In most cases, people want to do meaningful work and see their work having real impact on the world.
If you spent all your waking hours, eating ramen all the while, building something that has outsized impact on the world, you won't be burnt out - quite the opposite!
OTOH, you can even get burnt out doing almost nothing all day, getting paid six figures if your work got thrown away into the trash can regularly (had no impact at all on anyone).
Burnout happens when that expectation doesn't match reality.
> these companies just sound like another high-performance meat grinder
> but are constantly under scrutiny. And smaller companies are all going pseudo-agile to try to squeeze every last ounce from their developers
As bad it sounds, high-performance, being under scrutiny, squeeze every last ounce aren't really the root cause of burnout - these are great symptoms of a broken process that leads to unmet expectations that then causes burnout.
The broken process usually are:
Scenario 1: At big company, BigCo, your work rarely has a direct impact on the customer, or when it does, it could have been years since you actually made the change that makes it way to the customer if it ends up that route at all or you have no way of getting customer feedback.
Result: You question yourself whether working so hard or putting in that overtime was really worth it because you have no idea where that effort ended up
Scenario 2: Someone ambitious, NewGun, at BigCo figured out a shortcut to a promotion - build a new product which is "better, faster" than an existing product CashCow. BigCo is too afraid to make major changes to CashCow (because it is the cash cow) and there's already a layer of management that's known to steward CashCow. NewGun is an outsider so even if NewGun pulled off those major changes to CashCow, they won't get most of the recognition - the CashCow stewards will.
Action: NewGun convinces BigCo to give them a bunch of devs, works them to the bone (because they need results yesterday), skips actual customer research and discovery (because that takes too long and they need results yesterday), makes up usecases and fictitious users
Result: Product flops badly because it doesn't appeal to anyone real. Devs question themselves whether working so hard or putting in that overtime was really worth it
Scenario 3: This is a very close cousin of Scenario 2, except that NewGun is a fresh entrepreneur who convinced some investors to give them money. Result is the same
Micromanagement, scrutiny, squeeze every last ounce are all symptoms of bad management. They by themselves aren't a strong differentiator - great management can do scrutiny, squeezing as well when it's tactically critical, but that's the exception, like pulling the handbrake to avoid an accident, than the norm.
The real solution to not being burned out: work with a team that actually knows what they are doing and have strong fundamentals. How do find those teams is perhaps a separate post as this one's way too long already