I have a desktop which has / had a similar problem.

Originally I built it with a g-series Ryzen which has integrated Radeon Vega graphics. Upgraded to a 3060 and wanted to run Linux for gaming instead of windows.

I couldn’t get a distro to reliably use my graphics card without the issues you describe. Stuttering, crashing, generally unusable.

Garuda was the answer (to be fair I’d try Bazzite too but I just didn’t get there as Garuda worked). In fact, it worked out of the box for me and I enjoyed it so much I made it my work OS.

I like the GUI utilities they’ve made for front-ending a bunch of Arch CLI utilities and I’ve been saved by BTRFS snapshots more than once.

What is success here? The few founders and VC get filthy rich as the larger population dumps their money into Discord stock while the users and teams with limited foresight, who’ve moved their communities to discord, suffer?

I mean yeah I guess that’s the success Cory Doctorow warns us about again and again.

But that’s not my definition of success.

For context I’ve been on the receiving end of an IPO and the founders and investors made out like bandits while a fair number of employees were stuck holding the bags thanks to lock-ins, dilution and over priced shares.

So maybe we’re kinda staring at two sides of the same coin. Because yeah, you’re not misrepresentin my point.

But wait there’s a deeper point I’ve been trying to make.

You’re right that I am also saying it’s all bullshit - even when it’s “right”. And the fact we’d consider artificially generated, completely made up text libellous indicates to me that we (as a larger society) have failed to understand how these tools work. If anyone takes what they say to be factual they are mistaken.

If our feelings are hurt because a “make shit up machine” makes shit up… well we’re holding the phone wrong.

My point is that we’ve been led to believe they are something more concrete, more exact, more stable, much more factual than they are — and that is worth challenging and holding these companies to account for. i hope cases like these are a forcing function for that.

That’s it. Hopefully my PoV is clearer (not saying it’s right).

Ok hear me out: the output is all made up. In that context everything is acceptable as it’s just a reflection of the whole of the inputs.

Again, I think this stems from a misunderstanding of these systems. They’re not like a search engine (though, again, the companies would like you to believe that).

We can find the output offensive, off putting, gross , etc. but there is no real right and wrong with LLMs the way they are now. There is only statistical probability that a) we’ll understand the output and b) it approximates some currently held truth.

Put another way; LLMs convincingly imitate language - and therefore also convincing imitate facts. But it’s all facsimile.

Really?

I read your reply as saying the output is (can be) libellous - which it cannot be because it is not based on a dataset which resolves to anything absolute.

Maybe we’re just missing each other - struggling to parse each others’ output. ;)

Surely you jest because it’s so clearly not if you understand how LLMs work (at the core it’s a statistic model - and therefore all approximation to a varying degree).

But great can come out of this case if it gets far enough.

Imagine the ilk of OpenAI, Google, Anthropic, XAI, etc. being forced to admit that an LLM can’t actually do anything but generate approximations of language. That these models (again LLMs in particular) produce approximations of language that are so good they’re often indistinguishable from the versions our brains approximate.

But at the core they cannot produce facts because the way they are made includes artificially injected randomness layered on-top of mathematically encoded values that merely get expressed as tiny pieces of language (tokens) - ones that happen to be close to each other in a massively multidimensional vector space.

TLDR - they’d be forced to admit the emperor has no clothes and that’s a win for everyone (except maybe this one guy).

Also it’s worth noting I use LLMs for work almost daily and have studied them quite a bit. I’m not a hater on the tech. Only the capitalists trying to force it down everyone’s throat in such a way that we blindly adopt it for everything.

It’s all hallucinations.

Some (many) just happen to be very close to factual.

It’s sad to see that the marketing of these tools has been so effective that few realize how they work and what they do.

While there are many reasons to dislike (or outright avoid) Apple - if you purchase music from them, it’s DRM-free and useable anywhere.

I believe they were one of the first official channels to do this.

Still, hadn’t heard of Quobuz and will check them out!

You’ve described Ghost. Subscriptions for content are a first class citizen.

It’s primary a writing platform with built-in monetization options and the ability to self host. We switched to it from Substack. It’s been fantastic to use and operate. Super slick.

But that’s exactly my point - we here in this bubble prefer Jellyfin - but it’s not ready for mass adoption. Even plex is a drop in the bucket.

I see Jellyfin suggested as an alternative to Plex here. I hope it is one day.

At the moment it’s nowhere close.

I’ve been running Jellyfin side-by-side Plex for two years and it’s still not a viable replacement for anyone but me. Parents, my partner, none of the possible solutions for them come anywhere near close to the usability of Plex and its ecosystem of apps for various devices.

That will likely change because plex is getting worse every day and folks can contribute their own solutions to the playback issues. With plex it’s more noise, more useless features. So one gets better (Jellyfin) and one gets worse (Plex).

But at the moment it really isn’t close for most folks who are familiar with the slickness of commercial apps.

Even from the administrative side, Jellyfin takes massively more system resources and it doesn’t reliably work with all my files.

Again, Jellyfin will get there it’s just not a drop in replacement for most folks yet.

And for context I started my DIY streaming / hosting life with a first gen Apple TV (pretty much a Mac mini with component video outs) that eventually got XBMC and then Boxee installed on it. I even have the forksaken Boxee box.

I’m not defending them because it’s a shit move, but they clearly understand the most valuable feature is not their ad-ridden free content but rather the original value prop of the service to make it easy to share personal media with small groups of friends.

We use NGINX’s 444 on every LLM crawler we see.

Caddy has a similar “close connection” option called “abort” as part of the static response.

HAProxy has the “silent-drop” option which also closes the TCP connection silently.

I’ve found crawling attempts end more quickly using this option - especially attacks - but my sample size is relatively small.

Edit: we do this because too often we’ve seen them ignore robots.txt. They believe all data is theirs. I do not.

Ran WireGuard on a Pi1 and it was fine for two users. Albeit WireGuard was the ONLY thing running aside from a Gitlab Runner.

A 4b should be more than enough for many use cases except things that cause torrents of packets - but even then YMMV. It really depends on the workload.

One bit of advice: if you can, use a storage device other than the micro-sd slot for the 4B. Again YMMV.

I think that depends on what you’re doing. I find Claude miles ahead of the pack in practical, but fairly nuanced coding issues - particularly in use as a paired programmer with Strongly Typed FP patterns.

It’s almost as if it’s better in real-world situations than artificial benchmarks.

And their new CLI client is pretty decent - it seems to really take advantage of the hybrid CoT/standard auto-switching model advantage Claude now has with this week’s update.

I don’t use it often anymore but when I reach for a model first for coding - it’s Claude. It’s the most likely to be able to grasp the core architectural patterns in a codebase (like a consistent monadic structure for error handling or consistently well-defined architectural layers).

I just recently cancelled my one month trial of Gemini - it was pretty useless; easy to get stuck in a dumb loop even with project files as context.

And GPT-4/o1/o3 seems to really suck at being prescriptive - often providing walls of multiple solutions that all somehow narrowly miss the plot - even with tons of context.

That said Claude sucks - SUCKS - at statistics - being completely unreliable where GPT-4 is often pretty good and provides code (Python) for verification.

Recently switched from VsCodium to neovim - but still use Codium for some specific tasks.

My setup customization focuses around Telescope, Treesitter, Trouble & Blink.

But the advice I got was to start with vim keybindings in VSCode. I used those for six weeks until I got the hang of the basics and it had gone from frustrating to somewhat second nature.

Then I made the move.

I still use Codium for Terraform work (I have struggled to get the Terraform LS working well in neovim and I don’t use it often enough to warrant the effort) and as a GUI git client - I like the ability to add a single line from multiple files and I haven’t looked up how to do it any other way - I’ve got other stuff to do and it’s not slowing me down.

But I grew to hate Codium / VS code tabs in larger codebases. I was spending so much time looking for open tabs ( I realise this is a me problem). While neovim has tabs, it’s much more controlled and I typically use them very differently and very sparingly.

If I need to look up a data structure I just call it up temporarily with Telescope via a find files call or a live grep call (both setup to only use my project directory by default), take a peak, and move on.

The thing is - security risks are going to exist anywhere you install plugins you haven’t audited the code for. Unless you work in an IDE where there’s a company guaranteeing all plugins - there are always going to be risks.

I’d argue that VSCode, while a bigger target, has both a large user base and Microsoft’s security team going for it. I don’t see the theme being compromised as much as problem because it got solved and also prompted some serious security review of many marketplace plugins. Not ideal, but not terrible.

As per my other comment - the algorithm is only part of it.

A big aspect however is the slickness and ease-of-onboarding for mega-Corp apps. It’s a thing that would relatively easy to begin work on.

I’ve seen first hand the amount of time and money even growth-stage startups spend on onboarding and have lots of first-hand reports from peers at the big girls - it’s a critical part of success. Make it easy to get started and easy to stay using.

It’s missing from most fediverse experiences. Pixelfed being a serious contender for an on-boarding rethink.

“time-to-value” - we want that as low as possible.

That’s a shame irrespective of the drama. Asahi is surprisingly good. Installation is (relatively) straightforward.

I’ve got it on my primary laptop. I don’t use it frequently because battery life is poor compared to MacOS and I can’t use an external display but it’s an impressive achievement and I’m sure it will only get better. I haven’t used fedora in 20 years but it’s slick and easy and most of it just works. It looks just like my Linux workstation desktop.

Agreed.

For us the mitigation is to do a little monitoring with alerts set to start casually at 29 days out and enter critical 13 days out (out from expiry).