If $f$ is irreducible over a perfect field, then $f$ has no multiple zeros. 
If $f (x)$ is an irreducible polynomial over a perfect field $F$, then $f (x)$ has no multiple zeros$(1)$.

But I have also $2$ theorems, which are contradictory to the statement above

$\textbf{Theorem1}$: Let $f(x)$ be an irreducible polynomial over a field $F$. If $F$ has characteristic $p$, then $f(x)$ has a multiple zero only if it is of the form $f(x)=g(x^p)$ for some $g(x)$ in $F[x]$.

for example $f(x)=x^{4p}+3x^{2p}+x^p+1$, then $g(x)=x^4+3x^2+x+1$

$\textbf{Theorem2}$: Every finite field is perfect.

So every field of characteristic $p$ is finite, hence perfect, combined with $(1)$, Theorem $1$ is wrong, or not ?
 A: $\newcommand{\F}{\mathbf{F}}$I believe you need to consider the following classical example.
Let $x, y$ be independent indeterminates over the field $\F$ with $p$ elements, $p$ a prime. Consider the fields $E = \F(y)$ of rational functions over $\F$ in the indeterminate $y$, and its subfield $K = \F(y^{p})$.
Then one can prove that the polynomial
$$
f(x) = x^{p} - y^{p} \in K[x]
$$
is irreducible in $K[x]$. And clearly $f(x) = g(x^{p})$, where $g(x) = x - y^{p}$.
Note that $f(x) = (x-y)^{p}$ has the single root $y$ in $E$, with multiplicity $p$.
The latter fact allows for an easy proof of the irreducibility of $f(x)$ in $K[x]$. In fact, if $h(x) \in K[x]$ is a monic, proper, non-constant divisor of $f(x)$, then since $K[x]$ is a UFD $h(x) = (x - y)^{k}$ for some $0 < k < p$. The coefficient of $x^{k-1}$ in $h(x)$ is $- k y$. Thus $-k y \in K$, and thus $y \in K$, which is easily seen to be a contradiction.
A: There is no problem. There is just no irreducible polynomial over a finite field of the form $g(x^p)$.
