There's absolutely no need for a full-quote of what I wrote. Please quote only what's specifically relevant for context. Full-quotes are annoying and commonly an exhibit of mindless lazyness.
It doesn't matter what all you have saved in that .txt file on your Windows desktop when you weren't able to prevent the file to be deleted. Of course, shit happens, occasionally I deleted files that I didn't want to delete, but that's what backups are for.
When was the file deleted? Recently or long time ago (the latter would be bad)?
I think the more expensive versions of Windows 7 had "Previous Versions" feature, where you could restore deleted files from the "Previous Versions" of the folder the file was located in. Have you tried to see if your Desktop folder has "Previous Versions" available (from Shadow Copies / System Restore points)?
Moreover, somewhere on the forum I read that people used GPUs to crack private keys that consist of numbers and letters.
Bitcoin private keys are 256-bit numbers. How you encode them to make them shorter is another story and in the cracking context only relevant if you're missing rather small portions of them, say in WIF format.
People were able to crack low-entropy private keys or when very bad PRNGs were used to create "bad" (less random) private keys (which would again be in the category of low-entropy private keys).
Provide sources of your claim when someone managed to crack a "good" Bitcoin private key!
I think that if they were capable of that, then why is an ordinary 48-digit numeric code considered difficult and unrealistic? Considering that Google has already invented the Willow chip, does solving it remain realistic in the next five to twenty years?
They weren't, do the math. You can't tackle a search space of magnitude 10
48, neither by time, nor by amount of energy needed, nor by expenses if you want to rent the required number of GPUs to be able to find a solution within your lifetime (you won't).
If we translate the problem to find a specific 48-digit decimal number to a binary number problem, we're dealing to find a 160-bit binary number. No. of bits is
log2(10
48) = 48*
log2(10) which is approx. 48*3.322, almost 160.
That could be equivalent to the puzzle #160 of the famous Bitcoin challenge puzzle (you can find the mega-thread to it in the Bitcoin Discussion board; search for yourself, I don't see the need to provide you a link, you might benefit from a little searching exercises).
Attempting solvers of the Bitcoin challenge puzzle struggle hard (for months) to solve puzzle #71 (with unknown public key) and that's a key search space in the order of 2
70, that's 1,237,940,039,285,380,274,899,124,224 easier than a 2
160 or 10
48 problem.
You have no idea of the magnitude of such numbers. Assume you have a rig that can try a billion or a trillion of numbers. Now, do the math how long it would take to search a specific number in the 10
48 space. DO the math...
