I’d say P vs NP is the easiest, Riemann hypothesis is understandable in the sense that its statement is not too technical, but understanding why it is important is another matter. NS smoothness is probably understandable to the average math or physics nerd who has seen some PDE’s. The rest require developing more machinery to even state the problems.

Navier–Stokes existence and smoothness

Is actually every easy to understand. NS is the basic fluid dynamics equation, it describes how fluids move.

Never mind the actual equation, you can think of it as “speed” depending on “pressure” and “inertia”. We can use it in parts, in CFD, which is splitting the problem into very small pieces and calculating the pieces and then adding them up bit by bit. But that is super expensive to calculate, not as precise as we would like and difficult to understand.

We would really prefer an “analytical” equation like position = speed * time. The process to get there is usually the integral operation. That’s what the problem is all about.

The problem with the equation we have, is that “pressure” and “inertia” variables are so mixed up in the equation, that we can’t do that integral operation on the equation we have. You end up with something that’s like

y/x = (y/(x-y)) + x

and

x/y = (x/(y-x)) + y

Idk if that explains anything if I put it like that… You can just look at the whole thing https://en.wikipedia.org/wiki/Millennium_Prize_Problems#Navier–Stokes_existence_and_smoothness

Both sides of that equation describe the same thing, one with forces and pressures, one with measures of speed and time. But BOTH sides contain both space and time parts.

The pieces are interlocked in a way, where we can’t isolate variables, can’t get an integral.

Understanding NS itself isn’t that hard either, but I couldn’t do it just from wikipedia, if you have a bit of help guiding you through it, it’s not that complicated.

The problem is finding / proving that an integral exists and is smooth.

As for how hard it is and how useful it would be, the 1$ million are a joke. The solution is worth billions and billions.

And the problem is old.

The equations were developed over several decades of progressively building the theories, from 1822 (Navier) to 1842–1850 (Stokes).