Comments by "dkosmari" (@dkosmari) on "Al "Slop" Bug Reports Hurting Python, Curl, u0026 Other Open Source Projects" video.

  1. AI bros, Quantum bros, Crypto bros, they all get riled up when their fake gods get blasphemed. They'll soon flood the comments with "just use AI to fight AI, let the AI filter out the slop" followed with "let the AI fix the bug", and "let the AI write the whole source code."
    21
  2. You're assuming the malicious agent is dead broke, and doesn't have a couple hundred bucks to sabotage your project.
    18
  3. The problem is that it doesn't look dumb. You'll get a detailed description of the problem in the code. It will all look correct. You'll spend some 15 minutes reading the bug report, the bug description, comparing with your code... then eventually you spot it's describing some bogus interaction between modules, that doesn't actually happen. You reject the bug report, you ban the sender. You think you've won, you spotted and rejected the fraud. In reality, you lost. You wasted valuable time dealing with a bogus report. Now imagine having 1 or 2 reports per week, then it increases to 100 reports per week. You know there are a couple of legitimate bug reports in there, but how can you tell which one?
    10
  4. Yeah, now the project needs to set up a payment system in order to even exist. And the legitimate user reporting a bug, needs to register with some payment system. You clearly didn't think this through, huh? That's a sure way to kill all community participation with the project.
    6
  5.  @pavel9652  You're clearly one of the scammers shown in this video. Yes, that report shown was malicious. You have to be very dishonest to claim otherwise.
    6
  6. And how long does it take you to tell it doesn't make sense? You get a full bug report, with "steps to reproduce", with "sample configuration", with "compilation options", with code fragments showing where the bug is manifesting, etc. How can you tell it's slop? How long does it take to notice something is off? You're thinking of 1 AI slop bug report out of 10 reports. What if it's 90 out of 100?
    6
  7. No, if the AI was actually intelligent, and could discern a real bug report from a fake one, the scammers would just ask that same AI to generate better, more convincing slop.
    2
  8. What a great idea, let's leave potentially critical bugs dormant for 2 years. What could possibly go wrong? Do you work for Microsoft?
    2
  9. You're talking as if there weren't VPNs, and bot networks, to bypass IP bans.
    2
  10. Which AI? The same AI that hallucinated the original bug report? If the AI could reason the bug report is slop, why wouldn't the scammer just ask the AI to "generate a bug report that's doesn't look like slop"?
    2
  11. You don't know how coding works. Separating a massive messy piece of software into well-defined components is always a productivity improvement. They didn't create multiple libraries, they separated code that should have never been mixed up together in the first place.
    1
  12.  @PetraleiTanyl  I wasn't going to call you out, but the fact you said "the X11 community" is good evidence that you're the one that doesn't know what you're talking about. If every Xorg maintainer decides to abandon the messy code base to start from scratch, who are you to tell them they're wrong? Wayland has many problems (mostly bad integration with everything else) but it's not because it's "5 times as much work."
    1
  13.  @daedalus547  This user @cheako91155 is just trolling. He's probably ChatGPT to write slop in the comments.
    1
  14.  @PetraleiTanyl  Let's test your ChatGPT skills, troll. Name the 5 projects.
    1
  15.  @PetraleiTanyl  Did you forget to type in "name the 5 projects" to your ChatGPT prompt?
    1
  16. AI bros will respond to this with "but it always improooooooving, tomorrow the stochastic parrots might finally be able to do math, you don't know what's gonna happen in the future!"
    1
  17.  @PetraleiTanyl  Okay, you're just a troll, I get it now.
    1
  18.  @PetraleiTanyl  Your ChatGPT is spewing slop again, you might want to redo that prompt.
    1
  19.  @pavel9652  Why are you coming up with fantasy numbers, when the video shows 1/3 being bogus reports?
    1
  20.  @uiopuiop3472  Why should I have to perform a programming assignment to motivate you to fix the bad code you wrote? All those bad ideas you guys are giving, serve to punish both the scammers and the legitimate contributors.
    1
  21.  @tiranito3715  Quantum Computing, just like Cold Fusion, is useful if real. And both lack published results that they actually work.
    1
  22. Open Source software was never ever about freedom. Did you get some AI to write that nonsensical comment?
    1
  23. A PoC requires much more specialized knowledge. Might as well request the submitter to provide their PhD thesis int Cyber Security too. Just because the submitter lacks the knowledge, skill or time to produce a PoC, that doesn't mean more motivated and dedicated malicious actors won't.
    1
  24.  @pavel9652  The report on buffer overflow made sense. You have no C coding skills, that's why you didn't understand it. It was only caught by the cURL dev because he looked at the source code, and found there's already a boundary check, that prevents buffer overflows.
    1