No, people who want something approaching chatgpt but local want to run at least deepseek V3 32B.

Qwen at least fares much worse for my usage as do deepseek V3 under 32B.

The Rust code isn’t closed source, but I’d strongly prefer a coreutils replacement to use GPL over MIT as well.

Already fixed, in software that’s existed for years and is used by millions. But Oh no, memory issues, let’s rewrite that in <language of the month>! will surely result in a better outcome.

Rsync is great software, but the C language fates it to keep having memory issues in spite of its skilled developers.

Preventing a bug from being possible > fixing a bug.

I fear moving away from GPL that moving to Rust seems to bring, but Rust does fix real memory issues.

Take the recent rsync vulnerabilities for example.

https://www.cyberciti.biz/linux-news/cve-2024-12084-rsyn-security-urgent-update-needed-on-unix-bsd-systems/#more-2215

At least this one in a Rust implementation of rsync would have very likely been avoided:

CVE-2024-12085 – A flaw was found in the rsync daemon which could be triggered when rsync compares file checksums. This flaw allows an attacker to manipulate the checksum length (s2length) to cause a comparison between a checksum and uninitialized memory and leak one byte of uninitialized stack data at a time. Info Leak via uninitialized Stack contents defeats ASLR.

I would love this news if it didn’t move away from the GPL.

Mass move to MIT is just empowering enshittification by greedy companies.

If Mozilla wants to limit their use of my input, why the do I need to give them a full, non-exclusive license?

If Mozilla wants to limit their use of my input, why the do I need to give them a full, non-exclusive license?

Signal very recently made syncing between devices possible:

https://signal.org/blog/a-synchronized-start-for-linked-devices/

People are definitely born cops and it’s an immutable characteristic.

Most social justice aware leftists don’t deny class issues, lol wut

Whats the story on GrapheneOS drama?

Not for KDE which aims to be good for beginners.

I’ll have to come up with some examples and write something more detailed I think to explore this.

Until NixOS I was very in favor of language specific package managers and things like flatpak.

You see the conclusion of that article is that flatpaks are not repeoducible after presenting solutions to make it reproducible right?

If you care about your software being stable and secure, you should care about how easy the programming language used makes and encourages that.

People aren’t robots and make mistakes often.

The issue was closed, but a draft PR was linked… potato:

https://github.com/zed-industries/zed/pull/14034

I did this before being in emacs made it so convenient to avoid, but got bit randomly by different versions or gnu vs BSD.

woman in emacs.

I also find info pages much nicer to use after an adjustment period given I grew up on vim and man.

So they should say that it is written with performance in mind. I don’t care how you achieved that. rust, c++, assembly, whatever.

I care because performant and secure C++ is much harder to achieve while rust “shepherds” you towards it.

See https://nibblestew.blogspot.com/2020/03/its-not-what-programming-languages-do.html

I care because I know the values of those programmers in a narrow scope and won’t be as annoyed when I inevitably have to go debug the rust code instead of C.

However, that values statement was challenged by automatic binary downloads without user confirmation.

Luckily the fix is already in progress, but its concerning it was ever implemented.