Plus, the problems you need to solve get more complicated, you have new compliance needs, paying customers are demanding new features that may require changes across your entire stack, performance is more of an issue (saving 3% of your compute costs didn't matter when you were paying $500/month for infrastructure, but saving 3% of 10 million dollars/month starts turning into real money) and writing and deploying software used by hundreds of millions of people is literally orders of magnitude harder than software a few million people use.
If you go to their about page, they show a little table, where they have 163 million monthly active users and only 966 employees. Not sure how good this comparison is with Google and Amazon because the kind of tools Amazon builds vs what Automattic builds (WooCommerce, Wordpress) are totally different.
Simply making a process 2% more efficient can save hundreds of thousands of dollars a year. Or increasing advertising conversions by 1%. etc etc.
There aren't many technical challenges I have encountered, when reduced to their smallest functional dimensions, that couldn't be fundamentally addressed by a single engineer. Obviously, efforts like writing an entire (practical) operating system are impossible for a single engineer to accomplish, but even then a single engineer could lead the entire effort or massive parts of the effort (e.g. the Linux kernel).
Conversely, I have seen far too many cases of "too many chefs in the kitchen" when it comes to a software project. This is the end of the spectrum where you end up stuck on Docker with a stochastic and otherwise insane architecture that isn't even consistent within each product, much less across the entire organization.
Say that a tweak to the page improves ad revenue by $50M. 98% of tweaks have neutral or negative effects. The expected value of the tweak is thus $1M, well over one engineer's salary. You need to have 50 of them going on in parallel to have a good chance of actually finding a successful tweak, though. (Well, technically about 150 to have a 95% confidence that at least one will prove useful, but there's also a good chance that more than one will be useful then.) That's what all the engineers are doing - independently implementing different changes and discarding 98% of the results.
When I did this kind of work, the vast majority of engineering time was actually spent measuring the results. It was not unusual for the actual implementation work to be done in about 3 hours, but the experiment design, approval, configuration, logging, data-collection, and analysis to take 3 weeks. And that was with a fair amount of existing tooling for measurement & process.
Hopefully Windows 10X puts an end to the "10x developer" meme.
Majority of its business value goes to real estate owners.
Also see the other child comment: Once you have a significant amount of revenue, investing the effort improve things by 0.1% (or even 0.01%) can have a not-insignificant impact on the bottom line.
WeWork has 466 thousand members.
The scale is quite different.
If they were a normal company, they’d focus on the top x projects that generate value and spend as little as possible on the rest.
I wouldn't be surprised to learn that the basic twitter.com tweeting/commenting/retweeting system isn't remotely the most complex system that Twitter even has.
If I were running Twitter, I would freeze the feature creep, have a robust security and site-reliability team, perhaps make small improvements to UI/UX by eliminating friction/impedance. Similar to how Craigslist and surprisingly how Ebay is today - it is pretty much the same experience as 2008, I vividly remember. Instead, engineering resources can be channeled to new ventures (or augmentation projects) for additional growth.