If that’s the reason for things freezing up, the OOM reaper will come along and sort it out eventually, although it isn’t exactly quick on its feet.

Nah, there’s plenty of variety. Keep in mind that not all distros ship with any graphical environment at all, and you’re not obliged to keep your distro’s default even if it did ship with one. Linux in general is just as happy with ratpoison as it is with KDE.

I doubt most of them could stick with the Gentoo installation procedure for long enough to make it to a usable system.

You can still compile a surprising number of modern programs and libraries without unicode support (that is, they provide an explicit compile flag to switch it off)—it’s just that no general-purpose distro does it by default. I’m not sure you can set up an entire unicodeless system using current software versions, but I wouldn’t bet against it, either. And glibc isn’t the only game in town—musl is viable and modern (it’s the default libc in Alpine Linux and an option for some other distros), and designed for resource-constrained environments. Those two things between them might bring down the size by considerable.

Thing is, most LLM submissions are low-quality as well as low-effort. If you forbid them, well-meaning numbskulls will hopefully not clutter your bug tracker by submitting them, and those who are more interested in adding a line to their resume than following the rules can be blacklisted immediately for breaking said rules. As for the odd undeclared one that’s not low-quality and slips through without being spotted, no big deal. By my understanding, they’re unicorns, though.

Because the submissions are so low-quality overall, chances are that projects requiring that submitters admit there was an LLM involved in their submission will end up effectively shadow-banning most such submitters because it isn’t worth wading through their tripe. That’s just a different version of non-transparency.

The endgame we want isn’t blacklisting LLM submissions into perdition, it’s the code version of xkcd 810. Currently, most LLM code submissions are about as useful and desirable as porn spam on a forum. Maybe in a few years, that’ll be different. If it is, policies can be reviewed.

Actively offensive branding might turn me off, but I have a high tolerance for low-quality artwork and such, or I wouldn’t be able to use FOSS at all. Honestly, most programmer-generated branding has never been very good—at best, you might get a witty project name, but it’s often accompanied by a cheesy tagline and a cruddy icon that doesn’t render well at small sizes and so is useless as an icon.

Seems awfully elaborate for something I can accomplish by backing up a few text files and directories of same. The only real issue I had with spinning up the new laptop this spring was due to an underdocumented and partially broken UEFI BIOS that I had to figure out how to work around. Once I got past that, transferring other packages and settings was trivial.

Nothing. The sole Windows key on my keyboard has never been intentionally pressed except by my cat.

In my case, part of it is that sudo is an extra installation for me on Gentoo, while su is part of the base system on any Linux. Given that all nontrivial software has bugs, every unneeded package you install adds very slightly to your security risk.

In terms of security, sudo is better in the environment for which it is intended: a system with multiple human users that has a dedicated sysadmin who curates /etc/sudoers and makes sure that no user has more permissions than they absolutely need. However, only a small fraction of all machines running Linux meet those criteria. On the typical home system that’s using some distro’s default sudo-with-user-passwords setup, you can get root authority with only one password, whereas with su you need the passwords for both a wheel account and the root account. That isn’t much added security, but every little bit helps. On the other hand, sudo can be set to require you to enter your password again after a period of time, while su will allow a root session to hang on unto infinity, which may matter if untrusted Linux-savvy people have physical access to your machine (I don’t have that issue).

In other words, the benefits are real but minor and situational.

(None of this holds if you’ve done something really stupid in your configuration, like always starting an SSH server that allows both password login and direct root login when the system comes up. Always follow current best practice—in this case, certificate login only, and no direct root login—when setting up something that can be accessed over the network.)

(Some people claim that sudo has stopped them from unintentionally running a command as root. I just assume any console I’m using has root privileges and I shouldn’t run dodgy commands in it to begin with.)

All non-trivial software has bugs, and it’s unsurprising that in a sudo implementation in any language, many of those bugs are security-related. This is still quite young software. Ubuntu was premature in making it their default, I think, but that just means it’s immature, not that it’s completely broken.

Then again, I use su exclusively and don’t even have sudo installed, so I’ve got no dog in this fight.

(As for Rust itself, I am neither for nor against. It’s a programming language. It has some issues that mostly seem to be related to how building and distribution is carried out in practice, rather than the core language design. I have never met a programming language without warts, and I’ve used several. If you’re experienced with the language—whatever it is—you learn how to handle them.)

TDE will run quite happily on pretty pathetic hardware (that laptop with 2GB RAM that I retired this year ran TDE). It can’t do anything about Firefox being a pig, though.

I wonder if Gentoo is going to become the last option standing for some of these obsolete arches. Of course, it may only be matter of time before Portage won’t work on wd40 arches either, given the increased uptake of Rust in Python packages.

The problem here isn’t that Framework failed to keep track of the ideology espoused by every major developer on the projects they contributed to or endorsed, which, to be honest, isn’t something I’d expect of them. The problem (as usual for a corporation) is how they handled complaints. Trying to sweep stuff under the rug in the Internet age just results in someone setting the rug on fire. If instead, Framework’s response had been “We’re sorry, we didn’t know, we won’t give money or free advertising to any projects this guy is involved with from now on,” the whole mess would have died down by now except for a few people grumbling about how they should do more research before sending money out.

If you’re talking about the issue I’m thinking of, it wasn’t really specific to ASUS, but could be triggered by overvolting an X3D CPU on any board. Most of the big manufacturers (ASUS included) have revised their mobo firmware to make that impossible now.

You actually can boot without one—the system I’m typing this on right now has none. However, you have to keep the system simple enough that the root partition can be mounted without requiring anything not built into the kernel. That means your root partition can’t be encrypted or RAIDed, and you can’t build the support for the filesystem it uses as a module, which can mean a custom kernel. Most people don’t want to bother.

The larger ones were flexible, not floppy—they could be bent without cracking the casing, but wouldn’t just bend under their own weight.

Nah, I’d end up explaining why floppy discs weren’t floppy, instead, and let the younger folks explain the CDs.

Nah, that’s APL.

I think it was supposedly New Zealand or something. It’s been a long time since I’ve read the full texts.

That series of RFCs (1149, 2549, 6214) keeps getting rediscovered by new generations of technical folk. Among other issues that have never been completely addressed are accidental encapsulation of packets in hawks, and whether the Internet is doomed to be slow in locations where the only avian carriers available are flightless.

There has been one successful implementation of the protocol to date. 55% of ping attempts went through.

(As April Fools RFCs go, the only one that’s arguably more popular than IPoAC is the Hyper Text Coffee Pot Control Protocol, the source of 418 I am a teapot).