I have the following problem:
Let $f(x) = \int_{3}^{x}\sqrt{1+t^3}\ dt$. Find $(f^{-1})'(0)$.
I know that if $g' = f^{-1}(x)$ that $g'(f(x))f'(x) = 1$. Obviously when $x$ is $3$, $f(x)$ is $0$. So I worked out that:
$\displaystyle\ g’(f(3))f’(3)=1→ g'(0)=\frac{1}{f'(3)}→(f^{-1})'(0)=\frac{1}{\frac{d}{dt}[\int\sqrt{1+3^{3}}dt]}=\frac{1}{\sqrt{1+27}}$ $$=\frac{1}{\sqrt{28}}.$$
My question is: since $f(x)$ was originally an integral evaluated from $3$ to $x$, shouldn't $x$ be $3$ across the equation, making the attempt to take the derivative of it useless? Would this result in the solution being "not defined" since you can't divide $(1)$ by zero? Or, does the derivative effectively nullify the limits?
By the Fundamental Theorem of Calculus, you have $$f'(x)=\sqrt{1+x^3}.$$
Alternatively recall that $x=f(f^{-1}(x))$, so differentiating both sides of the equation and using the chain rule we have $$1=f'(f^{-1}(x))(f^{-1}(x))'$$
so $$(f^{-1}(x))'=\frac{1}{f'(f^{-1}(x))}=\frac{1}{\sqrt{1+(f^{-1}(x))^3}}$$
then since $f(3)=0$ we have $f^{-1}(0)=3$ (as you mentioned) and thus
$$(f^{-1}(0))'=\frac{1}{\sqrt{1+(f^{-1}(0))^3}}=\frac{1}{\sqrt{1+27}}=\frac{1}{\sqrt{28}}.$$