I was just thinking "OCR" when writing that post, too. But yours is a straightforward solution, that I have not heard of before.
To make sure that YC doesn't get into stupid IP trouble, I took a random snapshot of a public book that I was reading on my PC kindle. Here is the result from google docs:
-------
[...] We neighbors are out of bed." This struck the eeuld net then empley him; but I foolishly
rest, and we s00n after had offers from one let him kngw as 3 Secret that I 50011
Of them t0 Suppl)’ U5 with 5tE1ti0I1e1"Y; but intended to begin a newspaper, and might as yet we did not chuse to engage in shop then have work him_ My hopes of bL1SiI16SS. success, as I told him, were founded on mention this industry the more this, that the then only newspaper, printed particularly and the more freely, tho’ it by Bradford, was a paltry seems to be talking in my own praise, that wretehedly manag'd, no way entertaining, those of my posterity, who shall read it, and yet was profitable to him;
----
So obviously there are some bugs. The whole thing took me like 1-3 min for the first time. Google docs is kind of slow on my system. No I would still have to correct all the mistakes. No idea of the GP's approach is more work. Still a nice idea.
Not sure which text size you're using when you take your screenshot, but it works 100% for me using Jane Austin's Pride & Prejudice. The only errors are where I clip the descenders and is much faster than typing:
I too had to type things from my kindle to send quotes to friends. Though, never that long. Thanks for the effort!