "junior level" because their pivot "requires all stars". And we all know, when everyone is a senior, no one is...