It's the natural result of "fire the bottom 10% every year".
If that's the rule in your organization, and you have a core group of people that actually know the systems and get the work done, you better make sure you have 10% padding every year, lest you layoff someone important and their friends all quit in disgust. And since competence and institutional knowledge is built over time, that implies a revolving door of new folks coming in and most of it not making it.
The easiest way to become a F500 company is to start as an F5 company, then financialize your whole business, stop innovating, fire anyone effective, and give the executives huge bonuses.
This is effective. Therefore, normalization of this plays into the workers' hands, gives them information, and gives economic advantage to honest agents.
I mean, you could compare it to any non-capitalist society, where such treatment of workers is declared unacceptable. But what does this translate into in reality? Such strategies are still effective and provide an advantage to those actors who adhere to them. But since firing workers for their relative effectiveness contradicts the proclaimed ideology, such workers are simply accused of random crimes against the country and executed.
Usually, companies value opex more than capex - opex is much more flexible. That's one of the reasons why printers, coffee machines, companies cars and other things are typically leased.
Not really that simple. Opex gets better tax treatment (you can deduct it in full every year) but people aren't always Opex depending on what they're doing (research tends to be Capex)
Also tax treatment isn't the only consideration for financial engineering: It's easier for a company with a huge capital spend to argue that they're investing in the future and CapEx doesn't hurt EBITDA. On the other hand, some companies get worried about reporting a high "capital ratio" (ratio of capital assets to income).
In reality you can't say categorically that companies prefer Opex to Capex.
Concretely speaking, the FAANG companies are all wildly slashing opex (us) for capex (data centers, TPUs) even as the capex costs skyrocket due to demand outstripping supply.
Programmers are usually capex to be honest, with current tax treatment. When companies switched from data center to cloud they ended up shifting a lot of their compute from capex (buying servers) to opex (paying a hyperscaler for compute by the hour).
Of course if you're in one of the five tech companies building datacenters rather than renting (MSFT, Google, Facebook, Amazon, Oracle) then things are different.
If that's the rule in your organization, and you have a core group of people that actually know the systems and get the work done, you better make sure you have 10% padding every year, lest you layoff someone important and their friends all quit in disgust. And since competence and institutional knowledge is built over time, that implies a revolving door of new folks coming in and most of it not making it.