Depends on what you define as “politics” but aside from “everything is politics”, my Lemmy feed is mostly tech stuff. Just subscribe to communities that fit your interests. That being said, many interests will be under-represented on Lemmy as I think the user base skews either technical or political or both.

I don’t agree. LLMs are by design probabilistic. Chainsaws aren’t designed to be probabilistic, and any functionality that is probabilistic (aside from philosophical questions about what it is possible to be certain about, YKWIM) is aimed to be minimised. You’re supposed to be able to give the same model the same prompt twice and get two different answers. You’re not meant to be able to use a chainsaw the same way on the same object and have it cut significantly differently. You’re inherently leaving much more to chance by using LLMs to generate code, and creating more work for yourself as you have to review LLM code, which is generally lower quality than human-written code.

Not comparable at all. Power tools work deterministically. A powered chainsaw is not going to have a 0.1% chance of chopping a completely different tree on the other side of the forest. Of course accidents happen; your hand can slip. But a proper comparison would be if you got a computer to look at a large number of powered chainsaws and then generate its own in CAD based on what it’s seen, and then you use that generated power tool. Which, for something as potentially dangerous as a powered chainsaw, you most likely wouldn’t want to do, and would want to have careful human oversight over every part of design.

I also got LLM vibes. However, some humans do just genuinely write like that. It’s particularly an issue with ESL speakers getting caught in false positives, although this author seems to be a white Australian guy who is probably a native English speaker.

I suppose if you were really bothered you could go back and look at his writing before the dawn of the vibes and see if his writing style is about the same. I don’t care enough to check.

People have their preferences for UI and UX. I use Aerc because I like modal editing (ie being able to write my emails in vim) and keyboard nav. Using a desktop email client rather than webmail client from a provider gives me that freedom.

Besides, I don’t actually have a webmail client I can use lol. I host my own email and host the IMAP server but I don’t host a web interface.

If you haven’t already, I recommend Watchtower (nickfedor fork—the original is unmaintained) which automatically pulls updates to Docker containers and restarts them. Make sure to track latest, although for security updates, these should be backported to any supported versions so it’s fine to track an older supported version too.

Notesnook notebook with whatever info I need to be able to administrate the system. e.g. what different ports are used for and why the firewall policies are what they are, sometimes write-ups after a troubleshooting session, etc.

The Notesnook instance is self-hosted too, but if the server goes down, the notebook will still be available locally.

I don’t see where I said any of the words you just quoted. Impressive if Rust can suck a dick I don’t have though, I’ll give them that.

You can embed Assembly in Rust. A lot of low-level Rust projects embed Assembly.

A competing Forgejo instance

Over the weekend.

I don’t see small talk or whatever as a sign of a lack of interest in someone. Sometimes there’s nothing that deep going on to talk about.

The relevance for me personally is whether or not they can be useful for programming, and if they’re accessible to run locally. I’m not interested in feeding my data to a datacentre. My AMD GPU also doesn’t support ROCm so LLMs run slow as fuck for me. So, generally, I avoid them.

LLMs consistently produce lower quality, less correct, and less secure code than humans. However, they do seem to be getting better. I might be open to using them to generate unit tests if only they would run faster on my PC. I tried deepseek, llama3.1, and codellama; all take like an hour+ to answer a programming question given that they are just using my CPU, as my GPU doesn’t support ROCm. So really not feasible for anything.

Depending on what you count as AI, I think some of the long-existing predictive ML like autosuggestions based on learning your input patterns are fine and helpful. And maybe if I get a supported GPU I won’t mind using local LLMs for some things. But generally I’m not dying to use them. I can do things myself.

Thunderbird is a K-9 reskin. They’re the same app. People pick based on what icon and app name they want I guess.

I just cook normal meals and put them in tupperware and take them to work. Doesn’t work if you don’t have a microwave, but if you don’t have a microwave, I’m not sure what you could do other than sandwiches. So whatever you like to eat.

How’s the firmware support/availability? For things like graphics tablets, graphics drivers, etc?

I don’t think OpenBSD has binary compat with Linux but most Linux software should just need a recompile for BSDs—I’m discouraged from porting given that when it’s not a simple recompile I’d have much less idea what to do.

The reasons people use Linux are for qualities other than the ones affected by AI use. AI use has implications for code quality, correctness, and security. But none of those are why people use Linux. People use Linux over BSD or other Unixes because Linux supports the most hardware, has the biggest software ecosystem, and being a monolithic kernel is much easier to get up and running with lots of hardware without needing to install separate drivers. Those qualities still need to be addressed by BSDs or whatever alternatives before people will start migrating from Linux.

I say this as someone who regularly uses and enjoys an OpenBSD machine. I couldn’t use it as my main machine because it just doesn’t have the same software availability and plug-and-use hardware support as Linux. Porting software to a new target is not a trivial task for most users. I package a few things for the AUR and that’s much easier as the software already supports x86_64 Linux; I just have to write a script to install it. I think OpenBSD is a nice OS but I highly doubt Linux users will migrate any time soon. Think about how many people were clinging onto X11 because Wayland didn’t support their super specific workflow. And a migration to an entirely different OS would be worse.

Sounds good. kwm and zrwm are on my shortlist. I’ve also seen tinyrwm and I’m heartened by how short the implementations are. I’ve been thinking about perhaps writing my own WM. I might take a read of tinyrwm to get an idea of what rolling my own might be like and give it a go.

Looks good! How is kwm? Especially curious compared to river-classic, and if you’ve tried any other River WMs. I’m still on river-classic and hoping to move to 0.4 soon, but I need to decide on a WM first. Probably want something close to river-classic but some more programmability/extensibility would be nice.

You’re in the US. Mexico is next door and speaks Spanish. A much higher proportion of Spanish speakers in the US are from Mexico than from Spain, which is across the ocean. If you were in France and identified as a Spanish speaker, many more people would think you were from Spain.

I know about Tailscale. I don’t use it because I want my VPS to be exposed to the internet; some of my services are supposed to be public. And those that aren’t, have their own authentication systems that are adequately secure for their purposes. I just don’t need Tailscale so I’ve not bothered with the setup.