And their own Unix before that. And arguably their own Linux compatible kernel until they dropped that ball.

I don’t want to read vibe code, I want to see a humans code.

You know full well that AI vibe coding is a completely different level of assist to “forums for help”.

Can I suggest you have some rules around AI assisted code? Like explicitly marking them at a minimum?

Otherwise, sounds good, hope it takes off.

Can I introduce you to the concept of “fire” :D

A single bitflip wiping your novel is incredibly unlikely, to the point of being almost impossible. Modern OSs and filesystems are fairly resilient, and the data is likely all still there.

This is just the same “Year of the Linux Desktop” copium regurgitated.

I love using Linux, but there functionally is nothing wrong with Windows either. Much of the world is happily using W11. And Linux has some very rough edges that aren’t going away either, we just learn to live with them.

Some fine point tweezers are usually useful for getting lint out of the USB port, I have done it fairly aggressively and not done any damage yet.

It’s gamifying your exercise. Some get a dog to remind them, some use an app, some just have the habits.

Annoyingly, disk discovery. It refused to use my disks, claiming they didn’t have serial numbers. I could see the serial numbers in the frontend and the console, but their middleware just hated them.

I am using a USB multi-disk drive thing, which didn’t work properly on an old kernel, but it should have been fine with the new kernel.

I reported the bug, which didn’t really get addressed, and then had to build my array using the command line tools (which aren’t documented).

I dunno about recommending FreeNAS (Known as truenas now). It is basically an appliance OS, and unless you are using enterprise level hardware, they want nothing to do with you.

I’m currently using it, but it was a very unpleasant experience setting it up.

You’re not wrong, but the only difference (and time will tell if its sufficient), is that they intentionally and explicitly use the free tier for advertisement, so killing that kills their reputation overnight.

But if all else fails and they try to enshitify, their apps are BSD licenced on github, and headscale is adequate, so the free tier would just shift to headscale and carry on.

Lol, cut me’s some slack’s.

They are called heros. Flawed heros, but thats still a subset of heros.

Let them fight

Thank you [Honorary] Pedantry :D. “ad” just felt wrong somehow, but you are correct. I have fixed it, sorry for the pain and suffering I caused you :D

I’m not, it is normal, even in OSS development. Pushing every ticket is fine, but so is holding back until the work is done or until release. It is, and always has been, up to the project on when and how code goes public.

Its not their software until its released. They aren’t users until its actually released and actually deployed to a phone.

Development being private until release is NORMAL. I do it, every developer does it. All the changes live on my box until I’m ready for the world to see them.

Do you have a source for that? I can believe that select OEMs are getting preview internal access, but I strongly doubt they are releasing with ROMs cooked from those internal branches. That would open the OEMs to GPL requests/violations. And internal access doesnt mean doing the release processes that chew up time.

Giving preferential treatment to certain OEMs is its own issue though, but its a anti-competitive behaviour issue rather than a licencing issue.

Actually, it is open source. There is nothing wrong with developing in private and pushing public when its done. Every developer works that way to some degree or another. And there are good reasons not to push every commit public.

Of course. Although I can kinda see why they dont want to do that either. All the fly-by-night OEMs would be using dev and shipping half-baked ROMs (which I guess they do anyway, but it would be worse).

Cutting a release (and publishing) does cost significant time and effort. You effectively need to code-freeze, get all code merged into main, run all tests and QA, fix any breaking bugs, compete signoffs etc. On some of our small projects, doing a release could burn up to 2 weeks of real time, which on a monthly release cycle was killing us.

So I can almost buy their reasoning. But otherwise, agree that they can’t be trusted, and releasing once a quarter doesnt seem that hard.