“Falsehood flies, and truth comes limping after it, so that when men come to be undeceived, it is too late; the jest is over, and the tale hath had its effect: […] like a physician, who hath found out an infallible medicine, after the patient is dead.” —Jonathan Swift
They work as intended but aren’t modern.
The lack of a dogshit “modern” UI makes them even better. I don’t need a loading screen morphing into different rounded polygons like I’m a toddler having keys waggled in front of my face for three seconds of loading.
Yeah, you’re right. I was thinking of a combination of Rust’s debug mode, the fact it won’t compile if it detects a static under/overflow, and Swift’s default producing a runtime error on integer under/overflow. Corrected; I apologize.
Chris Sawyer really went off the deep end after programming RCT in assembly.
> “sounds like someone trying to push their preferred language onto others.”
> didn’t even name Rust in the first comment
I literally only talked about Rust when someone else mentioned it, because I’m not specifically talking about Rust.
Whereas yes, Rust is relevant to the Linux kernel for obvious reasons (matching C’s speed and being easily interoperable with it), I said “a modern language” because I’m talking generally about the use of old, unsafe languages and major exploits surrounding problems that modern languages fix cropping up every 5 seconds.
I’m not going to police my own tone about what are clearly negligent engineering practices. SWE should be treated with the same level of scrutiny as other engineering fields when its poor, outdated practices are constantly causing accelerating worldwide damage.
“Quit complaining about fixable systemic issues unless you’re able to personally go fix them yourself. Bridges collapsing due to negligent engineering practices? Go survey and fix them yourself. I am very smart.”
I mean yes, I do, given safe Rust inherently fixes seemingly every exploit listed in this article:
I feel like it’s the 1960s and I’m being told “we get it, you like seatbelts”.
Edit: Sorry, re: the first one. I was misremembering based on a mix of a few things (see below). It does this in debug mode (which in itself is good to have), and it won’t compile if it can detect an overflow statically, but it’s not checking over/underflow at runtime by default. It can do this if you want it to, but that’s an inherent speed penalty from C. I was thinking of Swift, which does automatically detect integer under/overflow at runtime unless you explicitly tell it not to on a per-operator basis.
In the spirit of not making you “trust me bro” since I was so wrong about the integer under/overflow one, here’s Rust’s overview of how safe Rust guarantees against data races.
SMB Direct remote code execution vulnerability due to integer underflow.
It’s a day ending in ‘y’ and we’ve found another RCE that wouldn’t exist in a modern language designed to catch things like integer under/overflow.
where linux
Well not really, given nothing they said suggested they didn’t read your 6-point excuse to comment on material you haven’t read (I did); you just asserted they didn’t.
“I didn’t read the attempts to explain this”
You’re one to talk. lmao
Removed by mod
Does the CPU have a serial number that’s visible to the system?
“I didn’t read the article that attempts to explain this and explicitly states CPUs don’t have a unique ID.”
Pictured: Lauren Boebert’s office.
If you strip out the encoder, then you’re getting closer.
That’s flat-out not true. BERTs, for example, are LLMs and are encoder-only, just like GPTs are decoder-only. Whisper is trained on a large corpus of text and does NLP tasks. That’s all it is to be an LLM.
A transformer model being encoder-only, encoder-decoder, or decoder-only has fuck-all to do with whether it’s an LLM. In fact, it being a transformer model at all isn’t required for it to be an LLM.
Whisper is not an llm, just because it is trained on big datasets does not make it an LLM.
The training dataset consists of 680,000 hours of labeled audio-transcript pairs sourced from the internet using semi-supervised learning.
Brother, I don’t think you know what a large language model is if you think it being trained on a human lifespan’s worth of audio transcripts isn’t enough.
Its job is NLP. It is an LLM, if not as general-use as something like ChatGPT.
So you’ve moved the goalposts from “that’s not just participation, it’s [things that participation entails]” to “well this doesn’t apply unless you need it to exist”*. GN needs a major platform like YouTube to actually reach an audience.
* By the way, the comic is illustrating a man with a car that’s seemingly from the 1920s or 30s, so there goes that tenuous argument.
contributing to, and directly benefitting from the very thing they’re complaining about.
You do understand what “participation” means, right? You understand that the original comic is specifically talking about contributing to and benefitting from the very thing they’re complaining about, right?
I agree with your other point that GN should offer more alternatives to watch their videos, but this one is creating a distinction without a difference.
“… appears to work on Linux with Proton 11 as captured in my screenshot above”
The emphasis in that point was the word “finally”, not “playable state”. As in “it’s finally working after x amount of time” like you’d expect with a project that takes any appreciable time to get working.
I could never have put it so succinctly; thank you.