In all seriousness, I have a high respect for Unix and Unix-like systems, particularly FOSS implementations like Linux and FreeBSD. When I first started using Linux in 2004 as a high schooler who grew up on Windows and who used classic Macs in elementary school, the power of the Unix command line and the vast selection of professional-grade software that was available for free, and legally with the source code, too, was mind-blowing. Not too long after that, I started learning about the history of Unix and about its design philosophy. I had a new dream: I wanted to become a computer scientist just like Ken Thompson or Dennis Ritchie working for a place like Bell Labs, or become a professor and lead research projects like the BSDs back when Berkeley had the CSRG. Downloading and using Linux in 11th grade pushed me away from my alternative thoughts of majoring in linguistics, mathematics, or city planning.

Sometime in graduate school I started paying more attention to the Xerox PARC legacy of the Mac, and I started realizing that perhaps Unix was not the pinnacle of operating systems design. I got bitten by the Lisp and Smalltalk bugs, and later I got bitten by the statically-typed functional programming bug (think Haskell and ML).

These days my conception of a dream operating system is basically an alternative 1990s universe where somehow the grander visions of some Apple researchers and engineers working on Lisp (e.g., Newton’s planned Lisp OS, Dylan, SK8) came to fruition, complete with Apple’s then-devotion to usability. The classic Macintosh interface and UI guidelines combined with a Lisp substrate that enabled composition and malleability would have knocked the socks off of NeXTstep, and I say this as a big fan of NeXT and Jobs-era Mac OS X. This is my dream OS for workstation-class personal computing.

With that said, I have immense respect for Unix, as well as its official successor, Plan 9.

IBM: everything is a record type, we have 320 different modules to help you deal with all 320 system-standard record types, and we can even convert some of them to others. And we have 50,000 unfixable bugs because other pieces of code depend upon the bug working the way it does ...

UNIX: everything is an ASCII byte. Done

I started writing code in the 1970s on TOPS-10, TWENEX, PLATO, and my undergrad thesis advisor helped to invent Multics. The benefits of UNIX are real, folks, the snake oil is the hardware vendors who HATE how it levels the playing field ...

This exact version of it was first published in v. 4.0.0, on 24 Jul 1996: http://www.catb.org/jargon/oldversions/jarg400.txt

That was then published as The New Hacker's Dictionary, third edition, 1996: https://books.google.com.vc/books?id=g80P_4v4QbIC&printsec=f...

[1] https://en.wikipedia.org/wiki/Jargon_File

To be fair, it’s not a stretch to suspect a company wants its competitors to be dependent on their product. The theory of planned security vulnerabilities sounds off the plot however.

That seems more applicable to Windows these days. If you graph CVEs vs. version, there is an interesting trend.

I remember watching a YouTube UNIX wars documentary that posits the exact opposite.

It argued that top brass at AT&T saw UNIX as a means to the telecommunications end, not a business in its own right. When it became more popular in the 1980s, it became obvious that they'd be bad businessmen if they didn't at least make a feeble attempt at making money off of it, so some exec at Ma Bell decided to charge thousands (if not tens of thousands; I can't find a reliable primary source online with a cursory search) per license to keep the computer business from getting to be too much of a distraction from AT&T's telecoms business.

That limited it to the only places that were doing serious OS research at the time: universities. Then some nerd at a Finnish university decided to make a kernel that fit into the GNU ecosystem, and the rest is history.

(This conspiracy may not be factually true, but it is teleologically true)