- cross-posted to:
- linux@programming.dev
- cross-posted to:
- linux@programming.dev
- Rust will save Linux from C’s inherent security weaknesses.
- Linux, faced with a flood of AI-discovered security problems, could use the help.
- Going forward, more and more Linux code will be written in Rust.
You must log in or # to comment.
A better title than previous, not sure I agree though. Not that I don’t think it wouldn’t improve memory safety. It is that LLMs can keep finding bugs on even a clean code base practically forever. Yeah they may not be critical bugs like those which are technically correct but makes you roll your eyes like you would on an intern that is trying a bit too much to impress. Infact if you incentivise the LLM enough, it will likely produce fake bugs. So better memory safety won’t stop some people dumping pages of LLM bug reports.
The only thing that will save Linux from AI is when programmers get their heads out of their asses, stop subscribing to scifi slop that AI is about take over and start asserting to everyone else how necessary human input is in their craft by forming unions.
The problem with programmers is they are quick to think their understanding of computers makes them understand everything and it created a massive blindspot of hubris big enough to allow the ruling class to destroy the potential quality of life that used to come from working as a programmer right in front of programmers eyes. I would be tempted to call it sad if programmers didn’t tend to be so condescending about believing their knowledge is a universal shortcut to understanding everything else.
Programming is an industry dominated by pushovers who aren’t willing to fight for anything that cannot be understood in the terms of automation, scaling or endless growth of technology. There is a willing blindness here and the rest of the world is getting really tired of it.
Fight, stop immersing yourself in stories about how AI will become sentient blah blah blah, outside of your bubble very few people actually believe this nonsense and it just blinds you to your own dehumanization that is happening right in front of your eyes for entirely human reasons that have NOTHING to do with technology working or not working and EVERYTHING to do with a failure of humans to organize and preserve the professionalism of their craft because they were convinced of delusions by the people ruling over them.
That problem isn’t even specific to computer science, it’s specific to most engineering fields. Programmers are just extremely good at automating, so they automate themselves out of existence.
At least in Germany, people with higher-education on-average also have more political awareness - but the more engineering their field contains, the more that awareness jumps out of the window (what I mean by that is that of all the people with higher-education, engineers have the highest rate of politically right-leaning people, which in my world-view equates to being incompetent).
Which imo comes from another problem: Computer Science had a huge boom, and now we ended up with a lot of businessmen who can hussle 70h/week to get their degree in 4 semesters, so that they can start their path to their first million. Basically, computer science got invaded by capitalists and the nerds just went to another corner in hopes not to be bothered. Not that they could have done much about it, there are way more capitalists than there are computer nerds.
It’s sci-fi that AI is going to take over, but we also need unions to prevent it? Hmm.
We need unions to prevent humans from the ruling class from abusing you because they have thoroughly convinced you and others like you incorrectly that robots can actually replace entry level versions of yourself.
It might appear that they can in the short term, but that is why we refer to AI as a bubble.
Wake up.
I don’t see why unions are necessary for that. If it turns out they are wrong and AI can’t reduce the need for humans they will quickly realise.
In any case, they’re pretty clearly just using AI as an easy excuse for layoffs they want to do anyway.
If it turns out they are wrong and AI can’t reduce the need for humans they will quickly realise.
No, they have not realized that has already happened.
- He described new C “guards” and scoped locks inspired by Rust
In other words, the improved safety will sometimes come from Rust code, and sometimes from C code. The important point being that safer practices are becoming more common now that Rust has called attention to them.
now that Rust has called attention to them.
More like now that Rust is embarrassing C.
A common pattern. GCC’s error messages only got good once Clang embarrassed it.
improved safety will ALWAYS come from rust code and SOMETIMES come from C code with AI help or these new magical unicorn C coders that have suddenly learned how to manage memory (etc etc) now that rust has appeared
yeah right
No worries, the amount of new slop code with hidden bugs is entering the ring to try to keep the balance and to ensure the overall security doesn’t improve by too much: https://www.neowin.net/news/linus-torvalds-declares-massive-ai-fueled-code-surges-as-the-new-normal-for-linux/
These are security patches for bugs found with AI tools - not AI-generated kernel code.
Phoronix commentors are going to be fuming at this news. And C programmers who don’t want to learn Rust won’t like it much either.
The best thing is that people are 100% going to use AI to port stuff from C to Rust and fix vulenrabilities. There is so much people can be mad about, we might just see some kernel forks
