I made up the following problem.
A curve $y=f(x)$ passes through $(0,1)$, and for all $a>0$, the arc length of $y=f(x)$ from $x=0$ to $x=a$ equals the $x$-intercept of the tangent to $y=f(x)$ at $x=a$.
That is, in the graph below, the red arc and the red line segment have the same length.
What is the area under the curve, $A=\int_0^\infty f(x)dx$ ?
My attempt
We have
$$\int_0^a \sqrt{1+(f'(x))^2}dx=a-\frac{f(a)}{f'(a)}$$
Differentiating both sides with respect to $a$ gives
$$\sqrt{1+(f'(a))^2}=1-\frac{(f'(a))^2-f(a)f''(a)}{(f'(a))^2}$$
$$(y')^6+(y')^4-(yy'')^2=0$$
But I don't know how to solve this differential equation.
Then I approximated the curve with a sequence of triangles as shown below.
Each of the colored line segments has length $\epsilon$, which approaches $0$.
Let $h_n$ be the height, and let $\theta_n$ be the lower-right angle, of the $n$th triangle from the left. Let $\theta_0=\pi/2$.
$h_n=1-\epsilon\sum\limits_{k=0}^{n-1} \sin{\theta_k}$
$\theta_n = \arctan{\left(\dfrac{h_n}{n\epsilon -\epsilon\sum\limits_{k=1}^{n-1}\cos{\theta_k}}\right)}$
I made an Excel simulation using these two equations, and it suggests that $A=1/3$. (The length of the longest side of the triangles seems to approach $1/2$.)
I tried to rearrange the triangles in some clever way to get a total area of $1/3$, but I haven't found a way.


Hint Regard $x$ as a function of $y$, so that the tangent line to the graph at $(x(b), b)$ is $T(y) = x'(b) (y - b) + x(b)$; then our condition is: $$\int_b^1 \sqrt{1 + x'(y)^2} \,dy = T(0) = x(b) - b x'(b) .$$ Differentiating with respect to $b$, renaming $b$ as $y$, and rearranging gives a more manageable differential equation, in fact a separable equation of first order in $x'(y)$: $$y^2 x''(y)^2 - x'(y)^2 - 1 = 0 .$$
Remark Inverting gives $y$ as a function of $x$ in terms of the Lambert $W$ function: $$y(x) = \sqrt{-W(-e^{-4 x - 1})}.$$