That's hilariously written. Especially the last part when you consider LGR successfully booted MS-DOS 6, bare metal, on a Ryzen CPU before he actually installed Windows.

Then create an array of 100 drives, which would put you at $180,000. Though you could probably get a bulk discount.
Loved that line, would almost justify the expense

I found it very entertaining and liked the style of writing.
Since the year "1984" fell in the link, here's something related. 😉

Nineteen Eighty-Four (1984)
(I know that reference is getting old, really old)
https://en.wikipedia.org/wiki/Nineteen_Eighty-Four

Timeline of DOS Operating Systems (1984)
"Digital Research ships Concurrent CP/M Release 3.1, featuring PC-Mode,
which allows users to run either PC DOS or CP/M-86 applications."

"Microsoft combined versions 2.1 and 2.01 to create MS-DOS 2.11 for other OEMs.
Version 2.11 was sold worldwide and translated into about 10 different languages"

"Microsoft releases MS-DOS 3.0,[K] after a difficult year and a half of
grappling with problems of software incompatibility, remote file management,
and logical device independence at the network level.
In laying the foundation for networking, the core team of five people led by
Zbikowski and Reynolds redesigned and rewrote the DOS kernel.
Redirector and sharer interfaces for IBM's network adapter card were added,
but the redirector itself, which interacts with the transport layer of the network,
wasn't ready.[24] Per Zbikowski: "The product was not ready for us to ship when IBM said,
'Fine, we'll take it.'"[17]"

"The IBM PC/AT, a computer built around the 6-MHz Intel 80286 microprocessor,
with a 16-bit ISA bus, new CMOS clock and 20 MB hard drive, is introduced.
It ships with PC DOS 3.0, which adds support for quadruple, or high density (80-track),
15 sectors per track 1.2 MB (1,228,800 bytes; 2,400 sectors) floppy disks. [..]"

"In Europe, AT&T and Microsoft release MS-DOS 3.1,[L] which added a new local area
network supplement Microsoft Networks 1.0 (identified earlier as MS-Net)[202] for
use on non-IBM network cards. Microsoft Networks services are provided by a file server
which is part of the Networks application and runs on a computer dedicated to the task.
[24] Neither MS-Net, nor its successor LAN Manager, was particularly successful competing
against market leader Novell, whose product Novell NetWare had a seventy percent market share.[44]"

"Accepting the emergence of PC DOS as a de facto standard, Digital Research announced Concurrent PC DOS,
which allows users to run up to four programs simultaneously using PC DOS and/or CP/M."

[..]

Source: https://en.wikipedia.org/wiki/Timeline_of_DOS … erating_systems

Loosely related:

https://en.wikipedia.org/wiki/MiniScribe
https://en.wikipedia.org/wiki/ST506/ST412
https://en.wikipedia.org/wiki/File_Allocation … e#Initial_FAT16
https://www.tiffe.de/Robotron/misc/Winchester-HD/WD1003.TXT

Actually, it's a very good question...

DOS/Windows are known for trying pretty hard to support older software.
So, I wouldn't be surprised if such systems would return 2 GB (or 4 GB) when asked about file size, even if the file is actually larger.

On the other hand, Unix systems avoid lying - if some program can't deal with the truth, it needs to be re-written/re-compiled...

A.1.1.1 "No Lies" Rule An error will be returned whenever a function cannot return the correct result of an operation. […]
Show full quote

A.1.1.1 "No Lies" Rule

An error will be returned whenever a function cannot return the correct
result of an operation.

Returning a "lie" to allow for common uses of a function (e.g. use of
stat() to determine if a file exists) could inadvertently cause a
correctly written application to operate incorrectly.

It is conceivable that returning a "lie" could keep an incorrectly
written application from malfunctioning in a way that creates a serious
problem, but no such applications are known to exist. (Of course it
would be easy to contrive one.)

PASC Interpretation reference 1003.1-90 #38 completed by the POSIX.1
interpretations committee confirms that POSIX.1 conforming
implementations are not allowed to lie to applications. This
interpretation explicitly states that if the file size will not fit in
an object of type off_t, fstat() must fail. In addition, PASC
Interpretation reference 1003.1-90 #75 went on to clarify that
EOVERFLOW would be a legal extension to report this condition.