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

  1. Too late. Bootcampers already inflicted a lot of damage and it will be hard to reverse it.
    1
  2. People who still use regular expressions in 21st century - who tf are they?!?
    1
  3. Exactly this kind of empty bloviators.
    1
  4. Compilers are simplest things in existence. You can write a compiler in anything. You don't even need a Turing-complete language to write a compiler. Just pretty much any language is ok.
    1
  5. So, an incompetent code monkey confirmed. GDB can be scripted, making your code monkeying in an interactive debugger truly laughable. GDB allows plugins that decode various complex data structures in a readable way, good luck doing it with your IDE. And, more importantly, any IDE with a debugging functionality is still just a UI with, most often, the same GDB inside. Code monkeys are so funny, especially when they are arrogant.
    1
  6.  @metaltyphoon  look, monkey, you're out of your depth Go away, pitiful kid, and don't interfere again when adults are talking.
    1
  7. not to mention that games are literally a toy code. Can't get more toy than this. But of course for a little dumb kid like you games are a pinnavle of software. Go away, play your dumb games, and never dare to voice your worthless opinions again.
    1
  8. You have no faintest idea of how to use LLMs. Where is your feedback loop? Where is critic? Where is code sandbox? Why did you fail to use SMT for reasoning?
    1
  9.  @mortiz20101  if you failed to comprehend what I'm talking about, you probably should not use LLMs. 1) Feedback loop: feed the result of testing the LLM output back into LLM, with all the relevant results (syntax errors, test failures, static code analysis output, etc.) 2) Critic: every time an LLM is producing an output, do it a few times with the same prompt, and then use another LLM prompt to criticise all the outputs and select the best one out of them. 3) Code sandbox: give LLM a tool to run arbitrary code in a safe sandbox. Use inference harnessing to ensure the tool is used immediately as the call appears in the output. 4) SMT, Prolog, etc. - LLMs cannot reason, obviously. But they can translate an informal problem into a formal language. Which can then be processed by an SMT solver, a Prolog interpreter, or whatever else you use as a reasoning tool. You have a lot to learn. Do it. Or stay ignorant.
    1
  10.  @BHBalast  this is exactly what he's doing wrong - using an inferior product, Devin, instead of building a rather trivial system on his own, out of open components, using the well known tools and methods I described. And not just using this inferior system, but also making long reaching conclusions about capacity of AI in general.
    1
  11.  @hungrybeaverontheleaver  how exactly? I build these systems routinely, so I have all reasons to laugh at those who think Devin is a proper AI coding agent.
    1
  12. It depends. There's nothing for Rust comparable in quality to C++ static code analysis tools. Rust gives you a better baseline safety for free, but for really high stakes safety it's not ready.
    1
  13.  @Goodwarrior12345  so, you don't understand mathematics, and you don't understand programming. Right. Logic literally is mathematics. There is no logic outside of mathematics. And programming, as it is dealing with formal systems, also a form of mathematics and nothing else.
    1
  14.  @Goodwarrior12345  only the formal logic is relevant to programming. And formal logic is mathematics. It's closely related to fields that are essential for programming as well: proof theory, type theory, denotational and operational semantics, term rewriting (and, therefore, cellular automata, lambda calculus and Turing machines). Physics is not mathematics, it's using mathematics as a language. But programming is all about dealing with formal languages. Also, it's quite funny that you think that mathematics is about numbers.
    1
  15.  @Goodwarrior12345  "You don't really need any of it in a practical context. " - only if your "practical context" is producing bloated, crappy, ultimately useless primitive software that should have never existed in the first place. "understand how computers work" - sorry, it's mathematics. Pure and simple. "Hell, you don't even need to know what a formal language is to do good programming. " nope, those who do not understand language theory are extremely crappy programmers who should not be allowed anywhere near any serious development. "As long as you know how a CPU works and how it processes instructions, you're pretty much good to go. " and you don't know how CPU works and how it processes instructions. If you don't know what Hoare triple is, you don't understand how CPU works. " Programming is all about dealing with the CPU by using formal languages as a communication tool." you don't really understand languages. Or programming, in general. It's not a "communication tool". It's a formalism, a way to express a problem solution. The real tragedy of this industry is hordes of self-taught or ill-taught "programmers" who believe for no good reason at all that they know how to program. And it's hilarious that even their trail of failures cannot convince them they know nothing and should have never even started programming.
    1
  16.  @Goodwarrior12345  Carmack is a game developer. Pretty good at it, and he is indeed using quite a bit of mathematics in doing it (optimisation is hard), but the moment he wanders away from games he fails massively. I've had a lot of fun and WTF moments with Oculus. Someone who think math first would have avoided all those issues. And mathematics is not a science, it is a language in which you describe languages and a tool for dealing with languages. No more and no less. If you cannot use this tool, you cannot use languages propetly.
    1
  17.  @StarForgers  it is the only relevant form of logic. Aristotle won't help you to write a code free from dead locks and race conditions. Formal logic and pi-calculus will.
    1
  18.  @luckerooni1153  Lol. Aristotle logic is *formal*, and therefore it is still mathematics.
    1
  19.  @BigAggLP  the thing is, in order to be able to solve problems, to see how things are connected, you need a systematic knowledge without gaps, without patches of magical thinking. I.e., a knowledge built from the first principles all the way up. Unavoidably having a lot of mathematics in the foundation. If you don't have such a foundation, no amount of googling will save you, you won't be fit for any problem solving job.
    1
  20. ​ @BigAggLP sure, but you still need a ton of fundamental knowledge in order to be able to even start exploring those possibilities.
    1
  21.  @BigAggLP  you do not need to know everything, or even a lot. You need to know what you know systematically, without mythical thinking gaps. This is impossible without a significant amount of mathematics. Once you have such a base, it is very easy to add new knowledge.
    1
  22.  @strappedwithkrylon  I am talking about the formal logic, the one that matters.
    1
  23.  @chrisdaman4179  bullcrap. Nobody should be using regular expressions since 2006, when Ford published his paper. Regular expressions are outdated and useless. Ignorant "developers" like you who carry on using the outdated and outright broken technology are the reason why software is so crappy. Ever heard of PEGs? No? Then go back to school.
    1
  24. More like the end of the code monkeying. Still not capable of doing any of the hard programming tasks.
    1
  25. Would you trust a surgeon who had the same ideas about college education to operate on you? No? Would you want to drive over a bridge that was built by an engineer who thought the same way about college education? Probably no. Then why do you think it's ok for a software developer?
    1
  26.  @nieczerwony  nobody would let a surgeon into a residency program without a very solid theoretical AND practical background from a university. It's not just books - it's a lot of lab work as well, and a lot of practical work. And no, you will never get this amount of knowledge this deeply embedded from just reading books. And no, software development is not any different. You can try to get the theory from books, but good luck with forming any systematic, solid understanding of it. This is what higher education is for, to guide you into forming a systematic, comprehensive knowledge. I never met a single person who could do it in any field from just books, without already having an experience of doing the same in any other field (i.e., at least some university level degree).
    1
  27.  @Dogth_xd  you make very little sense. Does it a lack of formal higher education show, or simply being deliberately obtuse? Software development is an engineering discipline, and programming IS mathematics. Period. If you don't recognise this as a fact, you simply know nothing about software engineering and programming. The fact that there is less opportunities to cause severe harm is irrelevant, and you're evidently underestimating the actual damage that undereducated code monkeys inflict on our society. As for Kalashnikov, guess who worked for him? One little guy Hugo Schmeisser. Rings any bells? You, uneducated people, are so far behind those who actually have a systematic knowledge that you simply have no mental capacity to comprehend how huge the chasm is, and how much you're missing, how incapable you are in comparison to anyone with a proper education. Sorry to break the news for you, I understand that you need some psychological comfort, some rationalisation for your ignorance, but I'm not in a generous mood and have no intention of feeding your rationalisation attempts.
    1
  28.  @32gigs96  you don't undertand macros, do you? Where do you think your "ergonomic" language features come from? Will be really hilarious if you admit that for you some pathetic crap like Python is somehow "ergonomic".
    1
  29.  @raffimolero64  DSLs place is exactly *everywhere*. If you're not using eDSLs for everything, you're inefficient and your code is bloated, unmaintainable and imperfect in so many ways that I just cannot list them all here. As for stability - only recently Rust started to stabilise, so it's clearly not the main selling point.
    1
  30.  @j.trades9691  there are no performance lossess for eDSLs, and maintenance of a hierarchy of languages is far cheaper than maintenance of a mess of leaky abstractions.
    1
  31.  @logicaestrex2278  code monkey detected. You should not be allowed anywhere close to any serious engineering, monkey.
    1
  32.  @logicaestrex2278  I can see you're butthurt Good. It must be really hard for you to perform all the mental gymnastics justifying your ignorance and incompetence.
    1
  33.  @SoftBreadSoft  Indeed, Julia is a proper meta-language.
    1
  34.  @TerminalAddict  did I ever say Rust is ideal? It is just better than C in one specific aspect - macros. And Rust macros are.still not quite what you'd expect from a proper language - see Common Lisp for a reference of how things should be done.
    1
  35. Do people who do not find it motivating and rewarding really should even be in this profession?
    1
  36.  @nobleradical2158  nope, C++ compiler does not know anything about any smart pointers (and it should not know). It can guarantee that it does not assume aliasing if two pointers point to two different data types though, which does not really help in most of the numeric applications. In plain C, `restrict` would be respected by the compiler and it will assume two pointers are not aliased and do not point to overlapping regions, allowing a lot more optimisations than if there was a possibility of aliasing.
    1
  37. I found that the best coding quality for LLMs is when they code in a language they never seen (and your RAG loop explainst it as they go). This way they use generic programming knowledge distilled from tons of code, but not awful practices of the majority of code pieces out there.
    1
  38. ​ @josevargas686 as if access to computers is enough to produce engineers. You need a proper higher education system, wirh established traditions and well staffed. Eastern europe lost most of it when the Warsaw Pact collapsed, which leaves us with India and China. And they are not enough to satiafy the current huge demand.
    1
  39. This approach only works if all you do is some very narrow code-monkeying. Try writing some hard real-time firmware for a reasonably priced MCU in Java...
    1
  40. We used to have Literate Programming... Now, few decades later, developers devolved to this sad state.
    1
  41. Not really. Things are still manageable, if you just ditch all the overengineered crap that bootcampers and self-taughts defecated over the years. There's very little value in all those layers upon layers of overengineered abstraction, and you'll be better of without them.
    1
  42.  @PiratePawsLive  lol, there is very little value in interactive debugging of this kind, especially for C++, where debugger will allow you to inspect the consequences of the problem (e.g., memory corruption), and will be useless in identifying the cause of the corruption that happened half an hour earlier of running the code under heavy load, in a very different part of the code.
    1
  43.  @PiratePawsLive  using an Arduino is almost cheating, as they (most of them) have built-in programmers with a USB interface. Although, if I remember correctly, Arduino IDE have a very limited debugging capacity...
    1
  44.  @nchomey  I'm not sure if I can name a single scenario where interactive debuggers would actually shine. Regardless of the type of the problems being solved. No matter what you do, you have data structures with complex semantics (e.g., if it's some GUI - you have your widget tree, etc.), while debuggers will only let you inspect the data structures on a very low level. Unless you invest a ton of time into implementing custom debug printers (e.g., see LLVM code base for quite a few examples of such). Even if the type of problems you're solving allows you to stop right where the problem happened, and not where it simply manifested itself, it's still a lot of extra steps you have to do vs. simply reading a well-formatted trace of the events that lead to the problem.
    1
  45.  @PiratePawsLive  automated code instrumentation is a very powerful tool for debugging (e.g., see Valgrind, or clang sanitisers). Until you know where the problem really is, you can instrument everything and log all suspicious events. Far easier to narrow down from having excessive information than to try finding a needle in a haystack without any relevant information at all.
    1
  46.  @nchomey  it's "print+trace+tons of other stuff" when bug happened. Logging is print+trace+tons of other stuff before the bug happened. See the difference? If you're preparing your comprehensive logging infrastructure in advance, you'll already have all the information you need to pinpoint the error when it happens. If you jump into an interactive debugging when you see an error, you start from scratch, you have nothing to work with.
    1
  47.  @PiratePawsLive  this is what logging tiers are for. You have an ability to log everything, ad nauseum, but in a normal operation you only log critical errors. Once an error happened - re-run with logging maxed out. You can also have logging tiers configurable per module, if you know where you want to look.
    1
  48. Meh. The average code quality is so low there's nowhere to fall anyway.
    1
  49.  @metaltyphoon  more like you don't have any idea. Let me guess, you're some kind of a web code monkey? Think of a case when your code runs on multiple MCUs, all communicate via various media, like RS485, I2C, CAN, etc. Protocols are time-sensitive. All MCUs workload is hard real-time. All logic depend on properly processing the data external to your system, that you have no control over. Now, good luck debugging any of this with your puny toy approaches, like stop and step through.
    1
  50. Entire web stack, starting from w3c standards, is fundamentally broken, so having any opinions at all on any of it is pointless. Just enjoy your suffering, web developers, it was your conscious choice to get into this awful mess.
    1