They’re expensive, risky operations of which development is only a part of.
This is probably why it makes the most sense to treat developers as a commodity in the industry. The level at which the consumer interacts with the developer isn’t as high as say, a website.
It's even worse when you consider that the engines are generally already made, you only need developers to do what you could almost call glorified scripting compared to engine development.
I think Unity and Unreal probably have to pay their devs a reasonable salary. But having devs work in Unity or Unreal to produce a game is not really all that hard. That's kind of the point of Unity and Unreal.
There are advantages and disadvantages for using these engines. They soften the "getting started" part of game dev for newer people, but for complicated projects it tends to be kicking the can down the road for dealing with deeper issues.
You'd be hard pressed to squeeze good perf out Unity while fully using the HDRP and not understanding the ins and outs of the pipeline.
I saw it more in that it's not the early '00's anymore where every company had their own custom engine. you can hire devs into a company with experience in UE4 or Unity and skip a significant amount of onboarding, especially if they are hired early into a project. There are certainly hard parts, but this means you can pay junior engineers to worry about "easier" parts of the game while letting the seniors/leads worry about architecture, infrastructure, and optimization. Not too unlike a front end web dev from what I've hears.
And for "street cred" I've worked in all 3 environments, unity projects, UE4 projects, and custom engines.
From OP.
Even with an engine game dev is much harder than front end web dev. The depressed salaries are due to the amount of passion game devs have.
Then again doing something in Unity is much easier than building an engine out. I credit Unity with teaching me to program.
That said, not all games are services yet so once the game ships anyone can make it to the chopping block.
Google/FB know that if they get the best engineers, they will have a top-tier product. But games don't have anything close to that same guarantee. You could have the most efficient, elegantly-written, perfectly abstracted game, but if the game-play sucks, it's not going to sell.