• 0 Posts
  • 30 Comments
Joined 3 years ago
cake
Cake day: August 2nd, 2023

help-circle
  • Because the goal is to get people to learn/think about something. We don’t care what you use as long as you retain knowledge taught in the course. If what helps you learn is LLMs, then go for it.

    Problem right now is there is a significant amount of people that are using these tools to do the thinking for them. And this is when Office Hours, Homework feedback, Email (I guarantee all students emails are responded to within 24hrs. Most are handled within 30 minutes) are all available and paid for (by tuition). I am even happy to schedule one-on-ones if privacy is a concern, but none of this is being utilized.



  • We are allowing LLMs for all of our homeworks. As long as you can solve the problems in the indicated way with a reasonable answer.

    In case you are not sure about the “indicated way”, there are practice questions with detailed step-by-step solutions for each hw problem that you just have to change the numbers/equations a bit and you’ll get points.

    What we’ve noticed is that the year-after-year averages are significantly higher, especially this year. However, students are bringing in details that we explicitly didn’t go over in lecture and putting that on the homework (e.g. Delayed branching in Computer Architecture, because it’s a random quirk of MIPS that even assembly programmers don’t have to deal with). None of these details are ever mentioned in lecture or the practice homeworks (in a few cases, they are mentioned with the explicit wording “do not worry about this now”)

    We can only assume people are copying the homework into LLMs and copying the results straight down. The latest exam had a question where students were asked to analyze a specific chunk of assembly code to deduce certain properties about it. Approximately 20-30% of the students didn’t know the FORMAT to answer it, despite it literally being item 1 on last week’s homework.

    And when I say format, I don’t mean exactly “you must write these exact words or you lose points”. It’s literally just point out “line A and B have this property X because of attribute Y”. Just including ABXY as shown in the practice homework is enough. But apparently people are too lazy to read a 10 bullet point answer…




    1. If your assumption is that X509 is trash, does that mean you hold the same amount of distrust to TLS?
    2. How do you propose the scaling of key management? Do you have a reasonable alternative to users blindly trusting every single key they come across?
    3. Back to my original question: what prevents a VSCode extension from stealing a private signing key (as opposed to an API key) and causing the same issues described here?

  • And how would apt help in this particular case? A supply chain attack can happen with any particular package manager. In this case, the compromised package was detected and mitigated within 93 minutes, affecting a total of ~330 users. Which is a lot better than how a lot of distros handled the xz breach last year.

    All reasonably secure package managers (and https) operate on a chain of trust. There is little that can be done if that chain of trust is broken.

    Based on this the cause was a malicious VSCode extension that stole credentials that were later used to trigger a deployment CI/CD pipeline. If there’s anything to learn from this, it’s probably to not use VSCode.


    1. All NICs already work off of DMA to access/copy packets into/from memory. Yes, even your $10 ones. So “would need DMA to stand a chance” doesn’t have any technical meaning other than putting a bunch of words together.

    2. The bottleneck for TCP is sequence number processing, which must be done on a single core (for each flow) and cannot be parallelized. You also cannot offload sequence number processing without making major sacrifices that result in corrupted data in several edge cases (see TCP chimney offload, which cannot handle the required TCP extensions needed to run TCP at 1Gbps). So no, “more offloading” is easy to say but not feasible.

    3. Who needs it: data centers trying to scale legacy software, or dealing with multi region data replication (rocev2 is terrible for long distance links). But no, no home user would need it












  • is definitely on every normal user’s mind at all times.

    That was the context. The problem wa connecting to Wireshark, which more and more people are doing thanks to general awareness of VPNs.

    and last but not least people like you going ‘hmm yes but akshually’ in sort-of-defense-but-not-really of the deliberately malicious and billion-dollar company.

    Huh? Where in my post did I defend MS? I was there when Balmer and crew decided to sue anyone with a pulse for using Linux. I was there when the Cathedral acquired the Bazaar (and I deleted my account for it), and I am still here using Linux and BSD for every single machine I own with the exception of one. I still hold a grudge against Mr. Bill “Jump on a roller to show how fit you are” Gates, and I refuse to purchase anything from their game catalog since 2011. Hopefully with this context, you would no longer misconstrue my point as “defending Microsoft”.

    Alas, normal users care about neither. The computer is just a tool that allows them to do work which allows them to put food on the table. If your assistance is just “boo hoo use Linux”. That’s not productive to them nor us. Joe Shmoe isn’t gonna care that you should save your documents as ODT instead of DOCX. They need that document working with no hassle NOW.

    Look at it this way: A normal dude with bad hair and questionable social intelligence isn’t getting up in the morning and deciding to fuck with a million or more users by making their computers unbootable. There is only good intentions.

    Case in video game modding: 1. GShade, where the developer deliberately made people’s game segfault if compiled on their own after an update 2. MultiMC, where the developer personally threatened to sue for trademark violation after packaging the application for a Linux distro 3. Bukkit, where one dev decided to DMCA and take down all instances of the project.

    Outside of video games: the entire university of Maryland, which attempt to inject backdoors into the Linux kernel that was not caught until they published a paper.

    Also, for the “good dudes part”: regardless of intentions, if the damage is done, the harm is done. If a suitcase falls from an airplane and kills me tomorrow, I wouldn’t care whether it was intentional or not. I would be dead.

    Going back to the original blog post: there is both a user problem and a technical problem here. The technical problem “could” be fixed by switching to Linux (assuming systemd or gnome doesn’t get to it first), but the user problem can’t. Calling out anyone who points out the user problem as “corpo drone” isn’t going to make it go away.