We are given that $55$ is a solution to $x^ 3 − 9 x + 8 \equiv 0 \pmod {2^{10}}$.
Find a solution to $x^ 3 − 9 x + 8 \equiv 0 \pmod {2^{19}}$ that is a lift of $55$.
I was going to try lift the solution to modulo $2^{20}$ using Hensel's Lemma, but I can't use that here because $f'(55) \equiv 0 \pmod 2$, so the method fails. But then if you check $f(55)$ modulo $2^{20}$, you get something that is non-zero. Doesn't this mean there are no solutions except $55$ itself?
Any help would be really appreciated, thanks.
Here is a simple computer search using sage.
So the lifts are there, and they are two for $55$, this answers the question.
Some comments on this maybe.
The given polynomial $f=X^3-9X+8\in\Bbb Z_2[X]$ reduces modulo $2$ to $\bar f = X^3-X=X(X-1)(X+1)=X(X-1)^2\in\Bbb F_2[X]$, it has a double root in $1$, and we want to lift both $0$, and the double root $1$. The derivative $f'=3X^2-9\in \Bbb Z_2[X]$ has then the values $f'(0)=-9=1\ne 0\in \Bbb F_2$, and $f'(1)=3-9=0\in\Bbb F_2$. As seen above, we have a unique lift of $0$ from $\Bbb Z$ modulo $2^1$ to $\Bbb Z$ modulo $2^{19}$.
What about lifting the $1\in \Bbb Z/2$?
First, we have $X^3-9X+8=(X-1)(X^2+X-8)$ over $\Bbb Z$, the $1\in \Bbb Z$ will remain root modulo any power of $2$, so we will look as an intermezzo at the quotient $g=(X^2+X-8)$. Then $g(1)=0$, $g'(1)\ne 0$ modulo $2$, so we expect unique lifts. And here they are:
Sorry for this, i would have hated such an intermezzo some years before.
The lifts of the root $55$ of $g$ start with $1 + 2 + 2^2 + 2^4 + 2^5 + \dots$ .
Now let us look at the lifts of $55$ for $f=X^3-9X+8$. They are:
Sorry again, it this is annoying, but this experimental issue makes one thing clear:
Indeed, working modulo $O(2^{N+1})$, $$ \begin{aligned} f(s)-f(t) &=(s^3-9s+8)-(t^3-9t+8)\\ &=(s^3-t^3)-9(s-t)\\ &=(s-t)(s^2+st+t^2-9)\\ &\in O(2^N)\cdot O(2^1)\ . \end{aligned} $$