chme

I assumed that they meant that they will not enforce it via technical means.

    #define EPROCRASTINATE 245 /* exhausted all output tokens with reasoning */

I'm just wondering why this isn't a European Citizen Initiative (ECI)...

I could not find any information on what kind of influence a online-petition on wemove.eu would have...

Threema is still vendor lock-in.

Element X is in some cases still a downgrade from Element. For instance there doesn't seem to be a way to create local key backups anymore. Also, that calls between Element and Element X are incompatible means both apps need to be installed in order to receive calls from all contacts.

Still, I love Matrix and hope that these issues will be resolved in time.

I had to deal with Intel Quark SoC X1000 on a Galileo board years ago, where the LOCK prefix instruction caused segfaults. Since the SoC is single threaded, the lock prefix could just be patched out from resulting binaries, before the compiler/build system was patched.

https://en.wikipedia.org/wiki/Intel_Quark#Segfault_bug

The problem being discussed is that the user of the script needs to validate it. It's great if it's validated by the author, but that's already the situation we're in.

The user is free to use a LLM to 'validate' the `install.sh` file. Just asking it if the script does anything 'bad'. That should be similarly successful as the LLM generating the script based on a description. Maybe even more successful.

Well... Maybe just have a BIOS on your system that fetches a markdown, pushes it to a LLM to generate a new and exciting operating system for you on every boot.

Wouldn't that be nice?

TBH, I doubt that this will happen...

It is much easier to use LLMs to generate code, validate that code as a developer, fix it, if necessary, and check it into the repo, then if every user has to send prompts to LLMs in order to get the code they can actually execute.

While hoping it doesn't break their system and does what they wanted from it.

Also... that just doesn't scale. How much power would we need, if everyday computing starts with a BIOS sending prompts to LLMs in order to generate a operating system it can use.

Even if it is just about installing stuff... We have CI runners, that constantly install software often on every build. How would they scale if they need LLMs to generate install instructions every time?

Maybe that is a reason for this approach. It changes the responsibility of errors from the person writing that code, to the one executing it.

Pretty brilliant in a way.

Even bringing down the "theory" to paper in prosa will be lossy.

And natural languages are open to interpretation and a lot of context will remain unmentioned. While programming languages, together with their tested environment, contain the whole context.

Instrumenting LLMs will also mean, doing a lot of prompt engineering, which on one hand might make the instructions clearer (for the human reader as well), but on the other will likely not transfer as much theory behind why each decision was made. Instead, it will likely focus on copy&pasta guides, that don't require much understanding on why something is done.

So... What you are saying is that we don't need 'install.md'. Because a developer can just use a LLM to generate a 'install.sh', validate that, and put it into the repo?

Good idea. That seems sensible.

Bonus: LLM is only used once, not every time anyone wants to install some software. With some risks of having to regenerate, because the output was nonsensical.

TBH. I never read prose that couldn't be in some way misinterpreted or misunderstood. Because much of it is context sensitive.

That is why we have programming languages, they, coupled with a specific interpreter/compiler, are pretty clear on what they do. If someone misunderstands some specific code segment, they can just test their assumptions easily.

You cannot do that with just written prose, you would need to ask the writer of that prose to clarify.

And with programming languages, the context is contained, and clearly stated, otherwise it couldn't be executed. Even undefined behavior is part of that, if you use the same interpreter/compiler.

Also humans often just read something wrong, or skip important parts. That is why we have computers.

Now, I wouldn't trust a LLM to execute prose any better then I trust a random human of reading some how-to guide and doing that.

The whole idea that we now add more documentation to our source code projects, so that dumb AI can make sense of it, is interesting... Maybe generally useful for humans as well... But I would instead target humans, not LLMs. If the LLMs finds it useful as well, great. But I wouldn't try to 'optimize' my instructions so that every LLM doesn't just fall flat on its face. That seems like a futile effort.

It really depends on the order of priorities. If the overall goal is to allow digital archeologist to make sense of some file they found, it would be prudent to give them some instructions on how it is decoded.

I just hope that people will not just execute that code in an unconfined environment.

nix possible with lix, if you can stomach nix syntax.

The spec is being implemented by Apple, who is sensitive to privacy issues.

I generally agree with your points, but I wouldn't trust Apple, or any publicly traded company, to have any kind of ethics. Just because their incentive to make as much profit as possible, leads to them trying to differentiate themselves from other companies, and thus they choose to temporarily align with privacy concerns doesn't mean they will not compromise on them, if they see better profits elsewhere.

I rather have privacy enforcing regulations like the GDPR or policies that go even further, than relying on publicly traded companies to protect their users.

Not a firefox dev, but pretty sure its 'main'

The "new" git default branch name is 'main' and 'autoland' existed before next to 'mozilla-central' and is the one where commits usually appear first.

Not only that, but the software also contained login information to the AWS backend archive servers:

https://www.techspot.com/news/107792-hacker-breaches-telemes...

This is also not surprising for BGS as was demonstrated with the multiple Skyrim re-releases, which didn't fix all issues patched by the unofficial patches, and even introduced more.

See the change log here: https://www.afkmods.com/Unofficial%20Skyrim%20Special%20Edit...

Only a very small amount of the issues fixed there where integrated into the official patch releases.

I was using Truecharts before k8s was deprecated.

The deprecation caused me to move to something more neutral and stay away from all 'native' apps of TrueNAS and migrated to ordinary docker-compose, because that seem to be the most approachable.

I was also looking into running a Talos k8s cluster, but that didn't seem to be as approachable to me and a bit overkill for a single-node setup.

But AFAIK, updates will overwrite everything, so installing anything is just temporary.

I also have a TrueNAS, but because of its limitations (read-only root file system), I came to the conclusion that, if I ever need to reinstall it, I would switch to Proxmox and install TrueNAS as one virtual client, next the other clients for my home lab.

I have found workarounds for the read-only root file system. But they aren't great. I have installed Gentoo with a prefix inside the home directory, which provides me with a working compiler and I can install and update packages. This sort of works.

For running services, I installed jailmaker, which starts a lxc debian, with docker-compose. But I am not so happy about that, because I would rather have an atomic system there. I couldn't figure out how to install Fedora CoreOS inside a lxc container, and if that is even possible. Maybe NixOS would be a another option.

But, as I said, for those services I would rather just run them in Proxmox and only use the TrueNAS for the NAS/ZFS management. That provides more flexibility and better system utilization.

OTOH with a large language model with open weights but without open training data - the closest equivalent to open source for software - you can still change its behavior very substantially with finetuning or remixing layers (from different models even!).

The closest thing to open source would be to have open training data. The weights are the binary, the training date is the source and the process of getting the weights is the compilation process.

Fintuning or whatever is just modding the binaries. Remixing different layers is creating a workflow pipeline by combining different functions of a binary software package together with components from other binary software packages.

There is also a massive union culture, which provides Germans with a great quality of life, but again, that's something probably negatively correlated to innovation.

Hmm... So if people struggle in life and have live from hand to mouth in multiple jobs to support their family and loved ones, they are more innovative?

I get your point partially. All these apps that companies put out in order to collect and manage shopping tokens or to contact their customer service would have been much better as a website.

However I still do like to have apps on my devices that just work offline, without distributing my data across services I do not control. And I also do not want to depend on a internet connection, when I am anywhere.

I like my offline Osmand/Organic Maps app to show me the trails when I am somewhere in the woods or mountains. I like my apps that instead on using some third party server, connect directly to my other local devices to share data.

IMO all (where possible) apps should be developed offline first, and only require internet when necessary, and those apps that cannot work without internet should be web apps, they do not need to be on my devices.

Maybe I am missing somehing, but I would assume that base network infrastructure like routers, firewalls and switches have a higher uptime, availability and reliability than ordinary servers.

Maybe just go with ansible or similar: https://github.com/ansibleguy/collection_opnsense

If you just want to measure user space software input lag, shouldn't it be more accurate if you print out the time when the event arrives in the kernel, and again when the mouse moves within a application window?

“The problem with sites that extract text from movies and other content is that they reduce people’s desire to pay a fair price for content, which can lead to people not seeing the official full-length movies, causing great damage to rights holders,” the anti-piracy group explains.

So a bad critique of the movie, might also reduce people's desire to pay to see the movie. So with this kind of reasoning any information you give others about a movie (or other copyrightable work) could be problematic. Pretty strange stuff...

So they either have to license their SDK with a GPLv3 compatible license as well, or have to change the license of the client to a non-GPL one.

In the latter case, IIUC their CLA (https://cla-assistant.io/bitwarden/clients) allows to do change the license unilaterally. (Not a legal expert, so please correct me if I am wrong.)

If so, then I feel strengthened again in my conviction that permissive licenses (as well as closed-source licenses) and CLAs are bad for both users and developers and should be avoided, if possible.