Laws are often purposefully vague to account for loopholes and changing circumstances/public attitudes though. It’s the task of courts to define the exact boundaries – and since jury trials aren’t a thing, the interpretations of any higher court will basically ammend the law for lower courts.

Did you notice they changed their display name to look like an alt account of yours?

What the actual fuck

I pity those who don’t get the reference/inspiration.

10/10 Kunstwissenschaftliche Analyse

Ah, alright. Fair enough.

Though in that case my earlier assumption about AWS doesn’t hold true either.

Could this violate Signal’s security model?

Your country would be able to determine exactly with whom you’re speaking by nature of IP’s being public.

It’s arguably more difficult if there’s a server intermediary, especially when speaking to someone in a different country.

For chat alone they are right.

For real time communication (like calls), you need global data centers. Hyperscalers aren’t required but make it easier to deploy globally since you rely only on a single provider.

Fortunately not too bad.

It’s just:

SUSE Linux Enterprise 16 supports the Model Context Protocol (MCP) standard and can bridge to any LLM provider.

So it doesn’t include AI but you can connect it to AI. Also openSUSE is seemingly unaffected.

If enterprise customers want to continue – and pay for – their AI circlejerk then let them.

You can also be spied on by me if you’d like to.

Just send your search queries to me.

For extra espionage, I’ll even make sure to use non-privacy respecting search engines (like Google, Yandex, Baidu) and AI for your queries.

Not the same group. The EU is very much heterogenous and not just a homogenous blob.

I’m pretty sure this group was also publicly against chat control.

Please use correct terminology. It’s the Death Disk. Or Pancake. Or Pizza.

According to the FAQ (warning: very wordy):

The goal of the project is not to slightly improve some aspects of insecure devices and supporting a broad set of devices would be directly counter to the values of the project.

The expectation is for people to buy a secure device meeting our requirements to run GrapheneOS. Broad device support would imply mainly supporting very badly secured devices unable to support our features. It would also take a substantial amount of resources away from our work on privacy and security, especially since a lot of it is closely tied to the hardware such as the USB-C port control and fixing or working around memory corruption bugs uncovered by our features.

Das ist K&R C, du Banause!

*dooppelvookaalen

I meant basic memorization, not any advanced stuff. If you have to re-derive everything basic from scratch again and again, you will be less effective at advanced stuff.

This is not to say the basic stuff should just be memorized. Rather, it should first be understood and only then be memorized.

And definitions must be memorized, otherwise you’re screwed. For instance, try proving something is a group if you forgot the definition of a group. Yes, the definitions have reason for being the way they are (which you will likely learn) but definitions just cannot be derived from your mind during an exam.

In OP’s example with memorizing multiplication tables instead of doing them on-the-fly: This is a core skill required for so much later on. You don’t want to waste time and energy thinking about how e.g. 7•8 = 7•2•4 = 14•4 = 14•2•2 = 28•2 = 56 because that’s a quick way to lose focus. Especially if you – like me btw – have to invert a 7x7 matrix with two variables x,y put in a bunch of positions (and linear combinations of them) in an exam.

Edit: substitute unescaped *s with •

Strongly disagree that memorization isn’t important. It’s THE foundation to be able to do effectively do more advanced stuff.

Take the equation (5678 • 9876). Use long multiplication and you only rely on doing a bunch of single digit multiplications and additions. It’s so much faster to be able to instantly know each step instead of having to recalculate these “atomic” steps again and again in your head.

You generally don’t need to be able to solve multiplications involving double digits in your head. It’s nice-to-have but otherwise useless, as long as you’re able to calculate the ballpark of the result.

For example, (38•63) is roughly 2400 and I can then calculate it on paper instead of in my head.

Head calculations are just so much more error-prone than written calculations. Don’t do them if you can avoid them. There’s a reason why math students (at a university) are infamous for being unable to make the simplest calculations in their head. It takes effort that could be spent somewhere else.

I think it can apply to the most general workflow with branches as well, where branches are used to develop features and then later merge them.

After all, any new branch is basically a “remaster” until merged back in, which is when the original master becomes the remaster.

Sure, the analogy isn’t perfect because in music the original master isn’t supposed to change – but the entire purpose of a version control system is to change the “master record”, i.e. what’s deployed to production.

I think master came from master record. It makes some sense, as all other branches would be derived from the master branch.

Processing would be quicker?

This would add significant processing overhead because of conversions everwhere. There is a lot of string processing going on inside your CPU already, I’m reasonably sure this would add a measurable overhead.

Plus I believe this would cause even more string related security vulnerabilities to emerge because of additional code that needs to be maintained.

This is somehow simultaneously the best and worst setup I’ve ever seen.

Absolutelynotme_irl