In his book on differential equations, Arnold writes that $x'(t)=x(x(t))$ is not a differential equation.
My question is: how can one solve it?
In his book on differential equations, Arnold writes that $x'(t)=x(x(t))$ is not a differential equation.
My question is: how can one solve it?
Here is a partial attempt: Let $x=At^n$ then $\displaystyle {nAt^{n-1}=A^{n+1} t^{n^2}}$. From which you get $n^2=n-1$ and $A^{n}=n$ resulting in some complex number answers that would need to be analyzed for being meaningful.
A general solution would be interesting though. The equation reminds me of a "delay differential equation" but with variable unknown delay.