Hardware != Software, but there is a huge overlap.
Some Dunning-Krueger candidates believe there is a fundamental difference, and that they can distinguish between them. 
Such people don't know what is inside a modern x86 processor or an FPGA, amongst many things.
I was going to walk away from this stupidity, but I just can't seem to let it drop. You claim that there is no "fundamental difference" between hardware and software. How far down are you going in order to make this claim? Yes, they are both called "engineering" and both require some common skills, but they diverge rapidly above that level. Are you talking about the activities, or the education?
As for the activity, here are a few design examples from my career, roughly ordered from all hardware to all software (not in time sequence):
- Hearing aid design (audio amplifier, filter, compressor), all analog circuitry
- "Wireless microphone" style radio transmitters and receivers
- EMI design and mitigation
- Switching power system noise analysis and cure in a timeslotted multiplexer
- Acoustic test equipment (some analog, some digital, no software)
- DS1 and DS3 framing, synchronization and switching ASICs (are you calling Verilog "software"?)
- 6800 processor board hardware
- Bootcode (68HC11) for system linecards
- Analysis of digital clock generation methods and time/spectral performance
Sure, some of those things could be done better digitally now, using software, but that's not an argument that there is no difference.
And I agree that a useful engineer should probably get experience in hardware and software, and avoid early specialization (I managed to avoid it my entire career).
Overall all you have done is mentioned a few spot examples; you have failed to supply a test that can be used to distinguish between hardware and software. Without such a test the differences are small
Apart from that, you have made many wrong and strange statements...
I've never thought of an electromagnetic field as "hardware"; it is physics! Even then, software is used to mitigate radio EMC problems and some EMI problems.
As someone that wears a hearing aid and expects to be eligible for a cochlear implant in one ear, I a, familiar how much you don't know about hearing aids. My first hearing aids had very limited "programmability"; my current has stunning capabilities partially implemented in "software". Cochlear implant controllers are even more sophisticated.
Is a neural net with all its internal weighting factors hardware or software? Consider Igor Aleksander's WISARD and Tesla's car controllers - especially the "prototype <cough> full self-driving" system.
I'm not familiar with the details of verilog, but VHDL certainly is software by most people's definitions. As wackypedia puts it "The VHSIC Hardware Description Language (VHDL) is a hardware description language (HDL) that can
model the behavior and structure of digital systems at multiple levels of abstraction, ranging from the
system level down to that of logic gates, for design entry, documentation, and verification purposes. ... The key advantage of VHDL, when used for systems design, is that it allows the behavior of the required system to be described (modeled) and verified (simulated) before synthesis tools translate the design into real hardware (gates and wires). ...
VHDL is a dataflow language". To aid that, VHDL has many attributes not associated with hardware, e.g. constructs to create/read/write files in an operating system.
As for 6800s, why not consder the contemporary AMD2900 family where microcode is a key implementation concept. Is the microcode hardware or software?
For a more modern example, consider Intel x86 processors from the P6 onwards. How do you think Intel changes the operation of its processors in installed systems? Is Intel Microcode hardware or software?
https://en.wikipedia.org/wiki/Intel_Microcode If you think x86 processors are hardware, how do they,
when powered down, detect and respond to "wake on LAN" packets to apply power to themselves to start operating again.
Analysis of clocks and time/spectral performance is no more hardware than systems modelling.