The Most Popular Blogs of Hacker News in 2025
Who were the most popular personal bloggers of 2025, and what made them successful on Hacker News?
The RAM Shortage Comes for Us All
Memory price inflation comes for us all, and if you're not affected yet, just wait. I was building a new PC last month using some parts I had bought earlier this year. The 64 Gigabyte T-Create DDR5 memory kit I used cost $209 then. Today? The same kit costs $6…
Minisforum stuffs an entire Arm Homelab in the MS-R1
The Minisforum MS-R1 uses the same Cix CD8180 Arm SoC as the Orion O6 I reviewed earlier this year. But everything else about this thing is different. What this thing should be, is a box that runs Linux and can compete with at least an Apple M1 Mac mini, or a …
JeffGeerling.com has been Migrated to Hugo
Since 2009, this website has run on Drupal. Starting with Drupal 6, and progressing through major site upgrades and migrations to 7, 8, 9, and 10, I used the site as a way to dogfood the same CMS (Content Management System) I used in my day job for over a deca…
1.5 TB of VRAM on Mac Studio - RDMA over Thunderbolt 5
Apple gave me access to this Mac Studio cluster to test RDMA over Thunderbolt, a new feature in macOS 26.2. The easiest way to test it is with Exo 1.0, an open source private AI clustering tool. RDMA lets the Macs all act like they have one giant pool of RAM, …
How to silence the fan on a CM5 after shutdown
Out of the box, if you buy a Raspberry Pi Compute Module 5, install it on the official CM5 IO Board, and install a fan on it (e.g. my current favorite, the EDAtec CM5 Active Cooler), you'll notice the fan ramps up to 100% speed after you shut down the Pi. That…
All Intel GPUs run on Raspberry Pi and RISC-V
We finally have Intel Arc GPUs working on the Pi somewhat stably—it required overcoming many small hurdles, but it looks like support could land in Raspberry Pi OS if we can get a simple patch upstreamed1. If that happens, all you'd need to do to use an Intel …
It's not that hard to stop a Trane
Six years ago, I replaced the old HVAC system that came with our house, a central forced air system installed in 19951. The new system is a Trane XR AC paired with an S9V2 96% efficiency forced-air gas furnace. And it ran great! Better efficiency, quieter, mul…
Testing the Mono Gateway, a custom-built 10 Gbps Router
Last month, the stars aligned for me to bring the Mono Gateway (a 10 Gbps router that YouTuber Tomaž Zaman and his team at Mono built from scratch) on a trip to Phoenix, and test it with one of the most OP network test boxes I've ever seen, at the ServeTheHome…
Converting hot dog plasma video to sound with OpenCV
When you ground a hot dog to an AM radio tower, it generates plasma. Your browser does not support the video tag. While the hot dog's flesh is getting vaporized, a tiny plasma arc moves the air around it back and forth. And because this tower is an AM tower, i…
NIST was 5 μs off UTC after last week's power cut
If you were 5 microseconds late today, blame it on NIST. Their facility in Boulder Colorado just had its power cut for multiple days. After a backup generator failed, their main ensemble clock lost track of UTC, or Universal Time Coordinated. But even if you u…
CM0 - a new Raspberry Pi you can't buy
This little postage stamp is actually a full Raspberry Pi Zero 2, complete with eMMC storage and WiFi. But you can't get one. Well, not unless you buy the CM0NANO development board from EDAtec, or you live in China. This little guy doesn't have an HDMI port, E…
The DC-ROMA II is the fastest RISC-V laptop and is odd
Inside this Framework 13 laptop is a special mainboard developed by DeepComputing in collaboration with Framework. It has an 8-core RISC-V processor, the ESWIN 7702X—not your typical AMD, Intel, or even Arm SoC. The full laptop version I tested costs $1119 and…
The Arduino Uno Q is a weird hybrid SBC
The Arduino Uno Q is... a weird board. It's the first product born out of Qualcomm's buyout of Arduino. It's like if you married an Intel CPU, and a Raspberry Pi RP2040 microcontroller—oh wait, Radxa's X4 did that. Arduino even tried it before with their old Y…
Air Lab is the Flipper Zero of air quality monitors
This air quality monitor costs $250. It's called the Air Lab, and I've been using it to measure the air in my car, home, studio, and a few events over the past few months. And in using it over the course of a road trip I learned to not run recirculate in my ca…
Why doesn't Apple make a standalone Touch ID?
I finally upgraded to a mechanical keyboard. But because Apple's so protective of their Touch ID hardware, there aren't any mechanical keyboards with that feature built in. But there is a way to hack it. It's incredibly wasteful, and takes a bit more patience …
Using AMD GPUs on Raspberry Pi without recompiling Linux
I'm working on a more in-depth test of some newer AMD GPUs on the Raspberry Pi, now that the 15 line kernel patch is (IMO) nearly ready for upstreaming. But this blog post shows how to quickly get almost any modern AMD GPU running on a Raspberry Pi 5, CM5, or …
Dell's version of the DGX Spark fixes pain points
Dell sent me two of their GB10 mini workstations to test: In this blog post, I'll cover the base system, just one of the two nodes. Cluster testing is ongoing, and I'll cover things like AI model training and networking more in depth next year, likely with com…
Nvidia Graphics Cards work on Pi 5 and Rockchip
A few months ago, GitHub user @yanghaku dropped a 15 line patch to fix GPU support for practically all AMD GPUs on the Raspberry Pi (and demoed a 3080 running on the Pi with a separate, unreleased patch). This week, GitHub user @mariobalanica dropped this (lar…
Big GPUs don't need big PCs
Ever since I got AMD, Intel, and Nvidia graphics cards to run on a Raspberry Pi, I had a nagging question: What's the point? The Raspberry Pi only has 1 lane of PCIe Gen 3 bandwidth available for a connection to an eGPU. That's not much. Especially considering…
Benchmarking NVENC video transcoding on the Pi
Now that Nvidia GPUs run on the Raspberry Pi, I've been putting all the ones I own through their paces. Many people have an older Nvidia card (like a 3060) laying around from an upgrade. So could a Pi be suitable for GPU-accelerated video transcoding, either s…
What is "good taste" in software engineering?
--
Unification across Microsoft Platforms
'Unification across Microsoft Platforms' written by davepermen
The Autoscaling Web
'The Autoscaling Web' written by davepermen
Is it worrying that 95% of AI enterprise projects fail?
In July of this year, MIT NANDA released a report called The GenAI Divide: State of AI in Business 2025. The report spends most of its time giving advice about…
Run Cleanup Tasks in Windows
'Run Cleanup Tasks in Windows' written by davepermen
How I provide technical clarity to non-technical leaders
My mission as a staff engineer is to provide technical clarity to the organization. Of course, I do other stuff too. I run projects, I ship code, I review PRs…
Testing the Mono Gateway, a custom-built 10 Gbps Router
We are in the "gentleman scientist" era of AI research
Many scientific discoveries used to be made by amateurs. William Herschel, who discovered Uranus, was a composer and an organist. Antoine Lavoisier, who laid…
Woohoo! Wireless LAN!
'Woohoo! Wireless LAN!' written by davepermen
2025 was an excellent year for this blog
--
Dell's version of the DGX Spark fixes pain points
Dell sent me two of their GB10 mini workstations to test:
Why it takes months to tell if new AI models are good
--
Nobody knows how large software products work
--
Endless AI-generated Wikipedia
--
Post from the cellphone
'Post from the cellphone' written by davepermen
The DC-ROMA II is the fastest RISC-V laptop and is odd
AI coding agents rely too much on fallbacks
--
Windows 10 UAP Extend Titlebar
'Windows 10 UAP Extend Titlebar' written by davepermen
Becoming unblockable
With enough careful effort, it’s possible to become unblockable. In other words, you can put yourself in a position where you’re always able to make forward…
An unofficial FAQ for Stripe's new "Tempo" blockchain
Stripe just announced Tempo, a “L1 blockchain” for “stablecoin payments”. What does any of this mean? In 2021, I was interested enough in blockchain to write a…
NIST was 5 μs off UTC after last week's power cut
If you were 5 microseconds late today, blame it on NIST.
Simple Immutable POCO Types in C#
'Simple Immutable POCO Types in C#' written by davepermen
MVVM in XAML — Binding of Commands to Root Element
'MVVM in XAML — Binding of Commands to Root Element' written by davepermen
What have we learned about building agentic AI tools?
In the middle of 2025, agentic coding finally became a thing: first with the release of Claude Sonnet 4, the first “smart enough to be useful” agentic model…
What I learned building an AI-driven spaced repetition app
--
CM0 - a new Raspberry Pi you can't buy
Software engineers should be a little bit cynical
--
What I use (12013 HE)
'What I use (12013 HE)' written by davepermen
Should LLMs just treat text content as an image?
Several days ago, DeepSeek released a new OCR paper. OCR, or “optical character recognition”, is the process of converting an image of text - say, a scanned…
What I use (12025 HE)
'What I use (12025 HE)' written by davepermen
Why doesn't Apple make a standalone Touch ID?
I finally upgraded to a mechanical keyboard. But because Apple's so protective of their Touch ID hardware, there aren't any mechanical keyboards with that feature built in.
You can't design software you don't work on
--
AI is good news for Australian and European software engineers
Right now the dominant programing model is something like “centaur chess”, where a skilled human is paired with a computer assistant. Together, they produce…
Mistakes I see engineers making in their code reviews
In the last two years, code review has gotten much more important. Code is now easy to generate using LLMs, but it’s still just as hard to review. Many software…
Only three kinds of AI products actually work
--
'Make invalid states unrepresentable' considered harmful
One of the most controversial things I believe about good software design is that your code should be more flexible than your domain model. This is in direct…
Big GPUs don't need big PCs
The Energy Lie
'The Energy Lie' written by davepermen
Why do AI models use so many em-dashes?
If you asked most people to name a defining feature of AI-generated writing, they’d probably say the em-dash — like this. Language models use em-dashes so much…
UWP App for Kiosk Mode
'UWP App for Kiosk Mode' written by davepermen
Netzwerk Reset auf Windows 8
'Netzwerk Reset auf Windows 8' written by davepermen
AI detection tools cannot prove that text is AI-generated
--
Writing for AIs is a good way to reach more humans
There’s an idea going around right now about “writing for AIs”: writing as if your primary audience is not human readers, but the language models that will be…
If you are good at code review, you will be good at using AI agents
Using AI agents correctly is a process of reviewing code. If you’re good at reviewing code, you’ll be good at using tools like Claude Code, Codex, or the…
The whole point of OpenAI's Responses API is to help them hide reasoning traces
About six months ago, OpenAI released their Responses API, which replaced their previous /chat/completions API for inference. The old API was very simple: you…
Tiny Game Engine Stuff
'Tiny Game Engine Stuff' written by davepermen
Grok is enabling mass sexual harassment on Twitter
Grok, xAI’s flagship image model, is now being widely used to generate nonconsensual lewd images of women on the internet. When a woman posts an innocuous…
To get better at technical writing, lower your expectations
Technical writing is a big part of a software engineer’s job. This is more true the more senior you get. In the limit case, a principal or distinguished…
GPT-5-Codex is a better AI researcher than me
--
Benchmarking NVENC video transcoding on the Pi
Thompson tells how he developed the Go language at Google.
Kenneth Lane Thompson, winner of the Association for Computing Machinery's A.M. Turing Award, discusses development of the Go programming language at Google in the early 2000s, its relationship the company’s cloud computing platform, and its eventual success a…
Was Daft Punk Having a Laugh When They Chose the Tempo of Harder, Better, Faster, Stronger?
Was Daft Punk Having a Laugh When They Chose the Tempo of Harder, Better, Faster, Stronger?
2026 will be my year of the Linux desktop
The meme will no longer be a dream.
Conic Sections
Clicks Communicator: the ultimate communciation companion
Clicks Communicator is phone purpose-built for taking action and communicating in a noisy world with deeper context, versatile input and greater control in a compact design.
A Two-Week Experiment – Blue Lemon bits
After many years as a Mac user, the last couple of years I’ve become increasingly frustrated with the macOS software direction and the continuous erosion of the historically vibrant indie developer ecosystem.
Giving Your Agent Eyes is not Enough
AI coding agents can build UI. But they can't verify it. Verification requires comparison—your agent doesn't remember anything.
Investigating and fixing a nasty clone bug
Recently I found myself battling with another nasty bug. It took me several hours to understand what is happening, and once I found it, it turned out that the cause of the bug is relevant to the Ergonomic cloning initiative that is currently being discussed a …
Year of reading
2025 was the Year of Discomfort for me. I found myself getting too comfortable with various facets of life. I wasn’t pursuing or sticking with things that made me uncomfortable. I think, and have been told, this comes with age. It’s the same sort of discomfort…
Why it matters - Jukka Niiranen blog
We don’t write just to deliver a payload of importance. We write to think. The sudden flood of "why it matters" formula in written content is not a positive sign for this.
The CEO cost function
Looking for an Algorithmic Executive
Getting Real With LLMs - Gilad Peleg
It's hard to tell from the internet what can one do with LLMs at their workplace. This blog tries to cover the current options and challenges
Representing Hierarchies
published on Jan 01 2026
Solving Agent Context Loss: A Beads + Claude Code Workflow for Large Features - JXØ
How external task state with beads transformed my AI coding workflow from constant babysitting to autonomous execution of 17-task epics. Includes the complete skill set to replicate this workflow.
Rent a Mac M4 Mini and Access It via SSH from Linux
I needed a Mac to run a Laravel test suite. On my ThinkPad T15g with Docker, the tests took so long I gave up after 2 hours. My client said it takes 5 minutes on their M4 Mac. I decided to rent one…
Why I switched away from Zig to C3
LowByteFox's Personal Website
The Next Thing Will Not Be Big
Disruption, too, will be disrupted.
Vibe coding killed Cursor
Cursor is dying because cost-optimization forces models into tunnel vision. RAG agents fail because they only see what they search for. The superior workflow for 2026 is massive context windows (gemini 2.5 pro) and manual control. Stop letting agents hide code…
Common prefix skipping, adaptive sort
The patent expired for US7680791B2 . I invented this while at Oracle and it landed in 10gR2 with claims of ~5X better performance vs the pr...
An experiment in vibe coding
For the holidays, I gave myself a little experiment: build a small web app for my wife to manage her travel itineraries. I challenged myself to avoid editing the code myself and just do it “v…
Try to Take My Position: The Best Promotion Advice I Ever Got
Published on 2026-01-02 by Andrew Graham-Yooll
From Web to Native: Building Apps with SvelteKit & Capacitor
A step-by-step tutorial on integrating CapacitorJS with SvelteKit. Convert your website into an Android/iOS app.
Dimensions: A Terminal Tab Manager
An interactive terminal tool that keeps tmux tabs organized into focused workspaces.
Why are weather forecasting sites so bad?
How AI and generative AI is disruptive for consumer weather forecasting services.
What I Actually Want To Say in Tech Interviews
Honest answers to common interview questions in the tech field.
A Socialist Now Runs New York: What History Predicts
NYC elected socialist Zohran Mamdani as mayor in 2025. Now sworn in, historical patterns show why societies turn to socialism and what the real solution is.
The Developer Is Dead, Long Live the Designer
With the advent of coding agents and their improvement, developing software as we have been doing it is, at least in many fields, already going the way of the Dodo. But bear with me for a second while I try to explain why I think this is great for developers a…
My setup, and why tmux workflow is better than ever | Raine Virta
Keybindings, popup tools, and workflows for a terminal-centric setup
Was it really a Billion Dollar Mistake?
TL;DR null pointer dereferences are empirically the easiest class of invalid memory addresses to catch at runtime, and are the least common kind of invalid memory addresses that happen in memory unsafe languages. The trivial solutions to remove the “problem” n…
MCP works great — until you actually ship.
Why MCP isn't ready for production agents, and what to do instead.
POSSE
POSSE is an abbreviation for Publish (on your) Own Site, Syndicate Elsewhere, the practice of posting content on your own site first, then publishing copies or sharing links to third parties (like social media silos) with original post links to provide viewers…