Debian, Arch, Slackware? Ubuntu, Open Suse, Mint? Knoppix, Tails, Parted Magic? KDE, Gnome, Cinnamon? Anyone who deals with Linux has to process and categorize dozens of names. A few safe claims can help as landmarks. If you want to install Linux, you really are spoiled for choice: There are currently around 250 distributions available for end users, the vast majority of them free of charge. Do you have to know or even try out 250 distributions to find the right one? Certainly not: 80 to 90 percent can be filtered out in advance. In this article, I’ll present the best systems and work out the differences, advantages, and weaknesses. Main strains with Debian dominance The only thing that unites all Linux distributions is the Linux kernel. On this basis, there are five main strains on which the vast majority of distributions (derivatives) are based: Debian Linux: around 125 active distributions, including Debian, Raspbian, Knoppix, Ubuntu with numerous other derivatives such as Linux Mint RedHat/Fedora: around 25 active distributions, including Fedora, RHEL, Alma Linux Arch Linux: around 20 active distributions, including Manjaro, Endeavour-OS) Slackware: around 10 active distributions, including Porteus, Slax, and a handful more, if you still want to count Open Suse as Slackware-based Gentoo Linux: around eight active distributions, including Redcore Linux In addition, there are numerous independent distributions such as Solus-OS, Clear Linux, or Puppy Linux and — for the sake of completeness — the Android mobile system. Debian therefore has far more successors than all other Linux strains combined. The more than 50 Ubuntu derivatives alone add up to more systems than any other main Linux strain. The main reasons for the spread of Debian are its compactness, flexibility, and stability (in the most widely used “stable” branch) and the reliable package management with a huge selection of software. Many derivatives such as Linux Mint, Elementary OS, Bodhi Linux, Zorin-OS, or Bunsenlabs do not reveal their Debian or Ubuntu ancestry in their names. Knoppix, Raspbian, or the NAS system Open Media Vault are also based on Debian. This picture is only intended to convey the quantities. These are half of the Debian derivatives with the Ubuntu node. Foundry The user and desktop area is dominated by comfortable Debian systems such as Ubuntu, Mint, or Elementary OS. The bottom line is that Debian systems are the first choice for beginners, but also for many pragmatic Linux connoisseurs. The only disadvantage of Debian and others may be somewhat outdated software versions. Most Gentoo-, Slackware-, Red Hat- and Arch-based systems are not suitable for the majority, but are islands for Linux connoisseurs and for specialized areas of application. The Gentoo base is practically dying out after the end of Sabayon and the switch from System Rescue CD to the Arch base. However, there are some notable exceptions with Arch, etc.: Arch Linux: There are two particularly popular distributions here: Endeavour-OS is a very fast Linux with a graphical installer, but requires some Linux experience in everyday use. Manjaro with its graphical installer and package manager is probably the most convenient Arch Linux, but is also not a beginner’s system. Red Hat: Fedora Workstation is focused on innovation, less on stability. The “Anaconda” installer used here cannot compete with the simpler Debian/Ubuntu alternatives (Ubiquity, Calamares). Slackware: Porteus is designed as a live system (no installation) and is the first choice for a mobile and fast surfing system. Open Suse is basically based on Slackware, but is now considered independent. For more than a decade, it was almost the only Linux aimed at the PC desktop with graphical operation and configurability. The distribution has lost importance and now tends more towards innovation (e.g. BTRFS file system) and less towards beginner-friendliness. Nevertheless, Open Suse (“Leap“) remains a rock-solid choice. Package formats and containers A desktop user may not care whether their VLC player or Office program runs under Debian or Arch. The software is the same in both cases. However, as all the main strains mentioned use a different package format and different tools when obtaining the software, the choice of system for the software used plays an important role. Once you are used to the DEB package format (Debian, Ubuntu & Co., Linux Mint) and the apt terminal tool responsible for this, the changeover to RPM (Slackware, Red Hat, Open Suse), Tar.xz (Arch) or even Portage (Gentoo) is a significant hurdle and vice versa. Package management differs technically in terms of recognizing package dependencies and also in terms of operation. Graphical software centers of desktop systems should not be relied on exclusively, as they only offer a subset of the software sources. Fundamental knowledge of the respective terminal package manager is therefore important. The package format for installations and updates differs significantly between the main Linux strains. Anyone who is used to Debian or Arch will remain loyal for this reason. Foundry Apt (DEB packages under Debian/Ubuntu), Zypper (RPM packages under Open Suse) and Yum (RPM packages under Red Hat) can be considered relatively simple. You have to get used to the very concise syntax of Pacman (Arch), although only a handful of commands are required for the essentials (update, installation, uninstallation, search). Familiarizing yourself with Emerge and Gentoo’s Portage package format will be too much for normal users. The Snap and Flatpak container formats require independent management. This, their package sizes, and the increased system complexity are annoying or even off-putting for many users. If you want to avoid Snaps, you have to avoid all official Ubuntus (Ubuntu, Kubuntu, Lubuntu, Xubuntu, UbuntuMate/Budgie/Cinnamon/Unity). With Flatpak, the situation is more relaxed because here, as a rule, only the offer in the form of the management software is available, but no binding pre-installed Flatpak software. Candidates with a pre-installed Flatpak environment are Linux Mint, Elementary OS, Endless OS, Fedora, Tuxedo-OS, Zorin-OS. Different release models All Linux distributions provide standard package sources to supply the respective operating system with software and updates. There are clearly differentiated release models, which are essential for up-to-dateness and stability, but are not always communicated as clearly as they should be. Fixed: This is the regular and predominant release model with a quasi-static standard system. The fixed model is not only typical for Debian/Ubuntu/Mint, but is standard almost everywhere beyond Arch Linux. The kernel and system remain conservatively in their original state and updates only correct the current security problems. In the case of LTS long-term versions, function and kernel updates are carried out via periodic point releases. In general, this model guarantees high stability for the desktop, and even more so for server systems. However, the application software from the package sources can become relatively outdated over the years (with the exception of browsers). Rolling: This model is the rule for Arch-based distributions (Arch, Endeavour-OS, Manjaro), but can also be found elsewhere as an optional variant: Examples include Debian Sid, the Ubuntu-based Rhino Linux, or Open Suse “Tumbleweed”. The independent Solus OS is also a rolling release. Rolling releases do not recognize system versions, but keep the Linux kernel, drivers, system, and software permanently up to date — with certain risks of incompatible components. Rolling releases are suitable for users who tend to be competent, who always want to stay up to date and can fix any problems themselves. Semi-rolling releases such as MX Linux, Antix, KDE Neon, or Tuxedo-OS are hybrids between fixed and rolling. Immutable: The young, extremely secure release model “Immutable Linux” is more restrictive than the fixed model and strictly separates the core system and software. Apart from updates, the core system is static and unchangeable for both users and software. For application software, the container formats Flatpak and Snap are used, which do not interfere with the core system. Prominent candidates are Fedora Silverblue, Endless OS, and soon a variant of Debian 13. However, the immutable model is hardly recommended for normal users: The range of software is limited, the read-only system is too inflexible for server tasks, and the same applies to driver updates. Origin and sustainability In the mass of distributions on offer, some candidates may appear to be the perfect solution at first glance. However, typical desktop users or even Linux beginners should not get involved with exotics. Linux projects from small development teams may quickly become obsolete or have shortcomings that are not immediately recognizable even after trying out the live system. A lack of language support or a mixed-language system are among the most common, but by no means the most serious shortcomings. Distributions and desktops A user-friendly interface is at least as important to many users as the familiar package format or release model. However, distributions and desktops are a complicated subject. Although the well-established statement that the desktop under Linux is merely an interchangeable software application is technically true, it is still not correct. If you choose a distribution with the wrong desktop, you will regularly be disappointed by a “real” desktop installed afterwards. This also applies to distributions without a default desktop with Netinstaller (such as Debian, Open Suse, Parrot-OS), which install the desktop selected by the user. As flexible as this may seem, the result is always an unambitious standard desktop that requires reworking and possibly further installations. Desktop and distribution selection: The widest possible range of graphical management tools is important if terminal knowledge is lacking. Foundry Practically speaking: It is advantageous to choose distributions that are clearly or even unambiguously committed to a desktop. Here you can assume that the interface is optimized and delivered with all the associated components. Examples of distributions that clearly serve a desktop are official Ubuntu flavors such as Kubuntu (KDE), Xubuntu (XFCE), Lubuntu (LXQT), as well as Elementary OS (Pantheon), KDE Neon (KDE), Bodhi Linux (Moksha), or Bunsenlabs (Openbox). Most distributions avoid the restriction to one desktop and therefore offer several interfaces, but favor at least one standard. It is practically always best to choose this standard desktop — and if this desktop is not desired, it is better to choose a different distribution. Examples of distributions that clearly favor a specific desktop are Linux Mint (Cinnamon), Solus-OS (Budgie), or Parrot-OS Home (XFCE). The Linux desktops If you expect — without a terminal — the most complete graphical use possible for software installation, system configuration, drive management, and desktop customization, you cannot choose just any Linux desktop and therefore not just any distribution. The desktop for Linux systems is in principle freely selectable, as here with the Debian Netinstaller. However, you are better off with a distribution with a pre-installed standard desktop. Foundry KDE Plasma: With its configuration centers and system tools, KDE is the undisputed leader among Linux desktops. However, KDE is complex and not always beginner-friendly. Other obvious distributions would be Kubuntu, KDE Neon, or Opensuse “Leap”. Cinnamon: This desktop is probably the best Linux interface at the moment, combining a wealth of functions with (still) clear operation. Linux Mint, Ubuntu Cinnamon, and others offer the latest and most complete Cinnamon. Gnome: This desktop is unconventional, but functional and complete, although the administration area (“Settings”) is more confusing than KDE or Cinnamon. Typical Gnome distributions are Ubuntu, Fedora, or Pop-OS. Mate: The Mate interface still ranks among the complete desktops in that almost all administrative tasks can be performed graphically. Nevertheless, it only serves as the standard desktop under Ubuntu Mate, although many distributions offer it as an option. Budgie: The Gnome-based desktop makes Gnome more traditional again, but has the same confusing settings center and its own customization tools that take some getting used to. The traditional distributions are Solus-OS and Ubuntu Budgie. XFCE: The conservative desktop is easy to use and customize, but has slight deficits compared to the “big” interfaces when it comes to system tools. Model distributions are Xubuntu, MX Linux, or Voyager-OS. LXQT: This desktop is standard in Lubuntu alone, but optional in many distributions. Although LXQT borrows configuration tools from its big brother KDE, complete graphical system administration reaches its limits here. Pantheon: The desktop with a Mac look is attractive, but very reduced. System settings and customizations only offer the essentials. Pantheon is developed by the Elementary OS distribution and is standard there. LXDE/Moksha/Openbox/Fluxbox: These interfaces are representative of a number of others that a user can select specifically because they like them or because they need to save resources. They offer few configuration tools and delegate system administration to the terminal. Distributions that rely on such desktops are generally optimized for economy or live operation, such as Knoppix (LXDE), Bodhi Linux (Moksha), Bunsenlabs (Openbox), and MX Linux (Fluxbox). More hard facts about Linux distributions Wikipedia with hardcore information: The article “Comparison of Linux distributions” provides technical details on a large number of Linux distributions in several individual tables.Foundry The English-language Wikipedia website “Comparison of_Linux_distributions” shows information on all important distributions in sophisticated tables. For example, the existence of an installable live system or a graphical installer, the general orientation, the standard file system, the standard desktop or the number of software packages for each distribution can be researched here. These tables are excellent decision-making aids for a strategic distribution search. The Distrowatch website always provides up-to-date information on all Linux distributions — including servers, exotics, and extinct dinosaurs. In addition to basic data on origin and orientation, there is always a brief, rarely in-depth system characterization. A simple distribution search by name can be found at the top left of the homepage. The real highlight, however, is the search filter at distrowatch.com/search.php. Provided you have some knowledge of Linux, there is no other way to get a quicker answer to the question of whether there is an Arch-based distribution with Netinstaller and Budgie desktop. Advanced search filters on Distrowatch: The well-maintained database enables technical selection filters for a targeted system search. Foundry Read...

AI firms tip 2026 as the dawn of a post-transformer era, with memory-led, workflow-specific systems overtaking general-purpose models. Read...

It’s not just on dubious trading platforms such as Temu. Even Amazon can be infiltrated by dubious dealers who advertise SSDs with large storage capacities as supposed bargains. These cheat packs do not come from Amazon’s own warehouse, but from external service providers who offer their counterfeit products via the favorable “Warehouse” deals (i.e. as returned goods). The quality of the counterfeit drives varies — from cheap fakes to professional counterfeits that cannot be recognized as such from the outside at first glance. In addition to genuine flash memory media that have simply been given a different sticker with product specifications, there are also completely counterfeit circuit boards with a fraudulent controller. This scam can be found on external drives with a USB connection. The controller tricks the operating system into believing a false drive size, simulates write operations, but only actually stores a fraction of the data. Foundry F3: Check for flash drives With Linux tools, counterfeit or simply defective flash drives can at least be clearly and reliably recognized. The F3 (Fight Flash Fraud) tool can scrutinize flash drives of all kinds and also detects subtle errors caused by normal ageing processes. It is not fooled by manipulated controller chips, but empirically determines the real capacity of a drive as well as its read and write speeds. F3 can also check the integrity of written data by writing and reading the entire space to identify old, unreliable flash memory. F3 is a collection of command-line tools that is present in the repositories of most distributions. In Debian, Ubuntu, and others, for example, it can be installed in the terminal with the command sudo apt install f3 For flash drives such as SD cards, USB sticks, external drives with flash media, but also for internal NVMe SSDs, the program f3probe is the most suitable tool in this collection of tools. It immediately determines the real capacity of a drive and is not fooled by controller chips. Check with f3probe: The test program determines the total capacity reported by the controller and compares written data in high memory areas to see whether they are actually readable.Foundry An initial test is also non-destructive — i.e. it keeps the contents of a drive intact. Like the other F3 tools, f3probe requires root authorization or a prefixed sudo, as well as the device ID of the drive, with the command lsblk -d The command then starts (example) sudo f3probe /dev/sdd the non-destructive drive test. The device must not be mounted for this. If everything is okay with a drive, the tool returns the below message after the check, which compares the actual size with the capacity reported by the controller by means of write and read operations: Good news. The device /dev/sdd is the real thing. Check with data loss: If the data on the drive is irrelevant, the “–destructive” parameter performs a faster test run. This also requires less memory, but overwrites the entire contents of the drive. The rudely opened housing of a fake external “SSD”: The circuit board contains a manipulated controller chip and only a small SD card! Foundry Also for HDDs: Thorough test While f3probe compares the information from the controller with the memory addresses actually available on the data carrier, the F3 tools f3write and f3read take a more universal approach. f3write writes files with checksums to a mounted medium until it is completely written, and f3read then verifies this data. Because these tools work at file system level, regardless of the type of disc being tested, they require partitions that are already mounted and writable. For example, if a drive to be tested is mounted under “/media/user/5EBD- 5C80/”, then f3write /media/user/5EBD-5C80/ executes the write process and f3read /media/user/5EBD-5C80/ verifies the checksums of the written files. In general, this check only uses the free, remaining space on a drive and is therefore not destructive. NAS hard drives: The recording technology is crucial The Network Attached Storage (NAS) device class should be permanently available in the network. The mechanical hard drives, which are still the most cost-effective solution for large NAS systems, must therefore be able to withstand continuous operation. The specifications of conventional SATA hard drives for desktop PCs do not provide for such continuous operation. Hard drive manufacturers therefore offer drives for these application scenarios with the label “NAS” and a surcharge. However, as many users have realized, even the addition of “NAS” is no guarantee of problem-free operation in the NAS. Only hard disks that do not use modern SMR (Shingled Magnetic Recording) are generally suitable for a RAID network of any kind and for the ZFS file system. With SMR, the internal hard disk controller uses the disks very sparingly: Read-write heads are aligned so that the tracks on a platter overlap in order to maximize storage density. However, this approach requires several passes before a data record is reliably written and read. For RAID, which performs a checksum comparison of written data blocks, this recording technique is unsuitable and leads to errors and long waiting times. The Linux kernel developers have documented the problems observed to date at raid.wiki.kernel.org/index.php/Timeout_Mismatch. According to this, only the conventional recording technology CMR (Conventional Magnetic Recording) with linear tracks is suitable for hard disks in a RAID network. In recent years, the hard drive manufacturers Western Digital, Seagate, and Toshiba caused a major scandal when hard drives with Shingled Magnetic Recording (SMR) were marketed as NAS hard drives without sufficient labelling. Unfortunately, a hard drive does not even reveal via SMART analysis whether SMR or CMR is used. At m6u.de/cmr, a NAS service provider collects a list of suitable CMR hard drives based on user information. Related content Best SSDs: From SATA to PCIe 5.0, from budget to premium Best external drives 2025: Top picks in portable storage Buying an external drive? Wait! 5 reasons to build your own instead Read...

According to US Attorney General Pam Bondi, who shared the footage, the tanker was used to transport sanctioned oil from Venezuela and Iran. Read...

The ability of AI services to generate images is well recognized and one of the most common uses of services like ChatGPT. It is also an area that has been surrounded by controversy – photographers, artists and filmmakers are upset that Open AI and other companies have “trained” their AI models on their copyrighted works. In this guide, I gather tips on what you can do and how to get better results. Create images from text The most obvious thing you can use ChatGPT to do with images is to generate something completely new. Just give prompts like “create a picture of two rabbits playing in a meadow” or “make a photorealistic picture of a woman sitting in front of a computer drinking coffee out of a cup that says PC for Everyone” and an image will pop up that you can download and use. In many cases the images ChatGPT generates are really good, or at least good enough to use, but sometimes they don’t match what you asked for, or have obvious errors. Today, errors like too many fingers or an extra hand in a group photo are not as common, but errors that cannot be ignored are still common. When this happens, you can either try to continue in the same chat and try to get ChatGPT to tweak the image until the result is better, or try again with a new or modified prompt. Which works best will vary and you will simply have to try it out. ChatGPT first generated the image on the left. When I asked it to change it so that she is holding the cup in her left hand, I got the image on the right – note that the gaze has been wrong.Bilder genererade av Chat GPT In my experience, the results of small adjustments rarely improve enough to be worth the effort and so-called prompt engineering, where you try out different formulations, is far from an exact science. Even an adjustment that is logically very small, like adding “one of the rabbits has a pink collar” or “she’s holding the cup in her left hand” to the examples above, can lead to totally different results – or be exactly as you hope. 5 tips for better photos with ChatGPT Skärmdump Describe what you are looking forHave a picture in your head of what you want? Describe it as if you were telling someone who can’t see what you see. “A girl with brown hair and pale skin sitting at a piano in an old house with old-fashioned furnishings” is better than “a girl sitting and playing the piano” if that’s exactly how you want the image. If you don’t have a clear picture yourself, ChatGPT has nothing to go on. Sure, it can produce interesting results from time to time to see what happens when the AI is given a freer rein, but if you’re after something in particular, you need to actually say what it is. Avoid overly detailed descriptionsA detailed description is important, but it can also be too much. If you write a whole A4 page with an extremely detailed description, there is a high probability that ChatGPT will lose the thread and produce something unusable. Include the essentials but let the AI fill in the rest. InPixell_Studio “Metadata”Do you want a wide image or a square one? Should it look like a photo or a painting? Should the colors be saturated or faded? How much of the image should the main subject take up? Should the light be warm or cold, sharp or soft? Tell ChatGPT how the picture should be made, not just what it should contain. Try it againDidn’t get it quite right? Ask ChatGPT to try again, or ask it to create some different suggestions. Change the prompt and see if it gives better results. Try a more detailed description – or vice versa, a simpler one, if your original prompt was already very detailed. Start from a sketchIf you can draw a simple sketch that shows the basic composition and content of the image you’re after, you can ask ChatGPT to turn it into a finished image in any style. How well this works varies widely. Common problems include illogical composition, facial expressions that don’t match the sketch, and most of all, people looking in the wrong direction. Editing and improving your own images In addition to generating brand new images, you can use ChatGPT to edit existing images. It’s important to note that this isn’t really editing in the usual sense of the word. Every time you ask it to make a change, it generates the whole image again, it’s just that the algorithm works in such a way that most of the new image will be identical to the original. Skärmdump Once ChatGPT has created an image, you can click on it to open the chatbot’s editing interface. There’s really only one tool here, plus buttons for undo and redo. Click on the edit tool and your mouse pointer will turn into a big circle when you hover it over the image. Click and drag to paint an area that marks the part of the image where you want ChatGPT to make the changes you then ask for. This could be things like removing a distracting object, changing the details of something (like changing the print on a jumper visible in the image) or adding something new. If you want to make more general changes, you can do so directly in the prompt without selecting anything. “Remove background” often works well in its simplicity, but other changes may need a bit more detailed descriptions. Anders Lundberg Sometimes ChatGPT gets itself to make more changes than requested. Then you can try to specifically tell it not to change anything else. For example: “Change the color of the umbrella to red. Do not make any other adjustments or changes to the image.” “Zoom, enhance” A very common trope in films is that a detail is needed from a blurry photo or still from a surveillance film, and all a “computer person” needs to do is zoom in on the image and click an Enhance button. Sometimes a high-resolution version showing the necessary details pops up instantly, but sometimes the story requires it to take time, and then the computer can keep thinking for a long time. Often part of the image is shown at a time and the tension is unbearable as pixel after pixel appears on the screen. This is science fiction, of course. Information that doesn’t exist can’t be ‘recreated’ no matter how advanced an algorithm or how powerful a computer. But with AI, it can be faked. Any feature that removes distracting objects or people and fills in the background uses machine learning of some kind, whether it’s called AI or not. Older techniques like Photoshop’s content-aware fill use simpler algorithms while some newer ones use the same algorithms that AI chatbots do when generating new images. I asked ChatGPT to enlarge and fill in the missing details in this photo.Anders Lundberg The result ChatGPT spat out – not too bad.Genererad av Chat GPT Enlarging an image works in a similar way, but instead of guessing what fits to fill in a larger gap, the algorithm guesses how many small gaps to fill in so that the image is sharper (shows more detail). Since some of the information is already there, the risk of the AI coming up with something completely wrong is much lower. If you can already see what a sign says in a low-resolution image and the AI just makes the text clearer, it hasn’t lied, although it can’t be said to have recreated lost information. The result will never be identical to what it would have been if the image had simply been taken at higher resolution or better sharpness, but in most cases that difference is an academic question – what matters is whether you can use the image at the size you want without it looking blurry. ChatGPT is particularly good at making people sharp.Anders Lundberg Another thing you can try is to ask ChatGPT to sharpen a blurred image, for example a photo where the camera focused wrong. This can work really well if the image is only slightly blurry, but if it’s very blurry it guesses wildly and then the person in the photo can look like someone else entirely. Apply a certain style to images ChatGPT has become known for being good at a particular kind of editing – turning a photo or other image into a picture with a particular style. You’ve probably seen examples of the trend to ask ChatGPT to make images in Studio Ghibli style, that is, with a cartoon style similar to films directed by Hayao Miyazaki. It’s very good at it, but be aware that the creators you make it mimic have in most cases been sharply critical of the move. Some have sued Open AI for copyright infringement. Less controversial is asking ChatGPT to change the image to a style that is not that of any individual artist, for example “turn this photo into a watercolour painting”, or asking for a style that belongs to a long-dead artist like Rembrandt. Foto: Anders Lundberg, målning genererad av Chat GPT You can also upload an existing image to have as a reference and ask ChatGPT to remake other uploaded images to match the style of that image. A trick you can try if this does not give satisfactory results is to upload the example in a new chat instead and ask ChatGPT to “generate a description of the image that could be used to ask ChatGPT to apply the same style to another image”. Paste the results into the chat where you have uploaded the image(s) you want to change the style of. Skärmdump Gallery In the top left column of ChatGPT, under New Chat and Search Chats, you will find the Gallery feature. It’s a repository for all the images you’ve generated with ChatGPT (technically only with the GPT-4o model, not images generated with the older Dall-e model). It makes it easier to find specific images you have generated, so that you can, for example, continue working or look up how you wrote the prompt at the time. Click on an image and then on Open in chat in the top right corner to go to the thread where the image was generated. Skärmdump Generate video with Sora In addition to image generation, Open AI has developed algorithms that can generate video, and is offered as a separate service called Sora, with its own website and app. Sora is not embedded in ChatGPT mainly because the service requires a more advanced user interface, and Open AI wants to keep ChatGPT’s simple interface. Sora is exciting and can create scarily realistic videos. Going through everything you might find useful about video generation would take up more space than I have in this guide. But you can start from the same basic tips as for image generation. My second tip is to try and play around with the service. But keep in mind that you can create a maximum of 15 10-second clips a day unless you have an expensive Pro subscription. Skärmdump Projects and GPTs Just like with text, you can use projects to keep all of your chats organised and add files and instructions to accompany any new chats in that project. This is ideal if, for example, you’re using ChatGPT to create image resources for a website or anything else where you want to stick to a consistent style. If you pay for a Plus subscription, you can also use the GPT feature to create customised versions of the chatbot, not to mention accessing GPTs created by other users, like the upscaling GPT I mentioned above. Read more about the benefits of projects and GPTs here. AI-generated images and copyright If you let ChatGPT or another AI service generate images for you, you have no copyright on them. It doesn’t matter how detailed your description was or how much you fiddled with the prompt. This means that others can copy ‘your’ images and use them, without asking you and without you being able to do anything about it. It is also illegal to claim that you own the copyright to an AI-generated image. However, if you take an AI-generated image and make major changes to it using a program such as Photoshop, it can become a “work of authorship”, which gives you the copyright to it. The same applies if you paint an image that the AI has generated – then it is your painting that you have the copyright to, not the generated original. The US Library of Congress has a good guide to AI and copyright, which also warns of the risk of an AI infringing someone else’s copyright. If you’re just using the images for personal use, there’s a low risk of you being sued, for example by Studio Ghibli if you’ve made a portrait of yourself “Ghibli-style”, but for those running a business, it’s more important to be careful. Read...

Vincent Arbuckle, Deputy Director-General of New Zealand Food Safety, confirmed that the investigation has found no ongoing food safety risks associated with the School Lunch Collective. Read...

At a glance Top AMD mini PC deal picks Acemagician S3A—$429 (20% off on Amazon) Acemagician Kron K1—$279 (12% off on Amazon) Bosgame P3—$440 (20% off on Amazon) Top Intel mini PC deal picks Acemagic V1—$259 (19% off on Amazon) Geekom GT2 Mega—$1049 (19% off on Amazon) Beastcom M1 Mini—$145 (19% off on Amazon) Whether you’re looking for a productivity mini PC, something for gaming, or just a budget-friendly machine with a small footprint, we’ve got you covered. The team at PCWorld continually sorts through all of the daily mini PC sales and puts together a curated list of the best deals available. Finding the right deals can be tricky, however, because there are so many things to consider, from the configuration to the price. Fear not! I’ve been covering mini PC deals for a long time — and PCWorld has tested thousands of PCs over four decades — so let me help you out. I select our favorite deals by thoroughly examining the list of specs, overall design, user feedback, and general value. The RAM, SSD, and CPU all matter significantly in mini PCs, and if you’re not careful, you may select a “bare-bones” rig that makes you bring your own memory and storage. All of the picks below are hand-selected for not only their performance, but most importantly, their value. I’ve also included some helpful answers to common questions about buying a computer at the bottom of this article. If you’re considering a laptop instead, be sure to check out our best laptop deals, or for more desktop choices check out best computer deals roundup instead—each are updated daily. Updated Dec 8, 2025 to reflect the latest deals and prices. Best deals on AMD mini PCs Amazon Acemagician S3A, AMD Ryzen 7 8745HS/16GB DDR5 RAM/512GB SSD/Radeon 780M iGPU/triple 4K support—$429 (20% off on Amazon) If I were to pick a single mini PC deal right now, I’d go with the Acemagician S3A for $429 because that’s a great price for the specs. It’s more than capable of handling your daily workload, browsing, streaming, and even light gaming thanks to the Radeon 780M integrated graphics. Acemagician Kron K1, AMD Ryzen 5 7430U/16GB DDR4 RAM/512GB SSD/triple 4K support—$279 (12% off on Amazon) The Acemagician Kron K1 is a great budget deal right now for only $279. With 16GB of RAM and a 512GB SSD, you’ll be able to accomplish all of your productivity needs. It would also work well as a streaming or multimedia device as well as a smart home hub—giving you centralized control without breaking the bank. Bosgame P3, AMD Ryzen 7 7840HS/32GB DDR5 RAM/1TB SSD/triple 4K support—$440 (20% off on Amazon) If you’re looking for a bit more juice, the Bosgame P3 for $440 is the mini PC to go for. Not only does it pack 32GB of RAM and a generous 1TB SSD, but the Ryzen 7 processor is ready to handle more intensive tasks should you need it. The fact that the RAM is DDR5 at this price point makes it even more of a great deal. GMKtec M7, AMD Ryzen 7 Pro 6850H/16GB DDR5 RAM/1TB SSD/quad 4K support, Oculink—$420 (25% off on Amazon) Personally, I think the GMKtec M7 deal here is a sneaky bargain. Sure, it comes with less RAM than the Bosgame above and a slightly weaker CPU, but it makes up for it with additional 4K and Oculink connectivity—meaning you can easily hook up an external GPU should you decide to get your game on. GMKtec M5 Plus, Ryzen 7 5825U/32GB RAM/1TB SSD/triple 4K support—$368 (8% off on Amazon) As an alternative to the M7 model, you could opt for the GMKtec M5 Plus deal for $368 instead. You won’t get the same gaming capabilities, but more RAM—even if it is DDR4—for less money is always a plus, especially if you’re looking to use it for work or productivity tasks. Plus, it comes with expansion slots for RAM and storage if you want to upgrade it even further in the future. Geekom A9 Max, AMD Ryzen AI 9 HX 370/Radeon 890M/32GB DDR5 RAM/2TB SSD/quad 4K support—$1,100 (8% off on Amazon) For a top of the line mini PC packed with powerful components, go with the Geekom A9 Max for $100 off. The Ryzen AI 9 HX 370 is a very powerful CPU that is not only capable of 1080p gaming combined with the 890M GPU, but also breezes through AI tasks. Ample RAM and SSD storage along with excellent connectivity complete the package. It’s still expensive, but $100 off is better than nothing, right? Best deals on Intel mini PCs Acemagic V1, N150/16GB RAM/1TB SSD/dual 4K support—$259 (19% off on Amazon) This deal for the Acemagic V1 is just a good overall bread-and-butter mini PC deal. The N150 processor and 16GB of DDR4 RAM means it’s powerful enough to handle your productivity and everyday work tasks. It comes with enough SSD storage to save just about everything you need—and includes an additional M.2 slot in case you want to upgrade later. It might not be fancy, but this is a good price for a mini PC that’ll be sufficient for most everyone’s needs. Geekom GT2 Mega, Intel Core Ultra 9 285H/Arc 140T/32GB DDR5 RAM/2TB SSD/quad 4K support—$1049 (19% off on Amazon) If you’re looking for an Intel mini PC that can do it all, then the Geekom GT2 Mega is the one to buy. Packed with a Core Ultra 9 processor, DDR5 RAM, and ample 2TB of onboard SSD storage, this will easily handle everything you throw at it. It’s a full desktop replacement with a smaller footprint and cheaper price tag. Plus, thanks to the Arc 140T GPU you can do all of the esports gaming you want or even play modern AAA games at 1080p with decent frame rates. Beastcom M1 Mini, N100/16GB RAM/512GB SSD/triple 4K support—$145 (19% off on Amazon) True to its name, the Beastcom M1 Mini mini for $145 is, well, a beast—an entertainment beast at least. Its triple 4K support is practically unheard of at this price. This is meant to be an entertainment mini PC thanks to its ultra-small footprint and excellent connectivity options. However, those looking for a productivity or work mini PC should probably look elsewhere. GMKtec K10, Core i9-13900HK/32GB RAM/1TB SSD/2.5G LAN WiFi6/quad 4K support—$600 (17% off on Amazon) The GMKtec K10 deal for $119 off is a solid bargain thanks to its powerful i9 CPU. Plus, the addition of a 2.5G ethernet port and Wi-Fi 6 support means it is adept at handling most networking tasks you might throw at it. To top it all off, it can support up to four 4K displays, one of which is a DisplayPort 1.4 capable of 8K as well. IdeaCentre Mini, Core 5 210H/16GB RAM/1TB SSD/dual 4K support—$600 (25% off on Lenovo) To the naked eye, this IdeaCentre Mini deal for $600 might not seem like a good value. But it comes with a couple of hidden extras that make it worth the higher price. Support for Wi-Fi 7, Bluetooth 5.4, and DDR5 RAM are just some of the advantages. Add to that the reliability and high build-quality of Lenovo machines and suddenly this becomes a more than worthwhile mini PC deal this week. Mini PC deals FAQ 1. What should you look for when buying a mini PC? The first thing you have to check off your list is just what configuration you need to get your job done. Are you going to use your new mini PC for browsing only? Then a cheap budget-friendly device with a decent CPU and 16GB of RAM will do the trick. Do you want it to be able to juggle a bunch of apps and not choke on 50 browser tabs? Then a more powerful CPU and at least 32GB of RAM would be better. Do you want to do some light gaming in your downtime? Look for a device that comes with a GPU or Occulink connectivity to hook up an external GPU instead. Are you looking to utilize the mini PC as a home server? Your priority then should be a large SSD—2TB or more. There are devices for each and every one of these scenarios, and they’re all quite affordable. The best part is that mini PCs are… well, small, so they can be placed practically anywhere depending on your intended use and available space. They also support multiple monitors out of the box, which makes it easy to replace your old, larger desktop and you won’t even need to pay for extra hubs. 2. Can a mini PC replace a desktop computer? For most people, the answer is “yes”! Mini PCs are definitely powerful enough to replace desktops for most users, especially if you’re going to use it for work, browsing, or streaming movies. If you want something for photo and video editing, or even gaming, there are models that allow you to do that, but they’ll cost a bit more. If you’re into hardcore gaming, however, you’re still better off with a proper gaming PC or laptop. 3. Can a mini PC run modern games? Mostly, yes. But it will ultimately depend upon the mini PC you buy and which modern games you want to play. Some high-performance mini PCs can run modern AAA games. You might not be able to run them at the highest settings though, and performance will depend heavily on the specific hardware. If you do want to game on a mini PC, look for powerful models with dedicated discrete GPUs such as an Nvidia RTX or AMD Radeon rather than integrated graphics. Alternatively, you can opt for cloud streaming services to game on less powerful mini PCs, which bypasses the need for high-end hardware. 4. Which mini PC brands are reliable? Some of the most reliable mini PC brands are Minisforum, Beelink, GMKtec, and Geekom. Of course Apple’s Mac minis are also excellent computers if you prefer using macOS. Brands such as Asus, Lenovo, HP, and Dell all make high-quality smaller-footprint PCs, but they might not fill the same niche and/or they are typically more expensive. Read...

The auditor-general`s annual report said the spending also included food for local families and in some cases travel for principals` partners. Read...

Te Pati Maori boss John Tamihere was unflinching at the praty’s annual general meeting, even as his leadership was called into question. Read...

Nvidia is reactivating a feature that many of you may have already written off: The new Game Ready driver 591.44 brings back PhysX support for selected older games. For owners of a GeForce RTX 50 graphics card, this is a very relevant change — especially if you enjoy playing classic games with elaborate physics effects. Why PhysX is important – and why it was removed PhysX has been a mainstay in PC gaming for many years, creating realistic effects for fabrics, smoke and particles in titles such as “Borderlands 2” and “Batman: Arkham City.” But Nvidia cancelled PhysX support at the beginning of 2025: New RTX-50 series graphics cards could no longer accelerate PhysX calculations in 32-bit games via GPU. As a result, the calculations landed entirely on the CPU — with significant performance losses compared to prior generation GeForce cards. Some gamers reacted with a hilarious, yet totally functional workaround: They plugged an older Nvidia card into their computer just to be able to continue using PhysX effects. Driver 591.44: PhysX returns – but only for select games With the current driver, there is now a turnaround. Nvidia is introducing a so-called “Custom Support” level. This is not a complete return of the old 32-bit support, but rather individually created profiles for a hand-picked selection of particularly popular PhysX titles. These nine games now benefit from GPU acceleration again: Alice: Madness Returns Assassin’s Creed IV: Black Flag Batman: Arkham City Batman: Arkham Origins Borderlands 2 Mafia II Metro 2033 Metro: Last Light Mirror’s Edge Nvidia has announced additional support for “Batman: Arkham Asylum” for the first half of 2026. Whether other games will follow remains to be seen. Nvidia writes that it has initially focused on the most popular classics that are still frequently played today. PhysX technology originally comes from Ageia, which developed its own physics processor unit (PPU) around 20 years ago. Nvidia took over Ageia in 2008 and integrated PhysX calculations directly into the CUDA cores of its own GeForce graphics cards. This made the additional PPU card superfluous, but also tied the effects firmly to Nvidia hardware. Driver also brings optimizations and bug fixes In addition to the return of PhysX, Nvidia’s new Game Ready driver offers further improvements: Optimizations for “Battlefield 6: Winter Offensive” and “Call of Duty: Black Ops 7”, especially in conjunction with DLSS 4. Fixes for graphics artifacts in “The Witcher 3”. Stability fixes for Adobe Premiere Pro. Driver 591.44 is available for download as usual via the Nvidia app or the Nvidia website. If you play classic games with PhysX effects, the update should bring noticeable improvements — often significantly higher frame rates and reactivated graphical effects. If, on the other hand, you only play modern titles, you will primarily benefit from the general optimizations and bug fixes. Read...