Comments by "Vitaly L" (@vitalyl1327) on "ThePrimeTime" channel.

  1. No, looks like some abstract register-based VM.
    1
  2. Or, even more likely, MISRA-C.
    1
  3. All accidentally designed languages suck. You cannot evolve a proper language from a pile of crap. And PHP was not even a proper language originally.
    1
  4. IT guys often care if some dimwit developers insist to use some wide open security hole like vscode which keep sucking unverified extensions like there's no tomorrow. In the more sophisticated workplaces they just ban this kind of crap indiscriminately.
    1
  5. LLVM IR is not platform-independent at all. There was an attempt to build the previous SPIR standard on top of LLVM IR, it failed. SPIR-V is the way now, and it's got nothing to to with LLVM IR. Now, even if you go a few levels up, shader source code, or OpenCL kernels, while semantically portable, are not performance-portable anyway. GPU architecture specific optimisations have to be done on a very high level and cannot be automated.
    1
  6.  @TheRealXartaX  found an uneducated dimwit who have no idea how to parse in 21st century. Ever heard of PEG, dimwit?
    1
  7.  @TheRealXartaX  found an ignorant regex enjoyer. You people are hilarious. So primiitve, so dumb, yet so confident.
    1
  8.  @M0J0-RL236  Huh? Every competent developer will prefer a PEG or a GLR to regular expressions. Only the incompetent ones do not know the alternatives.
    1
  9.  @M0J0-RL236  I do not work with incompetent people to start with. And I'd say, those who tolerate incompetence are indeed horrible engineers and awful teammates. I do not hire such people.
    1
  10.  @M0J0-RL236  look, people working for a company like Crowdstrike are supposed to be professionals. Educated, well-rounded engineers. Not some kind of bootcamp "graduates", high school dropouts or whatever else is now trying to sneak into this industry. They are supposed to know the basics. It is ok when you have to teach an intern. But when an "experienced" professional does not know fundamental things, it is a moral obligation to mock them viciously. Such people have no place in this industry and are a root of all its ills.
    1
  11.  @TheBusttheboss  there are some fundamental limitations in Emacs, the most evident one being its single-threaded blocking nature. Also, elisp itself is a bit outdated, though there are remedies for it.
    1
  12. The only people who complain about gatekeeping and "elitism" are those who absolutely should have been kept at the gate and never let in.
    1
  13. The problem here, they don't get the job done. They produce utter crap that can superficially pass if you don't look at it too closely, then crumble at a first attempt to use it as intended in production.
    1
  14. You'll end up with glaring gaps in your understanding, and without any ways of even figuring out where those gaps are.
    1
  15. Almost all software is written by incompetent, crappy coders. So it's pretty much obvious that most of the software is an utter crap. But hating it? It's like yelling at clouds. Crap software is just a force of nature.
    1
  16. Do you realise how far Java is from the basics?
    1
  17. Clean code zealots are funny, when pressured they always finally admit that their primary motivation is to have an excuse not to write comments.
    1
  18.  @VuxGameplays  he was talking about new developers in general. It's just that everyone default development to web development, which is funny and unnecessary.
    1
  19.  @nezu_cc  I did a ton of different things in the last three decades. That included scientific compute (analysing experimental data from CERN), along with building those detectors and hardware/software triggers for them, working on specialised CAD engine for ship building, working on compilers for GPGPU along with GPU hardware design, working on video compression and low latency video communication, high-frequency trading, industrial robotics. People I worked with had, of course, different paths, they worked in areas from game development to embedded automotive / aerospace, medical robotics, and many many more areas. Dozens and dozes. The world does not revolve around web. Web is the least interesting domain in IT, yet for some weird reason every new developer gravitate towards it.
    1
  20.  @VuxGameplays  there is a whole huge world outside of web. System development - operating systems, compilers, DBMS engines, etc. Embedded development - real-time, control, safety, etc., telecomms, robotics, HPC, CAD/CAE, automation in general, and many many more exciting and fun areas.
    1
  21.  @DirtyBigHoss  real developers are engineers, not code monkeys.
    1
  22.  @DirtyBigHoss  so, you do not see the difference between engineers and code monkeys?
    1
  23. Right, some people considered Andy Warhal an artist. The same people may get amused by the AI "art".
    1
  24.  @chrisdaman4179  the problems start when code monkeys start using "pattern matching for strings" to parse strings. And dim code monkeys cannot think of any proper way to parse. Very rarely people use regular expressions to answer the question "does this string match the pattern, yes or no?", they use them to parse the string or at least split it in parts. Those who do it are ignorant, incompetent and should have never been allowed anywhere near any coding.
    1
  25.  @TheRealXartaX  the funny bit here is that it is you who have no mental capacity to understand programming. Hilarious, is not it? Regular languages are outdated and have no use in the 21st century. Go back to school, you missed all the important bits of CS.
    1
  26. They picked the application that's deliberately very bad for Rust - bytecode interpreter. It's not good for Zig either. Plain C with computed goto extension allows to easily implement indirect threaded code interpreter, while Zig and Rust had to stick to puny slow switch statements. I'd never write a bytecode VM in Rust or Zig, they're simply not fit for this purpose.
    1
  27.  @ThePrimeTimeagen  there is no equivalent of a computed goto C extension, so there is no way to implement an indirect threaded code interpreter (see OCaml bytecode implementation for a nice and clean example of this approach). Threaded code is way faster than any switch can ever be.
    1
  28. Perl is powerful, PHP is not.
    1
  29. ​ @lukivan8 they should not be. In many counties engineering is a regulated profession. Just like some "self-educated" nobody does not have a right to call himself, say, a surgeon, and will be punished if he dare to, nobody should be allowed to self-style as an engineer.
    1
  30.  @17Codiferus  learning a new platform from dcratch is a negligible fraction of time of a development cycle of a mildly complex project for a competent engineer. If this is not the case, the engineer is not an engineer and should have chosen another career path.
    1
  31.  @ToastedBreadcrumbs938  this is what experience is for - to be able to pick up the best practices and the spirit of a given technology quickly.
    1
  32.  @dpgwalter  we need much more.gatekeeping in this industry. Incompetence cause a lot of real life suffering, and should not be tolerated. Would you tolerate a self-taught incompetent surgeon? Automotive engineer? No? Then why all this welcomimg and openness in software? Let everyone in, and be kind to their mistakes, yada yada yada.
    1
  33.  @svuvich  tell it to the Horizon survivors. I will do everything in my power to keep people like you away from this industry. Because incompetence is not without the consequences for the general public. Now, answer, why do you think the ignorant uneducated ones should be allowed into the industry? And try to apply your reasoning to medicine or civil engineering, to see how deranged any possible argumemts are.
    1
  34.  @svuvich  tell it to the Horizon survivors. Incompetence is not without the consequences for the general public. Now, answer, why do you think the ignorant ones should be allowed into the industry? And try to apply your reasoning to medicine or civil engineering, to see how massively wrong any possible argumemts in favour of your position are.
    1
  35.  @taragnor  a little slow system is not functional, because low latency is the main functional requirement.
    1
  36.  @danielburger2550  not on an immutable rootfs, not with an A/B bootloader, not with overlay updates.
    1
  37.  @danielburger2550  in some distribitions (in particular the ones for corporate use).such a configurarion is a default.
    1
  38.  @CallousCoder  there is no innovation in overengineering, it stems exactly from people not knowing what to do and applying a few tricks they know to build systems that should have been built from the very different blocks. They try to sledgehammer their few known methods into every hole they see, and it results in an ever growing complexity. Academics have a tendency for building systems based on the first principles, and much less inclined to overengineer things. Also, there's a bit of a culture and management problem here - for example, in Amazon the only way up is to create a new "product" and apply political pressure to push it. Resulting in literally hundreds re-implementations of the same old crap, competing with each other end eventually bestowed upon poor customers in cycles. Google is a little bit better than that, but still does a similar thing. I know quite a few people who escaped FAANG exactly because of their disdain for this culture.
    1
  39. If you cannot get an intelligent behaviour out of LLMs, it is a you problem.
    1
  40.  @carles.noguera  you're overcomplicating things. You don't need anything else if you have `subleq`.
    1
  41.  @gregorymorse8423  name a microarch on which it is 1 clock cycle. It is 4 on Skylake, and in general 2-5 clock cycles, as you'd expect. Even on GPUs it is 3 or more clock cycles, with all the huge LUT optimisations.
    1
  42.  @gregorymorse8423  and another thing for you to keep in mind, if you're talking about RSQRTSS and alike, they do even less than that, they only do LUT lookup for the first 12 bits of the result, and no consequent iterations, so the result is a total crap. The infamous Quake function did two iterations.
    1
  43. Sucking was the main design principle, so of course it sucks. By design.
    1
  44.  @123456crapface  you're funny, kiddo. Did not you know, kiddo, that real software is dealing with data that cannot be trusted to any 3rd party cloud? No real, mission-critical project will ever touch anything cloud. Leave it for kids and their dumb toys.
    1
  45.  @123456crapface  real software (e.g. in a HFT) is dealing with data that cannot be trusted to cloud and needs latencies far lower than anything that's possible with a cloud. Cloud is just a toy.
    1
  46.  @123456crapface  real software is dealing with a real data, that cannot be trusted to any cloud.
    1
  47. Nah. I'm elitist af, I believe there are people who are immensely better in what they're doing than the others, that a lot of hard work can make you orders of magnitude more capable than those who did not put the hard work in. I do not believe that such a thing as "talent" exists.
    1
  48. Java bytecode is verifiable and provably safe in load time. Rust code is partially safe somewhere in compile-time, and even that is not guaranteed. Go figure.
    1
  49. He forgets that there is no equivalent of a Total Functional Programming anywhere in OOP. Therefore, even comparing them makes no sense at all.
    1
  50. On the other hand, why work on something that is not fun?
    1