Agreed. There are many facets to this problem, so it’s difficult to get in one post, so I’ll try to reconcile the main points.

The core of what I’m trying to say, is don’t kill Linux trying to become Windows. Linux is great because it diverse, but it also has difficulties because of this. We should not change (nor destroy) the ecosystem for people who do not care to understand it.

That being said, we can also make it easier for people who do care and cooperate to make it over. But if we do this we, as Linux users, have to look at this from the right lens. The question is not “Linux users, what do you find difficult?”; this is survivorship bias. The question is “Windows users, why can’t you get Linux on your machine?”. From this framing, the real issues become a lot more apparent:

• Not savvy enough to set up USB stick
• Driver, and other hardware, issues
• Programs needed for work, or general daily usage, are unavailable
• Too much tinkering required (this is related to, but not the same as RTFM and CLI)

The first two points can be solved by purchasing a machine from a Linux OEM (i.e System76). If this is not possible, then you are going to have to do research; if this burden is too heavy, Linux is not for you.

AI has a good and valid use-case here, as it can significantly ease this process (even if it’s only right 60% of the time).

Linux may not have an alternative for your preferred programs; if this burden is too heavy, Linux is not for you.

Developers should follow open guidelines (i.e POSIX). If they refuse to, there is nothing Linux can (nor should) do about it.

The last point can be solved by distro choice, we completely agree here. The problem is finding said distro, which is difficult. For example, I’ve never heard of Ublue until your post. I appreciate distros that handle defaults and don’t push breaking changes. The community can make this better by having a dedicated website (with a decision tree) for choosing a distro, but this has its own set of issues.

No matter, the responsibility falls on the user to pick the right distro; if this burden is too heavy, Linux is not for you.

Linux Mint is a great distro, and I’m happy it works for you.

In terms of mass-adoption though, the fatal point is probably putting a Linux ISO on a thumb drive. Like I said prior, we must be aware of survivorship bias. You don’t care much for the terminal - but you made it through.

The people that didn’t make it through probably failed from the thumb drive step. I only say this from personal experience, because when I first installed Linux, I was very determined and came extremely close to giving up at this step. And I only got through because I happened to find an obscure forum about how Rufus needed a special setting for my machine.

P.S. I also was not tech savvy, but I wasn’t completely lost either - and I still struggled really hard here.

This is actually a really deep rabbit-hole. To avoid typing a novel, I’m going to cut out a lot of nuance.

Windows is installed by default on machines. Since people do not change defaults (many studies have been done on this), this is checkmate. As long as this is true, Linux will not have a major (20%+) market share.

So this has to start from the OEM. Several Linux OEMs exist (i.e Tuxedo Computers, System76, Framework) but they cannot compete with the Microsoft network. Those who are interested in Linux, but are not tech savvy, really really really should buy their device from a Linux OEM.

Driver issues are near non-existent on Linux OEM hardware. So software is the next step; and let me tell you, developing for Linux is rough. There are 2 window servers, 2 graphic stacks, 2 desktop environments, 2 coding standards, 2 C libraries,… you get the point. A lot of this can be abstracted, but it takes genuine work to do - and may be obsoleted in a month; meaning no company will do this.

All to say, creating “magically working” apps - even with a lot of monetary support - is a herculean task. Even Valve (who is FLUSHED with cash) gave up and just decided to make their own distro (SteamOS).

A lot of issues also just require personal tweaks due to open-source software being extraordinarily bad at setting sane defaults. With something like Windows, you can hire people to make this better. Who do you hire to fix the defaults for 300 independent projects? And will the devs even listen to them?

I could keep going, but you get the point, the buck is going to have to stop at the user for a lot of things.

The best solution (in my opinion) is to have specialized distros and have people choose from them. Want to game? SteamOS. Want to dev? Fedora. Want to surf the web? Linux Mint. Creating, and more importantly accurately listing, specialized distros will make lives easier. Leave the defaults to the devs, just download the “vibe” you want.

Windows users and Linux users are not seeking the same thing from their machines. The common mistake I often see from Linux advocates.

From personal experience, when I was a Windows user, I didn’t care (or even know) about privacy, open-source software, nor owning my machine. I didn’t care if I had to sign up for a Microsoft account, and I never changed defaults ever (except for my wallpaper). I just wanted the machine to turn-on, work, and play some games.

Why am I bringing this up? Because Linux requires the user care about their machine and defaults. You need to know your architecture, graphics card, and threat-model. You need to know what your apps are called and where they come from. You need to know what tools you need to troubleshoot (and devs will not help you). This is the biggest the pain-point of Linux. Do not succumb to the survivorship bias of RTFM or command-line.

This issue cannot be fixed from simplifying Linux interfaces (though we should do this anyway!). The soul of Linux is adventure, collaboration, and tinkering. To get the most from your machine, you’re going to have to interact with several communities. This is what makes Linux great, and frankly I do not think we should kill this for the general public - this is how you get enshittification.

The general public needs to understand that incompetence (being brain-dead) will lead to misery. It is simply the rule of the land. You need to care and you need to collaborate. We should not welcome nor accommodate users that refuse to do this.

I’ll leave the privacy policy here and let people decide for themselves.

They keep two types of logs. An identifiable one which is deleted in 24-48 hours (dns0 and quad9 also do this) and an anonymized one. There is no mention of “business partners”; and it also says explicitly that the information is not used to target ads.

As the privacy policy and service reads, it is not a honeypot. However, Google generally does not act in good faith, so there’s no telling if they have 100% adhered to the policy.

No matter, to make calculated and informed decisions, we should have all the facts in order.

Wouldn’t you just use AFS, CEPH, NFS, or 9p?

I really don’t want to be that guy, but isn’t SSHFS (FUSE) actually a terrible option when compared to an actual file-system? MacOS isn’t really missing out on much there.

The most painful part of MacOS (which makes it downright unbearable for me) is that system configuration files are XML. It’s an absolute nightmare.

Like I said prior, there is nuance to be had here.

We agree that Google products are generally a honeypot (good products that lure you in), but which products are honeypots are important.

You very likely want to avoid Chrome, Gemini, and Google Search - but 8.8.8.8 is not a honeypot, it is a loss-leader. You will be lured in from 8.8.8.8 if you say “huh. this is a great service. is there anymore?”, but 8.8.8.8 itself is not a malignant service.

Google does not automatically mean bad. It is dangerous precedent to blanket ban and remove nuance.

8.8.8.8 is an excellent service, and provides genuine privacy gains. The largest downside being that it is such a massive target for bad-faith and ignorant actors - like the Italian government.

Tor itself has a pretty good routing scheme that seems like it could replace DNS entirely. There are obvious (but surmountable) UX issues and there may be scalability issues - but it is 100% worth investigating.