Komunitas
mander.xyz
[Sorry for the double reply] The “numbers” template style would be considerably more useful if the palette was itself numbered. At least, while using that style. I’ve seen a lot of people struggling to find the template. I think that it deserves its own button. The dark mode is amazing. Seriously, I want it for the next years. I don’t think that it needs such a huge button though, when a simple half-black half-white sun icon would do the trick. On desktop the palette has an awkward shape, as a narrow 32x1 strip that you need to roll back and forth if the window isn’t maximised (fairly often, since people were doing other stuff while placing pixels). It would be great if it was a 4x8 somewhere at the right. A lot of people (incl. myself) were struggling to tell a few colours apart. Mostly dark grey vs. black vs. navy blue and dark chocolate vs. maroon. So it might be sensible to tweak the palette itself for the next years. But overall their hue distribution was really good, in no moment I thought “damn, I need more colours”.
Komunitas
lemmygrad.ml
I’ve contemplated this outcome, however revolutions are built and I’m not sure Russia’s internal conditions are such which predicate a resurgence of socialist ideas. Russia is (fortunately and unfortunately) proving capable of fighting off all of NATO in it’s current state. And if it wins I’ve no doubt the question would be resolved the other way, in the unquestionability of the new Russian bourgeois state…until crisis of course. But then raises the concern. Because a revolution must be built, and is not in preparation now, in the unlikely outcome of Russia falling to NATO, where would this revolution come from to re-resolve the situation into a favorable one? They would be balkanized, and as efficient as the Russian bourgeois have been at keeping our class down there, the US are masters of oppression. And so lies the situation at hand…fucked if they do, fucked if they don’t.
Komunitas
lemm.ee
Before using Rust I was using C++ for most projects and while it is a really powerful language there were some big problems: no standard build system, most projects use cmake or meson and vendor dependencies with the projects. These build systems were really hard to learn (especially cmake, meson is easier). There are package managers these days such as conan and vcpkg but there is not really one standard way to build programs like in rust. error messages were really hard to understand, especially when the project uses templates it felt like 3 languages in one, projects written before c++11 differ greatly from c++11 and up some of the new language features have really weird syntax, for example lambdas some people say that rust is hard, but modern c++ is considerably harder to learn, just look at the list of modern c++ features: https://github.com/AnthonyCalandra/modern-cpp-features, you have to know the different pointer types (unique_pointer, shared_pointer etc.), templates, rvalue references and move semantic, exceptions, constexpressions and the list goes on
Komunitas
sh.itjust.works
Dude, you can make a barcode of anything with literally a Microsoft word font, and you can scan with a phone. They can keep the same item IDs for heaven’s sake - I work for a manufacturer, and we wanted to add barcodes literally just for our own internal inventory counting (no on site POSs, less structured warehouse given nature of work). Our internal item codes have letters and symbols in them, like 25WIDGET-01. I was able to tweak the label printing template to add an additional line of the same item code in barcode font with the right open and close symbols added (/25WIDGET-01/) and then when we do inventory, we use a really cheap app that let me put together a scanning form that lets the user scan the code, converts it back to the string, and then lets them input a quantity, and sends it back to a master excel document without needing to manually transcribe written inputs or risking employee typos or visually misidentifying the item as 26WIDGET-01 instead. We also no longer have to have one group of counters scour the entire disjointed warehouse looking for every single location one of “their” items might be and hope they found all of them while other groups chaotically do the same - your group is assigned to scan EVERYTHING in this section. If you find something without a label you report it. No more crazy margin math, no more “oops I didn’t think I’d find boxes of our hottest seller in the storage trailer”, no more 30 paper copies of a form. Our total organizational cost to use this system in 2 countries and half a dozen locations was about 12 man hours of my completely underpaid time and about $30 per location per inventory in app fees on existing company phones. We made no changes to our ERP. It reduced inventory time, recounts, and mistakes by ~50%. Employees LOVED not having to waste time trying to exactly copy longass random strings for our weirder items or the “gotcha” items. Our vendors don’t barcode for us either - we do it on receiving, because we already have to label the items with our internal code anyway. And we’re doing a TON of custom manufacturing - often items that literally may never be sold under that code again - with large volume orders, we’re not asking cashiers to manually enter everything for stocked items sold repeatedly at small volume. We’re literally ONLY using this for inventory and get no other benefits. Their excuses are such hot bullshit it’s crazy. Barcodes also literally have nothing to do with discounts - as above, they are literally just a way of storing a string visually. Not exaggerating to call them a font meant to be read by computers. If you are storing pricing in a computer anywhere, and not making cashiers enter THAT manually as well, then the only difference between barcodes and not is that with a barcode they scan the item code into the computer and without they manually type it? And if the only place they store pricing is on the item sticker, that is a system so fucking vulnerable to fraud both internal and external it’s CRAZY. For an organization of their size??? No way.
Komunitas
lemmy.blahaj.zone
So it is not just “so stupid design” that “we don’t even feel devices are 10x faster than 15 years ago”, but deliberate design to use the hardware capabilities for the sake of other people’s computers. To my limited understanding, better use of hardware requires both some level of standardization of hardware and better optimized software, with lower level programming languages if possible. Vibe coded Electron apps are never going to be as well optimized as something made to run a particular chipset. But how do we accomplish that when most programming is being done on high level programming languages? It seems that the industry has prioritized human readable code and improved UX over hardware efficiency, which at a surface level isn’t the worst trade-off. I ultimately agree with this vision, and it’s a serious problem I’ve contemplated as well. That said, what’s the right balance between more efficient, repairable hardware and accessible, more readable code?
Komunitas
lemmy.today
Short answer: Custom Fedora Silverblue image through uBlue’s template, because it offers a relatively mature and easy to use distro with unique features in terms of stability and security that’s (almost) unmatched within the Linux space. Long answer: ::: spoiler spoiler which distro and why do you prefer it over others? Personally, I’m very fond of atomic^[1]^ distros. What they bring onto the table in terms of stability and “It just works.”^[2]^ can’t be understated^[3]^. I’ve been running Fedora Silverblue^[4]^ for the last one and a half years and it has been excellent barring some smaller issues^[5]^. While on the other hand, the distros^[6]^ I’ve experienced in the mean time through dual-booting happened to be a mess and I eventually couldn’t continue to use them as they accumulated issues all over the place. So far, it should be pretty clear why I prefer atomic distros over traditional ones. However, why do I favor Fedora Silverblue over the other atomic distros? Well, I try to be very security-conscious. And, unsurprisingly, this has influence on my choice. In this case; Fedora is the only one (together with openSUSE) that properly supports SELinux. While AppArmor is also excellent, it’s not ideal for the container workflow atomic distros are known for; which is probs one of the reasons why openSUSE has only recently started supporting SELinux while they’ve been supporting AppArmor for a long time. Furthermore, while both Fedora’s and openSUSE^[7]^'s offerings are excellent. Fedora has been working on theirs considerably longer and therefore their atomic distros are more mature. Thus, I ended up with Fedora. Silverblue, however, wasn’t actually initially preferred over Kinoite. I started on Kinoite, which I was attracted to for how KDE Plasma was relatively similar to Windows^[8]^ and for how it allowed easy configuration out of the box. At the time, Kinoite wasn’t that polished yet. So I had to rebase^[9]^ to Silverblue and the rest has been history. There are actually atomic distros that don’t heavily rely on the container workflow to do their bidding and thus don’t necessitate the use of SELinux over AppArmor. Those distros would be NixOS and Guix. These are on my radar and I might even switch to either one of them eventually^[10]^. Heck, I’ve even installed the Nix package manager on Fedora Silverblue through Determinate Systems’ Nix installer. But, to be honest, I’m most interested in Spectrum OS. Which I would define as the love child of NixOS and Qubes OS^[11]^. Perhaps more commonly referred to as ‘immutable’. Built-in rollback capability. No system corruption due to power outage or anything. Automatic background upgrades. Obviously, there’s a lot more I like about them. I won’t do a complete rundown, but the following is worth mentioning: (Some degree of) declarative system configuration. Reproducibility. Improved security. To be more precise; at first just the stock image, but I’ve since rebased to uBlue’s Silverblue image and more recently to my custom image using uBlue’s ‘template’. As much as I like Fedora, their repos could be a lot better; both in terms of available packages and how up-to-date the packages are. Furthermore, though more GNOME’s issue than Fedora’s, extensions add IMO excellent functionality to the table. However, they sometimes behave very unpredictable in an otherwise very predictable environment. For example, enabling the blur my shell extension somehow forces me to log out right after I try to unlock my screen; probably caused by the gnome-shell crashing for some random reason. Which would be EndeavourOS and Nobara. Which would be openSUSE Aeon and openSUSE Kalpa. Fedora Kinoite was indeed my first experience on Linux 😅. Which actually felt like a magical experience for how easy and effective it is. After their infamously steep learning curves have been conquered. Best desktop OS in terms of security and privacy. :::
Komunitas
beehaw.org
It’s great to see they’re putting some effort into some really powerful string templates. I like the versitility in the approach they’ve gone with. Just a couple things stood out to me. The problem with interpolation is that it’s dangerous as a global feature, because it doesn’t allow for validation or sanitization when constructing the final string. This exposes it, for example, to SQL or JavaScript injections. Actually no, not having interpolation just makes it more verbose to create those vulnerabilities. Also, SQL injections aren’t solved by validation and sanitization (well I guess in theory they can be if you’re aggressive enough), they’re solved with parameterized queries. These issues exist regardless, but lack of interpolation makes things significantly more verbose for the 99.9% of other cases where you aren’t sending commands to a server. The cause of these vulnerabilities is inexperienced (or sometimes careless) developers trying to concat all their queries together, and this will help mitigate that but is unlikely to solve it. The problem with interpolation seems to me to be a purely cultural thing. Pointing fingers at arbitrary reasons isn’t really helping the argument against it, and it’s completely valid to say “it just didn’t feel like it belonged in Java until now, and we wanted to make sure we got it right”, which I would much more readily believe. There has been a lot of discussion around the choice of the expression format. Due to the existence of many libraries using $, # or {} as expression delimiters, the choice was for a format that is not valid outside String Templates This seems backwards to me. Wouldn’t you want them to feel as familiar as possible to users of these libraries and make migration easy for many users? It’s not like the syntax for string templates would compile before anyway (unless there’s something I’m missing) since they all require the template processor to be specified. Skimming through the rest of the features, it seems like there’s a lot of usability improvements, which is awesome to see!
Komunitas
kbin.earth
Mostly Moo Deng number five at home, my fam is lovely. Eight when outside trying to avoid rude and entitled people going through me as if I don’t exist, taking the piss. If the shopping cart behind me is parked in my ass I contemplate number one, but I am too kind to go rabid.
Komunitas
lemmy.ml
DeepSeek-V3.1 is a hybrid model that supports both thinking mode and non-thinking mode. Compared to the previous version, this upgrade brings improvements in multiple aspects: Hybrid thinking mode: One model supports both thinking mode and non-thinking mode by changing the chat template. Smarter tool calling: Through post-training optimization, the model’s performance in tool usage and agent tasks has significantly improved. Higher thinking efficiency: DeepSeek-V3.1-Think achieves comparable answer quality to DeepSeek-R1-0528, while responding more quickly. The tool calling improvements are very welcome
Komunitas
lemmy.world
Don’t use LLMs in production for accuracy critical implementations without human oversight. Don’t use LLMs in production for accuracy critical implementations without human oversight. I almost want to repeat that a third time even. They weirdly ended up being good at information recall in many cases, and as a result have been being used like that in cases where it really doesn’t matter much if they are wrong some of the time. But the infrastructure fundamentally cannot self-verify. This is part of why I roll my eyes when I see employment of LLMs vs humans presented as an exclusionary binary. These are tools to extend and support human labor. Not replace humans in most cases. So LLMs can be amazing at a wide array of tasks. Like I literally just saved myself a half hour of copying and pasting minor changes in a codebase by having Copilot automate generating methods using a parallel object as a template and the new object’s fields. But I also have unit tests to verify behavior and my own review of what was generated with over a decade of experience under my belt. Someone who has never programmed using Copilot to spit out code for an idea is going to have a bad time. But they’d have a similar bad time if they outsourced a spec sheet to a code farm without having anyone to supervise deliverables. Oh, and technically, my example doesn’t actually require you to know the correct answer before asking. It only requires you to recognize the correct answer when you see it. And the difference between those two usecases is massive. Edit: In fact, the suggestion to replace the nouns with emojis came from GPT-4. Even though it doesn’t have any self-introspection capabilities, I described what I thought was happening and why, and it came up with three suggestions for ways to improve the result. Two I immediately saw were dumb as shit, but the idea to use emojis as representative placeholders while breaking the token pattern was simply brilliant and I’m not sure if I would have thought of that on my own, but as soon as I saw it I knew it would work.