An ordinary equation like 2x=10 hides a number; a differential equation hides a whole function. dtdy=0.03y does not say what y is. It says how y changes: at every moment, y grows at 3% of its current size. You solve it by finding functions whose derivative behaves that way. y=Ce0.03t works for every constant C, since its derivative is 0.03Ce0.03t=0.03y.
That is why solutions come in families. dxdy=2x is satisfied by x2, x2+1 and x2−7, because shifting a curve up or down does not change its slopes. Each value of C gives a different curve that obeys the same slope rule. An initial condition picks one: if y(0)=3, only x2+3 passes through (0,3). A rate law together with a starting value is called an initial value problem, and it has a single answer.
Check a proposed solution by substitution, the way you check a root. Does y=e2x satisfy y′′−4y=0? Here y′′=4e2x, and 4e2x−4e2x=0, so it does. The order of a differential equation is its highest derivative, so that one is second order. Order is not about powers: (dxdy)3=y is first order even though it contains a cube.
Think of it like this. A differential equation is a recipe for change with no starting point: "every moment, grow by 3% of whatever you have". Start with $100 or with $500 and the same recipe produces different histories; together they form the family of solutions. The initial condition says how much you began with, and that picks out one history.
Core idea. A differential equation relates a function to its derivatives; solving it means finding the function(s) that fit.