You have, but general adoption of Linux-lineage OS's is focused in the mobile tech space. Desktop adoption is still exceedingly low.
Statscounter shows Linux installs about on par with ChromeOS installs [https://gs.statcounter.com/os-market-share/desktop/worldwide]. Both are beaten out by the "Unknown" category; there are literally more people running we-can't-tell-what OS than a Linux-derivative OS.
I think these days, the roadblock isn't reliability; it's Linux being off-mainstream. Network effect, essentially. Most of the (non-Internet) stuff the average person hears about in tech media and by word-of-mouth is not guaranteed to be available on Linux and is basically always guaranteed to be available on Windows, MacOSX, or both. Linux archs have done a decent job of solving the "Can I see my screen with the latest graphics card" problem but are still behind the curve on the "Can I buy Photoshop off the shelf at Best Buy and run it on my computer" curve.
. . . are still behind the curve on the "Can I buy Photoshop off the shelf at Best Buy and run it on my computer" curve.
Or I can use GIMP for free.
You are correct that Linux on the desktop is not mainstream and that is largely down to network effects, but I don't understand the next logical step in the argument. How does that make Linux on the desktop "not ready" or "incomplete?" In my opinion (and I recognise that this is just an opinion), Linux has been a superior experience to Windows for a long time. All common problems have several robust solutions. Uncommon problems often also have a good solution (which isn't guaranteed on Windows either, at least not without spending a lot of money). The idea that Linux isn't "Desktop Ready" is a tired trope. Just because people don't, doesn't mean they can't. Linux doesn't have a marketing department (and I wouldn't want it to).
But again, that's the problem to solve. You can use GIMP for free; most people can't. It's a space-alien UI relative to the Photoshop they know. And while I love GIMP and think it's a fine drop-in replacement for 90% of use-cases, it's still not good enough to fly (not fast enough, not robust enough, not the same size of plugin ecosystem) for the 10% that are using Photoshop in high-volume professional work settings. Businesses love cheap, and if GIMP could be substituted, they'd have already forced substitution on their art-houses.
You can even probably get Photoshop running on your box with the right cocktail of emulators and libraries. Most people can't.
The problem of picking up software from any brick-and-mortar store, taking it home, and running it on my Linux machine has no good solutions. Alternatives that you and I know work just aren't palatable for the average user. The average user still can't.
I'm not sure this is, practically, a solvable problem; I'm just identifying that it is the problem for adoption of Linux on desktop outside of the wonks like us that are willing to learn a lot of computer stuff off the beaten path.
Modern desktop Linux is great, but it's a whole commerce-adoption model away from being part of the beaten path.
Statscounter shows Linux installs about on par with ChromeOS installs [https://gs.statcounter.com/os-market-share/desktop/worldwide]. Both are beaten out by the "Unknown" category; there are literally more people running we-can't-tell-what OS than a Linux-derivative OS.
I think these days, the roadblock isn't reliability; it's Linux being off-mainstream. Network effect, essentially. Most of the (non-Internet) stuff the average person hears about in tech media and by word-of-mouth is not guaranteed to be available on Linux and is basically always guaranteed to be available on Windows, MacOSX, or both. Linux archs have done a decent job of solving the "Can I see my screen with the latest graphics card" problem but are still behind the curve on the "Can I buy Photoshop off the shelf at Best Buy and run it on my computer" curve.