The limit of the alternating series $x - x^2 + x^4 - x^8 + {x^{16}}-\dotsb$ as $x \to 1$ I want to calculate the limit of this sum :
$$\lim\limits_{x \to 1} {\left(x - x^2 + x^4 - x^8 + {x^{16}}-\dotsb\right)}$$
My efforts to solve the problem are described in the self-answer below.
 A: From here is the amazing solution:

Since $S$ satisfies the functional equation
  $$S(x) = x − S(x^2)$$
  it is clear that if $S(x)$ has a limit as $x$ approaches 1 then that limit must be $1/2$. One might guess that $S(x)$ in fact approaches $1/2$, and numerical computation supports this guess — at first. But once $x$ increases past $0.9$ or so, the approach to $1/2$ gets more and more erratic, and eventually we find that
  $S(0.995) = 0.50088\ldots > 1/2$.
  Iterating the functional equation, we find
  $S(x) = x − x^2 + S(x^4) > S(x^4)$.
  Therefore the fourth root, 16th root, 64th root, … of $0.995$ are all values of x for which
  $S(x) > S(0.995) > 1/2$.
  Since these roots approach $1$, we conclude that in fact $S(x)$ cannot tend to $1/2$ as $x$ approaches $1$, and thus has no limit at all!
  So what does $S(x)$ do as $x$ approaches $1$? It oscillates infinitely many times, each oscillation about 4 times quicker than the previous one; If we change variables from x to $\log_4(\log(1/x))$, we get in the limit an odd periodic oscillation of period 1 that's almost but not quite sinusoidal, with an amplitude of approximately $0.00275$. Remarkably, the Fourier coefficients can be obtained exactly, but only in terms of the Gamma function evaluated at the pure imaginary number $\pi i / \ln(2)$ and its odd multiples!

A: We can define a rotation point when the curve switches its direction from increasing ordinates to decreasing values as x advances.

So when we try to locate those points in our polynomial, the derivative of this function must be 0 in that abscissa.
$$fn'(x)={1-2x+4x^3-......(-|+)2^n x^{2^n-1}}$$
This polynomial has (n-1) factors, which means that $fn'(x)=0$ has $(n-1)$ solutions which implies that there are $(n-1)$ deviation points in our graph.
When $n$ tends to infinity we have limitless deviations as shown in this graph:

Well, consequently, our limit is hovering endlessly between $0.5+λ$ and $0.5-λ$ with $λ$ extremely small -- it may even be zero!
Finally, we can say that the function is diverging, and the limit doesn't exist.
A: Based on a paper "Summability of alternating gap series" by J.P. Keating and J.B.Reade in year $2000$,
(an online copy can be found here)
, one can use Poisson summation formula to show
$$
S(x) =\sum_{n=0}^\infty (-1)^n x^{2^n} = \frac12 x + \frac{2}{\log 2}
\Re\sum_{n=0}^\infty\left(
\frac{\Gamma(\alpha_n i)}{\lambda^{\alpha_n i}}
- \sum_{k=0}^\infty \frac{(-1)^k}{k!}\frac{\lambda^k}{\alpha_n i + k}
\right)$$
where $\alpha_n = \frac{(2n+1)\pi}{\log 2}$ and $x = e^{-\lambda}$.
As $x \to 1^{-}$, $S(x) - \frac12 x$ will be dominated by the first term ( the term for $\alpha_0$ ) which oscillate with amplitude 
$$\frac{2}{\log 2}\left|\Gamma\left(\frac{\pi i}{\log 2}\right)\right|
= \frac{2}{\sqrt{\log 2\sinh(\pi^2/\log 2)}} \sim 0.00275
$$
and periodic in $\log_2 \lambda = \frac{\log\log\frac1x}{\log 2}$ with period $2$.
Please look at the paper mentioned above for more details.
A: This is a copy of my answer in mathoverflow which I've linked to in my OP's comment but it might really be overlooked

This is more an extended comment than an answer.
I refer to your function 
$$ f(x) = \sum_{k=0}^\infty (-1)^k x^{2^k} \qquad \qquad 0 \lt x \lt 1 \tag 1$$ and as generalization
$$ f_b(x)= \sum_{k=0}^\infty (-1)^k x^{b^k} \qquad \qquad 1 \lt b \tag 2$$

[update 8'2016] The same values as for the function $g(x)$ as described below one gets seemingly simply by the completing function of $f(x)$ -the sum where the index goes to negative infinity, such that, using Cesarosummation $\mathfrak C$ for the alternating divergent series in $h(x)$ , we have that
  $$h(x) \underset{\mathfrak C}=\sum_{k=1}^\infty (-1)^k x^{2^{-k}} \qquad \qquad    \underset{\text{apparently } }{=} -g(x) \tag{2.1} $$
(Note, the index starts at $1$) .
  The difference-curve $d(x)$ as shown in the pictures of my original answer below occurs then similarly by
   $$d(x)= h(x)+f(x) \underset{\mathfrak C}= \sum_{k=-\infty}^\infty (-1)^k x^{2^k} \tag {2.2}$$
  Using the derivation provided in @Zurab Silagadze's answer I get the alternative description (using the constants $l_2=\log(2)$ and $\rho = \pi i/l_2$ )
  $$ d^*(x) {\underset{ \small
 \begin{smallmatrix}\lambda=\log(1/x) \\
 \tau = \lambda^\rho \end{smallmatrix}}{=}}  \quad {2\over l_2 } \sum_{\begin{matrix}k=0 \\ j=2k+1 \end{matrix}}^\infty    \Re\left[{\Gamma( \rho \cdot j ) \over \tau^j}\right] \tag {2.3}$$
  which agrees numerically very well with the direct computation of $d(x)$.
(I don't have yet a guess about the relevance of this, though, and whether $d(x)$ has also some other, especially closed form, representation)
[end update] 

To get possibly more insight into the nature of the "wobbling" I also looked at the (naive) expansion into double-series.
By writing $u=\ln(x)$ and expansion of the powers of $x$ into exponential-series we get the following table:
$$ \small \begin{array} {r|rr|rrrrrrrrrrrrrr}
 +x & & +\exp(u) &&= +(& 1&+u&+u^2/2!&+u^3/3!&+...&) \\
 -x^2 & & -\exp(2u)& &= -(& 1&+2u&+2^2u^2/2!&+2^3u^3/3!&+...&) \\
 +x^4 & & +\exp(4u)& &= +(& 1&+4u&+4^2u^2/2!&+4^3u^3/3!&+...&) \\
 -x^8 & & -\exp(8u)& &= -(& 1&+8u&+8^2u^2/2!&+8^3u^3/3!&+...&) \\
 \vdots & & \vdots & \\ \hline
 f(x) & & ??? &g(x) &=( & 1/2 & +{1 \over1+2} u &+{1 \over1+2^2} {u^2 \over 2!} &+{1 \over1+2^3} {u^3 \over 3!} & +... &)
 \end{array}$$ 
where $g(x)$ is computed using the closed forms of the alternating geometric series along the columns.
The naive expectation is, that $f(x)=g(x)$ but which is not true. However, $$g(x)=\sum_{k=0}^\infty {(\ln x)^k\over (1+2^k) k! } \tag 3 $$ is still a meaningful construct: it defines somehow a monotonuous increasing "core-function" for the wobbly function $f(x)$ , perhaps so to say a functional "center of gravity".
The difference $$d(x)=f(x)-g(x) \tag 4$$ 
captures then perfectly the oscillatory aspect of $f(x)$. Its most interesting property is perhaps, that it seems to have perfectly constant amplitude ($a \approx 0.00274922$) . The wavelength however is variable and well described by the transformed value $x=1-4^{-y}$ as already stated by others. With this transformation the wavelength approximates $1$ very fast and very well for variable $y$.       
For the generalizations $f_b(x)$ with $b \ne 2$ the amplitude increases with $b$, for instance for $b=4$ we get the amplitude $A \approx 0.068$ and for $b=2^{0.25}$ is $a \approx 3e-12$ 

Here are pictures of the function $f(x)$, $g(x)$ and $d(x)= f(x)-g(x)$ :
 
The blue line and the magenta line are nicely overlaid over the whole range $0<x<1$ and the amplitude of the red error-curve (the $y$-scale is at the right side) seems to be constant.
In the next picture I rescaled also the x-axis logarithmically (I've used the hint from Robert Israel's anwer to apply the exponentials to base $4$):
 
Having the x-axis a logarithmic scale, the curve of the $d(x)$ looks like a perfect sine-wave (with a slight shift in wave-length). If this is true, then because $g(0)=1/2$ the non-vanishing oscillation of $f(x)$, focused in the OP, when $x \to 1$ is obvious because it's just the oscillation of the $d(x)$-curve...
For the more critical range near $y=0.5$ I've a zoomed picture for this:

But from  here my expertise is exhausted and I've no tools to proceed. First thing would be to look at the Fourier-decomposition of $d(x)$ which might be simpler than that of $f(x)$. Possibly we have here something like in the Ramanujan-summation of divergent series where we have to add some integral to complete the divergent sums, but I really don't know.              
