I share his vision. For this exact same reason I get angry, when people start rambling about software engineers being needed no more. It’s exactly the same abstraction layer extension, as compilers were, albeit the one that grants even less control over execution flow.

vim, awesomewm, mbsync (isync),

As for recent discoveries: dwl — I was surprised on how robust it is, and how well it works.

I’ve been running ministral on CPU on a home-server: works pretty nicely, not very performant for everyday tasks and the savings were not sufficient for it to make sense. It still was cheaper and faster to just use Mistral API and get better models.

IMO, the only way the switch will actually happen is not the RAM or GPU, but if more manufacturers will start selling new/refurbished devices with Linux on it. Reason being: most people treat computers exactly the same way as I treat my car. I won’t bother replacing firmware, engine, transmission, or even tires on my car if there is no good reason for it. Same goes for people: they won’t be replacing OS, CPU, RAM, disk, etc, except there will be a really good reason for it.

Though, one thing to be mentioned: I generally see the trend that Linux is becoming more popular as more and more popular people adopt it (and are vocal that they have better time with it than with the alternative): PewDiePie, Linus Sebastian, etc. I think the trend might potentially accelerate, as more and more people are really unhappy with how well Windows works these days, and not everyone is ready to buy Mac (though the most recent Neo release success is a great indicator of how bad situation on the market is overall)

There are folks form KDE who are trying to implement the Plasma Bigscreen solution: https://plasma-bigscreen.org/. Seems promising to me :)

Though, I’d still recommend to use an external device to avoid breaking the TV OS up ;)

The web ui with integration of email ecosystem for all those things are one of core selling points of https://sr.ht/

I’ve been following the software forge federation some time ago, and didn’t feel to pick up even when it was discusssed initially. It is a neat idea on high-level, though it requires forges to implement it, which has a risk of not picking up (just look at how much iterations of social media federation protocols was there, until ActivityPub arose).

On the other hand, all of the forges are based on a distributed technology out of the box: git. Most of the “modern days” comforts there are, are just built on top, and there are different ways to approach it.

As an example, you can send patches directly to the author in email. Is heavily implemented and suggested by https://sr.ht/ (1) — a software forge, which focuses on building a federated workflow by using email for communication (which is federated by design). This way, you can create “Pull Requests” without having account on the forge — all you need to do is just submit a patch. Author is very vocal about supporting it (2), and provides quite useful guides to learn (3), (4)

Generally, I’d say that e-mail is the only federative implementation you can get so far :)

But on macOS it just uses Apple’s own WebKit fork, so it is very expected: WebKit is very optimised towards Apple hardware on macOS and iOS.

Each time I see anything like that, I just disengage with the content

I was going a long way, until I built a perfect AwesomeWM configuration for myself, and have not changed it for a while now. I am willing to switch to Wayland-based solution now, as it seems to be a bit more performant, but I just can’t make myself to do it: my config is really cozy and working

I like systemd overall. The ease of use, uniform interface and nice documentation is awesome.

Though each time I try to run it on outdated hardware (say, my Thinkpad X100e, which is, well, a life choice xD) — it makes whole system much slower. IMO, openrc is not as bad, and in some ways it gives some capabiilties of systemd these days.

Wow, thanks for suggestion of Tdarr — that project indeed looks very nice. What is. your experience using it? Any quirks?

My understanding of Keybase is that it was some kind identity aggregator. You were able to link identities not just by keys, but also by external services, like Twitter (at a time), email and other things.

I’d prefer not to dual boo, but it might be the safest way to start? If I dual boot, get used to Linux and (hopefully) get everything I need working, can I then go from dual boot to erasing the Windows partition and recombining so I then only have Linux installed and can keep the work and programs I already installed on Linux?

My personal experience says: try dualbooting first, because it will make you to have a working machine continuously. Taking into account that all Linux-based OS behave vastly differently from MS Windows, it is possible to break things, when learning a new way of doing things.

The drives for my server are NTFS. Does anyone have experience with this format on Linux (I use Emby)?

I’ve been using an external NTFS drive for compatibility and big files storage: works as charm. The worst case scenario is you will need to install an ntfs-3g driver, although it is usually included with the distro.

As for production: I don’t have much experience with that, although I can recommend you looking around tooling that solves the problem. You will need quite a bit of patience and trying things, because switching platform will definitely require you to make some shifts in usual processes you have now. Don’t expect things to be obvious 100% replacement: unfortunately lots of people have this expectation, and get frustrated.

As for hardware, just looking the model up on the internet with adding “linux”, or “ubuntu”, or “fedora” should do the trick of figuring out if it will work.

Exactly my feeling each time I get back on personal PC/laptop after whole day of working with Mac.