Not really. The OS syscall interface, ABI and the fact everything is abstracted via your C compiler normalises the differences between the machines pretty well meaning you only end up dealing with portability issues.
Portability issues is where real hardware benefits. It's where you have battles of unusual register sizes, endianess, host/network order differences, different memory models and memory protection, different performance characteristics, different timings and different exploits.
Unless the emulation is 100% accurate, including timing, which is a really difficult thing to do (look at the effort MAME goes to), then the benefits over real hardware is moot.
Emulators are also expensive to write due to the above, have their own bugs and don't always recreate the bugs in the real hardware (which are sometimes exploitable).
There are varying degrees of corrupt and totalitarian.
We're not quite Russia bad yet, but we have however run "quick processing overnight courts" to deal with hundreds of people, thrown people in jail over Twitter jokes, kept people in jail for writing articles critical of the state, used extraordinary rendition like soap, murdered people critical of the government, shot unarmed civilians, started big wars, entertained numerous fascist dictators, operate large concentration camps to keep immigrants in, control media interests through the old boys network, force slavery on unemployed people by making them work for nothing and generally act like unsociable arseholes on the world stage.
This is why I still like old Unix systems (early BSD/v7 etc). It is possible to get a complete understanding of the system in a relatively short amount of time and they're still pretty productive.
Actually most can't be emulated because emulators don't exist and those which can typically can't be emulated correctly (Sparc emulators for example which are notoriously sparse and bad quality).
Some of the architectures also have different endianess and incredibly complicated peripherals to the cost effective host machines as well meaning that it's actually more power efficient to run native. A headless 100MHz VAXstation for example draws less power than the equivalent host that would be required to provide a full, accurate emulation with peripherals. These aren't arcade machines.
This fundamentally goes against the Unix philosophy though which is to provide small well-defined parts from which you can construct a complete solution from.
If you need a complex manual for a complex program, something is wrong.
In 2009, our development team lost a whole 10 hours to a degraded Linux mdadm RAID1 that wouldn't rebuild due to an obscure error after a digger severed our power and internet connection. No internet access as power came up first so no access to online help. mdadm is buggy. Documentation sucks. Error messages suck. Only recourse was a full restore from tape which took a long time. This was the last straw after over a decade of dealing with this crap from network dropouts, laziness, half-arsed features, distro wars, politics and churn.
Some previous Unix experience in the late 1990s with OpenBSD on an old SparcStation 5 (the only thing that would run on that machine nicely) jumped into my mind on the way home. It had that warm, fuzzy, well-engineered, well-documented feeling about it, like an old HP RPN calculator. Got home, downloaded it and installed it on my laptop, replacing Ubuntu.
4 years down the line: one happy person with the same laptop running 5.4 still with that warm, fuzzy, well-engineered, well-documented feeling.
Not once has it let me down. Not for a minute in the 4000+ hours I've been using it. It just works.
It's fine for surveying but for dynamic applications which are far more common, it's a pain.