Simon 𐕣he 🪨 Johnson

they/them

Lord, where are you going?

  • 1 Post
  • 47 Comments
Joined 2 months ago
cake
Cake day: April 22nd, 2025

help-circle



  • Now I’m super curious about Gentoo and Portage. You don’t hear so much about compiling your own stuff anymore (probably because there’s less architectures around).

    “Nobody” runs Gentoo anymore because most distros have taken the 80% optimizations you can do and just mainlined them. This was back in 2000’s where some distros weren’t even by default compiling with -O2. Gentoo usage just proved out that the underlying code was effectively -O3 safe in the 80% case and nobody was sneakily relying on C/C++ vagaries.

    I have much less time to tinker, but my favorite new bag is Fedora Atomic (currently using Bazzite on my main desktop). I’m incredibly interested in figuring out Nix though, but I haven’t had the time. Immutable distros are honestly something incredibly useful for both power users and normies. The main issues I’ve had with Fedora Atomic have really been around vagueness in the “standard” but they’re still figuring things out as far as I can tell.




  • There are plenty of sha1 implementations that are more readable and sensible and less readable and sensible. This portion is simply an manually unrolled loop (lmao these gcc nerds haven’t even heard of Gentoo) of the hash chunk computation rounds. Hash functions aren’t “impenetrable” they’re just math. You can write math programmatically in a way that explains the math.

    The point of this post is actually things like x[(I-3)&0x0f]. It’s entirely the same concept as coercion to manipulate index values this way. What’s funny is that void pointer math, function pointer math, void pointers and function pointers in general are typically seen as “beyond the pale” for whatever reason.

    Beyond that if you know C you know why this is written this way with the parens. It’s because C has fucked up order of operations. For example a + b == 7 is literally “does adding a + b equal 7”, but if you write a & b == 7 you would think it means “does a AND b equal 7”, but you’d be wrong. It actually means does b equal 7 AND a.

    Furthermore a & (b ==7) makes no sense because b == 7 is a boolean value. Bitwise ANDing a boolean value should not work because the width of the boolean is 1 bit and the width of the int is 8 bits. ANDing should fail because there’s 7 void bits between the two types. However the standard coerces booleans in these cases to fit the full width, coercing the void bits to 0’s to make bitwise ANDing make sense.

    Beyond that asking what the memory size of a variable in C is a fools errand because the real answer is “it depends” and “it also depends if someone decided to ignore what it typically depends on (compiler and platform) with some preprocessor fun”. Remember how I said “void pointers” are beyond the pale? Yeah the typical “why” of that is because they don’t have a known size, but remember the size of something for C is “it depends”. 🤷

    Almost every language has idiosyncratic stuff like this, but some let you make up your own shit on top of that. These kinda low hanging fruit jokes are just people virtue signaling their nerddom (JS bad am rite guis, use a real language like C), when in reality this stuff is everywhere in imperative languages and typically doesn’t matter too much in practice. This isn’t even getting into idiosyncracies based on how computers understand numbers which is what subtracting from 0x5F3759DF (fast inverse square root) references.


  • I thank god every day people who make these comics are too stupid to open gcc’s sha1.c because they’d see shit like:

    #define M(I) ( tm =   x[I&0x0f] ^ x[(I-14)&0x0f] \
    		    ^ x[(I-8)&0x0f] ^ x[(I-3)&0x0f] \
    	       , (x[I&0x0f] = rol(tm, 1)) )
    
    #define R(A,B,C,D,E,F,K,M)  do { E += rol( A, 5 )     \
    				      + F( B, C, D )  \
    				      + K	      \
    				      + M;	      \
    				 B = rol( B, 30 );    \
    			       } while(0)
    
          R( a, b, c, d, e, F1, K1, x[ 0] );
          R( e, a, b, c, d, F1, K1, x[ 1] );
          R( d, e, a, b, c, F1, K1, x[ 2] );
          R( c, d, e, a, b, F1, K1, x[ 3] );
          R( b, c, d, e, a, F1, K1, x[ 4] );
          R( a, b, c, d, e, F1, K1, x[ 5] );
          R( e, a, b, c, d, F1, K1, x[ 6] );
          R( d, e, a, b, c, F1, K1, x[ 7] );
          R( c, d, e, a, b, F1, K1, x[ 8] );
          R( b, c, d, e, a, F1, K1, x[ 9] );
          R( a, b, c, d, e, F1, K1, x[10] );
          R( e, a, b, c, d, F1, K1, x[11] );
          R( d, e, a, b, c, F1, K1, x[12] );
          R( c, d, e, a, b, F1, K1, x[13] );
          R( b, c, d, e, a, F1, K1, x[14] );
          R( a, b, c, d, e, F1, K1, x[15] );
          R( dee, dee, dee, baa, dee, F1, K1, x[16] );
          R( bee, do, do, dee, baa, F1, K1, x[17] );
          R( dee, bee, do, dee, dee, F1, K1, x[18] );
          R( dee, dee, dee, ba, dee, F1, K1, x[19] );
          R( d, a, y, d, o, F1, K1, x[20] );
    

    And think, yeah this is real programming. Remember the difference between being smart and incredibly stupid is what language you write it in. Using seemingly nonsensical coercion and operator overloaded is cringe, making your own nonsensical coercion and operator overloads is based.

    That’s why you should never subtract things from 0x5F3759DF in any language other than C.





  • Lol the suggested hardware for usable performance is ~$50k MSRP for the GPUs alone and that’s an SXM5 socket so all proprietary extremely expensive and specific hardware.

    My PC currently has a 7900 XTX which gives me about 156 GB combined VRAM, but it literally generates 1-3 words per second even at this level. DDR5 wouldn’t really help, because it’s a memory bandwidth issue.

    TBH for most reasonable use cases 8 bit parameter size quantizations that can run on a laptop will give you more or less what you want.


  • Neo-liberalism was supposed to pause all of these conflicts, but because America is no longer a leader in the world, because it sucks

    Lmao. You realize that this idea in the modern era is based on one gigantic moron who’s only job it is to be a stenographer for power had this big brain moment where he wrote down “No two countries with McDonald’s have ever gone to war”.

    Then he changed it to any country that has a company that sells stuff to Dell won’t go to war with each other.

    The capitalist peace theory has never been true.



  • Just a reminder that Reddit was once difficult for people to understand.

    I honestly don’t believe this at all.

    Snapshat was popularized by a generation that grew up only using apps, and it was designed to be obtuse, mysterious and difficult to learn in comparison to other apps as a feature. It grew regardless.

    To be honest though, I’m a bit disappointed by the other users here. The quality of comments is really poor, both idiotic and adversarial. I’m talking fox news comment section level.

    Yeah so is reddit. The best moderation and engagement in fediverse typically exists in the highly moderated communities that people constantly complain about not respecting their freeze peach and antisocial tendencies.






  • It seems naive to believe that the Chinese firewall acts purely as a benign protector of the assaulted Chinese citizen. Chinese people are not like stupid children in need of protection, they are smart and strong.

    Yeah it’s equally naive to believe that the Chinese firewall acts purely as a hostile censor, Chinese people aren’t uneducated, oppressed, impoverished individuals, they are accomplished, politically active, and well to do. The Chinese people have comparatively derived a larger individual and collective benefit from their government than Americans have in the last 50 years.

    If you read actual comparisons of “censorship regimes” there are tons of commonalities that are just ignored by Westerners and their Chinese counterparts are made out to be uniquely evil and beyond the pale. For every news article you read about how the National Security Police invites a satirist to “drink tea” you’re ignoring all of the times the FBI does the exact same thing, and uses various psychological tactics to escalate into a position of legal authority to get around their limited authority to collect evidence.

    You know why it’s “soooo hard” for the cops to arrest rich people even if they know where they are? It’s because the tactic of escalatory arrest (an arrest that happens without a warrant as the result of an “investigation”) doesn’t work on rich people, they have gates, intercoms, staff, and know their rights. They aren’t easily cajoled into the position of opening their home to a cop, or allowing a cop access to their body. Isn’t is very strange that these very technical legal distinctions aren’t told explicitly to the “freedom loving people” of America? Meanwhile the agents of “evil Chinese government” don’t need to play games like this, because the cards are all on the table.

    People in other countries get “dissapeared”, but when ICE or the Department of Corrections shuffles prisoners around for political purposes such as Mahmoud Khalil. People in other countries are “political prisoners” but in America we have the WGAD which is a nice rhetorical trick so that the government can “honestly label” it’s political prisoners (upon a opaque and deliberatley difficult review process only undertaken by those who actually want to go through it for the benefit of being labeled a political prisoner. WGAD has not authority to enforce anything.

    People in other countries get thrown in jail because of political corruption, in the US saying such a thing is insulting the honor of the judiciary as a whole, a judiciary that allows the same practices the jailed Stephen Donzinger for the crime of taking on a legal case against Chevron in Ecuador. Furthermore it’s processes are abused to provide legal procedural punishments for missteps in engaging with the system such as the contempt charges the Donzinger case. Donzinger is still disbarred and cannot leave the country, despite winning all of his appeals. All at the behest of a corporation that doesn’t want to create a precedent that it must pay for poisoning people.

    The reality here is that you’re not actively comparing things, you are just going on hunches or whims, and if you take a look that’s how a lot of information you receive is actually structured. That is what allows labels like “authoritarian” to have a spooky evil weight. In essence the US has simplify codified the abuse into law, which is how it gets around these icky little moments of “Are we the baddies?” the reply is a thought terminating cliche of “No we’re all just following legal orders, in the freest country in the World”. China doesn’t need to Nuremburg because it’s goal of social cohesion ensures that people understand how and why things are happening to them.