You try and build a PC to play the most demanding game you have in your library, but that's not the case for me. I built a system costing around $2,000 over a couple of years, and it's primarily used to play a game that barely scratches the surface of the available performance. That game is OldSchool RuneScape, and I've been playing (on and off) for approximately 18 years, though I've been a fan of the MMORPG for almost 30 years, since starting as a wee lad with friends at high school (that's sixth-grade to many of you).

For years, Adobe Acrobat, Foxit PDF, and Nitro PDF have been the go-to solutions for many. However, unless you manipulate large PDF files frequently, these paid PDF editors feel overkill. What if I told you there's a powerful, versatile, and completely free way to handle all your PDF tasks, right on your own terms? After trying out dozens of free tools, I came across Stirling PDF, which ticks all the right boxes for me.

I’ve spent years tweaking my home lab setup, and while my main server handles the heavy lifting, I’ve come to appreciate how useful the Raspberry Pi can be for running specific services. Offloading tasks to a Pi reduces the load on my main server, helps isolate potentially risky software, and makes it easier to recover from crashes. It's also been a great way to experiment without messing up anything critical.

Dipping my feet in the self-hosting trend is great. But when a browser pops the "not secure" error, I worry about running my humble stack over HTTP, even though it's on my local network. Since I share some of them with my family, it's hard to discourage them from clicking on suspicious links while my apps run on plain HTTP. I tried enabling HTTPS with self-signed certificates for my self-hosted apps and services, but I kept hitting the wall.

The future of everything from smartphones, to military equipment, to electric vehicles hangs on 17 rare earth minerals and the magnets that they're made into. And China, the world's largest refiner and producer, is tightening its grip and threatening the US' largest automakers. Over the last 30 years, China has methodically cornered the market on mining and refining rare earth minerals, which are used to produce a variety of common items like passenger vehicles and everyday electronics. In the wake of US President Donald Trump's increasingly aggressive trade war, China is leveraging its position as the world's largest producer, at the expense of the American auto industry. "We're within 90 days of this becoming a critical problem for everybody," say Ambrose Conroy, founder of Seraph Consulting and a major investor in Democratic Republic of Congo mining operations."We're within 90 days of this becoming a critical problem for everybody."According to analysts, more than 90 percent of the world's supply of those 17 elements at the bottom of the periodic table are mined, refined, and turned into rare earth magnets in China. After Trump announced tariffs of up to 145 percent on …Read the full story at The Verge.

Sony’s original PlayStation wasn’t just a game console but a full-blown cultural reset. While it would go on to spar with Microsoft’s Xbox in later generations, the PS1’s earliest battles were against Nintendo’s N64 and Sega’s Dreamcast. What’s wild is that we almost got a Sony-Nintendo hybrid with a disc drive, but that weird alternate timeline never came to pass. Instead, Sony went solo — and ended up making history.

Home Assistant has gone on to be one of the must-have recommendations for any home lab setup, even if this simply consists of a single network-attached storage (NAS) enclosure. The company behind this impressive platform has launched a preview device called the Home Assistant Voice. It's essentially a smart speaker with some powerful microphones for picking up and using your voice to carry out commands. Unlike Alexa (and some other smart assistants), everything can be processed and carried out locally on your local area network (LAN).

At the time of writing, there are two graphics cards in existence that support FSR 4. That's been the case for two months, and although we may get another GPU that supports FSR 4 soon, the lineup still looks slim. It looks even slimmer when you consider how difficult it has been to find AMD's latest GPUs in stock.

If you're looking to get the most out of your CPU and need the best cooling solution, an AIO is the way to go. Luckily, the one that we love, the Arctic Liquid Freezer III 360 A-RGB, is not all that expensive right now, thanks to a sweet deal from Amazon that brings the price down to just $100. If you're quick, you can save 31% off retail. So get it while you still can.

Windows 11 may not be a perfect operating system out of the box, but one of the things that makes it great is the ability to use third-party programs to change its looks and behaviors to fit your needs and preferences. Among those tweaks, I personally find taskbar customization to be one of the most interesting things you can do, because it's something that's on screen at all times (usually).

We've had the celebratory party from May's Patch Tuesday, and now we're suffering the nasty hangovers that come afterward. A few days after a new big patch goes live, it's not unusual to see lots of people make reports that their operating system isn't working as well as it used to. So, let's check out some of the reports that are coming in, and if you're experiencing any of these, it's worth noting that there's a high chance the issue lies with Windows before you start blaming hardware or other software.

Chatbot upgrades, hot and fresh.The latest versions of OpenAI’s multimodal GPT AI models are now rolling out to ChatGPT. OpenAI announced on Wednesday that GPT-4.1 will be available across all paid ChatGPT account tiers and can now be accessed by Plus, Pro, or Team users under the model picker dropdown menu. Free users are excluded from the rollout, but OpenAI says that Enterprise and Edu users will get access “in the coming weeks.”GPT-4o mini, the more affordable and efficiency-focused model that OpenAI launched last year, is also being replaced with GPT-4.1 mini as the default option in ChatGPT for all users, including free accounts. Users on paid accounts will also see GPT-4o mini replaced by GPT-4.1 mini in the model picker options. Both GPT-4.1 and GPT‑4.1 mini are optimized for coding tasks and instruction following, and outperform GPT‑4o and GPT‑4o mini “across the board,” according to OpenAI. Both of the new models support a one million context token window — the amount of text, images, or videos in a prompt that an AI model can process — that greatly surpasses GPT-4o’s 128,000-token limit. OpenAI says that speed improvements also make GPT-4.1 more appealing for everyday coding tasks compared to the powerful OpenAI o3 & o4-mini reasoning models it introduced in April.GPT-4.1 and GPT-4.1 mini were made available to developers last month, alongside the GPT-4.1 nano model, which OpenAI described as its “smallest, fastest, and cheapest” offering to date. GPT-4.1 Nano is absent from this model rollout announcement, and it’s unclear when — or if — it’ll be generally available in ChatGPT.