Margaret Atwood onstage at Detroit Opera House on January 26, 2026 | Photo: Monica Morgan/Getty Images Maraget Atwood, the storied author of The Handmaid's Tale and The Blind Assassin, was interviewed as part of the Babell Literary and Cultural Festival in Porto, Portugal. As it usually does at these things, the issue of AI came up, and Atwood didn't mince words. According to Deadline's recap, Atwood said she'd used an AI chatbot exactly once, Anthropic's Claude, and came away unimpressed. She was looking for information about the British detective series Father Brown and, well:"Claude gave me the wrong answer, or it lied. Of course, it didn't know it was lying because it's not a human being; it's a large language model… It had skimmed a …Read the full story at The Verge.

Before "just ChatGPT it" was a thing, my first stop for answers wasn't Google. It was YouTube. Watching someone demonstrate how to do something just clicked in a way that a wall of text never did. That probably sounds ironic coming from a writer whose job is writing the kind of stuff people are supposed to read instead of skipping to a video, but I know I speak for the majority when I admit a quick video almost always wins.

Playing Windows games on an Android phone has gone from being a party trick to a practical reality at some point in the last few years. Most of the credit goes to Winlator and the open-source stack underneath it: namely Wine, Box64, DXVK, and an increasingly mature set of community GPU drivers. Apps like GameHub and GameNative built consumer-friendly launchers on top of that same stack, and for a lot of people, the experience is now good enough to be a genuine alternative to the likes of the Steam Deck. It doesn't help that GameHub has been marred by controversy, either.

Cursor has been one of my favorite AI coding tools for a while, but its recent deal with SpaceX instantly made me look at the market differently. It made me curious about its strongest alternatives and whether any of them already feel more useful, polished, and practical for my daily workflow. So I tried the top Cursor rivals to see which one actually deserves a spot in my coding setup. After spending time with all of them, one option clearly stood out.

I've lost count of the number of note-taking apps that I've used. From dabbling in SimpleNote to using Notion as a database for my notes, and, of course, the basics like Google Keep. Some have been great for collecting ideas, and others had an excellent interface. Elsewhere, a few promised to organize everything for me, but you know how that story goes.

Many of you are probably still using DDR4 RAM, but DDR6 RAM is already on the horizon. Poised to surpass even the blazing-fast speeds of DDR5 memory, DDR6 promises 8,800–17,600 MT/s speeds, thanks to a new 4 x 24-bit sub-channel architecture instead of the 2 x 32-bit seen on DDR5 memory. With higher frequencies come greater signal stability challenges, which is why the familiar DIMMs are going away, making way for flat CAMM2 sticks. These modules will be bolted onto the motherboard, lying parallel instead of the perpendicular layout we've been seeing for decades. With server deployment targeted for 2027 and consumer adoption soon after, the ongoing memory shortage couldn't have come at a worse moment. The 2028–2029 window that was planned for DDR6's consumer rollout might not hold anymore, considering the stratospheric prices will only conflate the already sky-high launch prices of this new technology. More memory-efficient LLMs could bring down enterprise memory demand, but that isn't guaranteed to lower RAM prices.

Teenage Engineering has already issued multiple substantial updates for its surprisingly capable $329 EP-133 KO II sampler. Its latest is one of the biggest yet. OS 2.5 adds audio over USB, selectable sample rates for lo-fi fun, sample reverse, an arpeggiator, equal-length autochopping, and it extends the maximum length of a sample from 20 seconds to 40 seconds by capturing mono, instead of stereo, audio.Sample reverse is such a simple feature that it's shocking it wasn't implemented earlier. An arpeggiator doesn't always make a ton of sense on a sampler, but the KO II sounds so incredible repitching samples (like the SK-1 successor I've a …Read the full story at The Verge.

I'm not here to argue about the controversies surrounding image or video generation handled by generative AI models. Artists should be paid for their work, end of. But if you're going to use LLMs and their ilk, running the tools locally is the way to go. It keeps your data private, lets you generate images that commercial tools won't let you, and you don't have to pay a monthly subscription fee.

The current state of hardware is so great that we've actually hit a ceiling. We have flagship graphics cards that are exponentially faster than the hardware of a decade ago. Yet when a brand-new AAA game releases and you launch it at ultra settings, it struggles to maintain a consistent 60fps at native 1440p or 4K.

Apple is looking to alleviate some of the pressure on its supply chain by seeking an exception from the Trump administration to buy RAM chips from CXMT, a company blacklisted by the Pentagon over ties to the People's Liberation Army, according to the Financial Times. The skyrocketing prices of RAM and storage have driven Apple to raise prices on almost all of its products this week, so it makes sense that it would seek alternative sources. Legally, Apple isn't barred from buying chips from CXMT, but doing business with a company tied to the Chinese military would carry serious reputational risks. It's possible that CXMT could still find it …Read the full story at The Verge.

GPU passthrough is the Proxmox trick everyone wants to talk about first. It makes for better screenshots, better forum posts, and a much more exciting project than naming storage pools or cleaning up old backups. I get the appeal, because passing a GPU into a Windows VM feels like the moment your little server finally starts showing off. But after you’ve lived with Proxmox for more than a weekend, storage is the part that decides whether the whole setup feels dependable or exhausting.

If you haven't heard the news, Fox just bought Roku for a whopping $22 billion. The acquisition isn't finalized; in fact, it isn't expected to officially close until sometime in the first half of 2027. Still, the news was enough to push me towards finally putting the finishing touches on my Jellyfin server that I'd been putting off.