Posts

RISC OS Open Turns 20

Image
RISC OS turned 20 last month, which is the kind of milestone that usually gets celebrated with corporate retrospectives and vague "legacy matters" platitudes. But here's the thing: this isn't just another aging codebase limping along on nostalgia. The open-source effort behind RISC OS has quietly been doing something most retro-computing projects never attempt—actually moving forward. Take memory cards alone. In year six, the team got native SD/MMC support working on BeagleBoard hardware. That's not just about wrangling ancient drivers into shape. It means the OS could suddenly read modern storage without kludges or adapters—a real capability, not a museum piece. And then there was the Raspberry Pi port in year seven, which didn't just prove the OS could run on dirt-cheap hardware. It proved it could scale from a niche retro oddity to something that could plausibly power your next hobby project. The traffic numbers back this up: their website hit recor...

Vocabulary Shift in US Workforce

Image
I've been looking at the words used to teach English language learners, and what's striking is how much they've changed over time. It's not just that new words have been added - though they have - but that the overall character of the list reflects fundamental shifts in the US workforce and society. Take a look at some of the words that are now considered essential: mortgage, corporation, analysis. These aren't the kind of words you'd typically associate with everyday conversation, but they're the kind of terms that are increasingly relevant in a service-oriented economy. When you think about it, this shift makes sense. By 1957, just four years after the original list was published, white-collar workers outnumbered blue-collar workers for the first time in US history. The nature of work was changing, and the language we use to describe it was changing too. But what's interesting is that this change wasn't just about adding new words - it was ...

Go 1.27 Overview

Image
I've been covering the Go programming language for years, and I've got to say, the upcoming 1.27 release has some interesting changes. One thing that caught my eye is the new ability for method declarations to declare their own type parameters - it's a feature that's been a long time coming, and it's going to make certain types of code a lot more expressive. For example, if you've got a generic container and you want to create a map that wraps lines, you can now do that in a way that's both type-safe and flexible. What's nice about this release is that it's not just about adding new features - it's also about making the language more consistent and easier to use. The official release notes are, as usual, a bit dry, so I thought it would be helpful to put together a hands-on guide with runnable examples that show what's changed and how the new behavior works. We'll take a closer look at some of the key features, including the new m...

MI355X GPU vs B300

Image
I've been following the developments in AI hardware, and one thing that's caught my attention is the MI355X GPU. It's offering surprisingly competitive performance to the B300 when it comes to running Kimi K3, and that's largely due to its massive 288GB of VRAM. What's really interesting, though, is that the MI355X is managing to keep up with the B300, which is a much more established player in the field. Running Kimi K3 at around 952 tok/s/node, AMD is continuing to make a strong case for itself as the winner in terms of performance per dollar. Over the past few months, we've seen a huge explosion in the capabilities of open source models, and it's clear that the hardware is starting to catch up. One of the key factors in getting Kimi K3 to its current throughput number was the use of speculative decode, which required some additional work to get right. I'm still trying to wrap my head around what this means for the future of AI hardware, thoug...

Linux Hits 10% Market Share

Image
I've been watching the Linux desktop market for years, and I have to say, the latest numbers caught me off guard: Linux desktop market share has surpassed 10% in North America. That's a significant milestone, and it's got me wondering what's driving this growth. Is it the usual suspects - gamers, developers, and power users - or is something more at play? The fact that GNU/Linux is free and open source software is certainly a draw for many users. It's a collaborative effort between the GNU project, which was formed back in 1983, and the development team of Linux, the kernel. This collaborative approach has led to a robust and customizable operating system that's attractive to users who value flexibility and control. But, I've seen Linux market share fluctuate over the years, so I'm not convinced this growth is sustainable - at least, not yet. What's really interesting to me is that this growth is happening in North America, where Windows and...

AI Financial Advice: How Machine Learning Outperforms Humans

Image
I've been following the development of AI in finance for a while now, and I have to say, the latest findings are pretty surprising. Researchers have discovered that AI can provide remarkably accurate financial advice when asked the right questions, even outperforming human advisors in certain scenarios. This isn't just about AI being able to crunch numbers faster than humans - it's about the quality of the advice itself. I've seen some of the results, and it's impressive: in some cases, AI is able to identify investment opportunities that human advisors might miss, and provide more personalized recommendations to boot. What's interesting is that this isn't just a matter of AI being better at analyzing data - although that's certainly part of it. It's also about the fact that AI can ask questions, and learn from the answers, in a way that human advisors can't. For example, an AI system might ask a user about their investment goals, risk to...

Diátaxis Framework

Image
I've been writing tech docs for years, and I've seen my fair share of frameworks and methodologies come and go. But Diátaxis, a documentation framework that's been gaining traction lately, has really caught my attention. What sets it apart, in my opinion, is its focus on creating user-centered content - it's not just about writing clear instructions, but about understanding what users actually need to know. The idea behind Diátaxis is simple: it's a framework that helps tech writers create documentation that's tailored to the user's journey. It's based on four main principles: tutorials, which help users get started; explanations, which provide context and background information; reference materials, which offer detailed technical specs; and troubleshooting guides, which help users resolve common issues. I've been exploring Diátaxis for a while now, and I'm impressed by its potential to streamline the documentation process - but I'm a...