I agree with your overall opinion, but I just don’t agree with how the problem was presented. Your statement, with more of the surrounding context:

… lemmy.ml, works more like that than you realize. e.g. a change is soon going to give lemmy.ml veto power in what communities are allowed to be acknowledged as existing to new instances …

The key words here are “allowed to be acknowledged as existing”. Not acknowledging a community’s existence means not federating it. .world does that with db0’s piracy community because of EU laws, and it’s basically an instance-imposed community ban. Pyfed has/had a hard-coded denylist of community names in the source code that stopped them from being federated, and the result was none of the instances running unmodified Piefed were able to access them.

I wouldn’t have an issue with if you said a change in Lemmy “gives lemmy.ml exclusive control over promoting what communities show up as popular in other instances”. They don’t have the ability to censor the existence of communities that go against their views just the ability to censor their promotion. That’s a big problem, but it’s not as catastrophically bad as them having the power to censor the actual content on other instances.

I dislike centralization as much as the next person and have my issues with lemmy.ml being allowed to control anything outside its own instance, but I think the way you phrased it is misleading.

what communities are allowed to be acknowledged as existing to new instances

That suggests .ml has the ability to prevent communities from being acknowledged at all by other instances, while the anti-feature is actually about them being the sole source of truth for what counts as a “popular” community.

They can censor and curate that list to their authoritarian-apologist desires—which is a problem—but it only affects discoverability when browsing for popular communities, and instance admins can (and should) turn that off.

Your source is 3 months old and doesn’t back up your claims.

what does “hardcode lemmy.ml as a source to pre-fetch popular communities” mean in practice.

It is an attempt to pre-populate new instances with some popular communities which is seen as a way to improve discoverability. I find the general concept of using “popularity” for that to be somewhat problematic, but the main issue I have with the actual implementation is that it uses lemmy.ml as the source of truth for that, and there is no way to change that*.

Step 1: Make personal computing unaffordable.
Step 2: Rent “personal” computing as a service.
Step 3: Boil the frog by continuously restricting what people can do with the service.
Step 4: Wait for local computing to die.
Step 5: Stop LLMs from running on rented computers.

Hardware won’t last forever. Once they have full control over what people can do with computers, they have full control over the information people are exposed to. LLMs won’t help with that if the only ones that are accessible to the layman are sanitized and censored.

Oh boy. We’re already at 13 out of 14 on the checklist.

Sign a 5-year contract for one baby a month, outsource the first 8 deliverables at a loss while using the time to ramp up production staggered such that it produces a baby on a monthly basis. /s

Any distro works.

Any non-LTS distro works*

Using a distro release based on a 2 year old kernel with brand new hardware is asking for a horrible experience. For gaming especially, you’re also losing out on months/years of improvements to Mesa.

Mac is very similar to Linux in that it comes with bash (these days zsh) and a lot of the command line tools you’d expect on Linux, including gcc

No it doesn’t.

The gcc command is a wrapper for clang, and the clang command is a stub that runs an executable used to install the “Xcode Command-Line Tools”

It also uses the BSD coreutils, rather than the GNU coreutils present on most Linux distros. The two are only compatible up to functionality defined by the POSIX standard, and anything beyond that is an inconsistent mess.

Windows is more difficult. The command line is very different (it inherits from DOS instead of Unix like both Mac and Linux). It doesn’t come with Python pre-installed

If you limit yourself to not using WSL, sure. WSL 2 runs an actual Linux kernel with the same Linux executables you would find on any other distro.

It’s still Windows and full of telemetry and AI garbage nobody wants, but it somehow manages to have better Linux compatibility than macOS.

Even Musk, for all his recent evil got rich trying to reduce our dependence on gas cars.

Everybody else already covered his role in Tesla, so let’s look at something else that demonstrates his concern for the environment and his fellow species:

He has a datacenter in Memphis running 35 “temporary” methane generators to power Grok, the self-described “Mecha Hitler” AI. All but a dozen of them are being used without permits for permanent generators, and none of them have air pollution filtration systems installed. Oh, and it’s near a low-income community that was already plagued by air pollution.

Agreed. The call trace shows it occurred as part of a drm_ function, which is related to the DRM (Direct Rendering Manager) subsystem.

There’s a chance it might not be the root cause, but the more obvious answer is that the Nvidia driver managed to corrupt a kernel data structure.

A thief doesn’t loudly announce that they just stole something.

Yeah, they’re are. I used sbctl to enroll and manage my own keys, and I chose to include the MS ones to ensure dual booting still worked properly.

Because of that hard-bricking motherboard problem, choosing to not include the MS keys is actually more effort due it being gated behind a flag and a mountain of warnings.

these games only accept the secure boot setup where the root key is that of microsoft’s.

I have a PC where I could actually test this. Custom MOK but with all the MS signatures in the database. I can boot into Windows through the BIOS using only the MS-signed bootloader instead of GRUB or any chain loader, and Windows itself considers Secure Boot to be enabled successfully.

Do you know if it would immediately reject the game from launching, or would I be flagged and banned later as some kind of ban wave?

The latter is something I would prefer to avoid.

Oh, the comment I directly replied to is absolutely justified in its downvotes. I actually meant to reply to a different comment of theirs.

There’s a lot of FUD and disinformation around Secure Boot and TPM 2.0 in general. When it comes to anticheat and those as requirements, people are dog piling IMO. The comments being cynical or exaggerating the security risk of the TPM to the user are getting more upvotes, while the comments that disagree with those are getting pushback.

secure boot enabled with machine owner keys wouldn’t be enough either for these games

They should be able to check which signing keys were used for every part of the boot process. Unless they want to be colossal assholes and check the MOK as well, they could still verify what they need without flagging Linux Secure Boot dual-booters as cheaters.

You quoted the end of my comment, so you must have read this part:

Together, they make it possible for anticheat to tell if something (like cheating software) tried to rootkit Windows as a way to evade detection.

For the threat model of anticheat software, verifying system integrity is not an unusual requirement.

The other reason is that I’m kinda verbose.

Some say, “why many words when few work?”

Others express their freedom to choose exactly to what extent of verbosity and verbiage they consider necessary in order to accurately and effectively communicate their previously-unspoken thoughts either through private correspondence or statements to some subset of the general populace.

Amazing. Crowdstrike did end up providing some benefit to users after all.

Once you delve into the technical specifics of Secure Boot and the TPM, it’s actually not that unusual. I wrote more detail in another comment on this post, but the TLDR of it is that Secure Boot is meant to enforce the integrity of the boot procedure to ensure that only approved code runs before the Windows kernel gets control, and the TPM 2.0 is meant to attest to that. Together, they make it possible for anticheat to tell if something (like cheating software) tried to rootkit Windows as a way to evade detection.

I don’t agree with the requirement, but it’s not a pointless requirement or some grand conspiracy to make people buy new hardware.

Sorry to see the downvotes on your comments explaining the technical stuff. You aren’t wrong, but people are cultish and like dog piling.

The entire idea of Secure Boot is to verify the boot chain using signature checks to ensure that nothing “unauthorized” runs in the boot process before control is handed off to the kernel. It’s meant to stop lower bootloader stages from silently modifying or hooking later stages.

In theory, it’s supposed to stop rootkits from being able to exist above the OS, hiding themselves while stealing information or influencing programs. In practice, there’s a shit load of badly implemented EFI programs and bootloaders that are signed and later turned out to be vectors for arbitrary code execution (this is why you need the DBX list to be updated frequently).

Cynically, Microsoft probably came up with Secure Boot because that whole rootkit-and-fuck-with-the-kernel thing used to be one of the ways people cracked Windows 7.

As for TPM 2.0, the whole point of it being used for anticheat is because it stores an immutable log of the Secure Boot process and attests to the integrity of the system. If I installed my own Secure Boot certificates and rootkitted Windows for the sole purpose of cheating, the TPM would see that a self-signed executable was used during boot and refuse to say the system was unmodified.

Edit: The downvote button is not a “I disagree” button. There is an actual technical reason why Secure Boot and TPM 2.0 are used in anticheat crap. I don’t agree with it or that they demand it as a requirement to even open the game, but it’s not some grand conspiracy to make you buy new PC hardware.