Using LLM for confidence scoring is a bad idea
LLM-based agents can do a lot of tasks. Or, at least, they can appear to do them. For example, they write code, summarize papers, plan trips, and fill in forms. If you ask one for a probability or a confidence score, it will give you one of those too. I have seen many agent designers use LLMs to generate confidence…
I turned my RSS feeds into an e-ink newspaper to stop reading on my phone
I built a small tool that turns my RSS feeds into an e-ink newspaper, so I can read my favorite blogs on a pocket e-reader instead of my phone.
Building a TB-303 from Scratch
Code music in the browser.
Mxc: Microsoft Execution Containers version 1.0.0
Agents are unlocking enormous productivity gains for customers, but their ability to work across files, networks, and applications can introduce new security risks. This can leave customers feeling like they only have two choices: give agents unrestr
We have ipinfo at home or how to geolocate IPs in your CLI using latency
TLDR: I made a CLI tool that can resolve an IP address to a country, US state and even a city. https://github.com/jimaek/geolocation-tool It works well and confirms ipinfo's findings. Recently, I read how ipinfo finally proved what most technical people assumed: VPN providers don't actually maintain
Async/Await on the GPU
GPU code can now use Rust's async/await. We share the reasons why and what this unlocks for GPU programming.
A decades-old bug in Knuth's long division (TAOCP Vol II, Algorithm 4.3.1D)
A long division story
Image Compression
Exploiting the quirks of human vision to make images smaller, without looking awful.
“Tweet” and the bird logo apparently enter the public domain
This case involves Project Bluebird, a social media service previously named “twitter.new” and renamed to tweet.app immediately after this ruling (for reasons this post makes obvious). Project Bluebird’s service is designed to reimagine the Twitter service that Musk imploded when...
Three important steps in my maturation process
My father passed recently, and he was twice my age. I am approximately the same age that he was when I was born, and I am now “the old generation” - there’s no one left in the generation above me.
If DSPy is so great, why isn't anyone using it?
Any sufficiently complicated AI system contains an ad hoc, informally-specified, bug-ridden implementation of half of DSPy.
Web Security is Too Hard
It started innocently enough. I saw a tweet about a new product offering from one of my favorite companies, Cloudflare. Neat! I clicked through to the site and there it is: And huzzah!, my preferre…
Claude Platform on AWS
The Claude Platform on AWS is now generally available, offering a new way for AWS customers to access the full set of Claude platform features with AWS authentication, billing, and commitment retirement.
Chess puzzle I found in my dad's old book
Drag or tap a piece, then choose a square. Tap or click a placed piece to remove it.
FLUX.2 [Klein]: Towards Interactive Visual Intelligence
Introducing FLUX.2 [klein], our fastest image models. Unified generation and editing in under a second, on hardware you already own. Apache 2.0 on 4B.
The "Windows XP Box" (2003)
I needed a small Windows XP machine and a Mini-ITX board was the obvious choice. So I decided to build my "Windows XP Box" in a Windows XP box. The external dimensions of the box are a tiny 243mm x 200mm x 48mm.
WATaBoy: JIT-Ing Game Boy Instructions to WASM Beats a Native Interpreter
A Game Boy emulator runs faster than a native interpreter by dynamically recompiling SM83 to WebAssembly, effectively making JIT possible on iOS.
Fable Took My Job
Fable is cool yo, it takes jobs yo, - Dario "steve" Altman
We won't "just hang out"
On human connection after the Singularity.
Where did all the starships go?
In this Weekly Chart, we look for sci-fi books and their changing titles.
Maladaptive Frugality
I recently decided to, finally, have my iPhone fixed, only to realize a few hours later that my AppleCare could have covered it. I was in a low mood until my partner suggested that I was robbing myself of a good decision. The iPhone needed fixing, and procrastinating on it wasn’t useful. Deciding to do […]
Just Use Go
Hey, dipshit. You know what compiles in two seconds, deploys as a single binary, and doesn't shit itself when a transitive dependency gets yanked from npm at 3am? Go. The same way HTML has been sitting there since the dawn of the goddamn internet waiting for you to stop overcomplicating the frontend, Go has been…
Let's compile Quake like it's 1997
The first batches of Quake executables, quake.exe and vquake.exe were programmed on HP 712-60 running NeXT and cross-compiled with DJGPP running on a DEC Alpha server 2100A. In June of 1996, having shipped their title but concerned with NeXT stagnation, id Software switched their development stack.
Light Mode InFFFFFFlation
Back in the day, light mode wasn’t called “light mode”. It was just the way that computers were, we didn’t really think about turning everything light or dark. Sure, some applications were often dark (photo editors, IDEs, terminals) but everything else was light, and that was fine. What we didn’t notice is that light…
Direct Win32 API, weird-shaped windows, and why they mostly disappeared
The nice thing about Win32 is that it does not try to talk you out of any of this. It just gives you the messages, the handles, the drawing APIs, and enough rope to ...
Claude now creates interactive charts, diagrams and visualizations
Ask Claude to explain a concept or analyze your data, and it can respond with interactive charts, diagrams, and visualizations — rendered inline as part of the conversation.
Only 17% of all 64-bit Integers are products of two 32-bit integers
In software programming, the product between two integers is often computed to a fixed number of bits with overflow. Consider 8-bit integers. If you multiply 127 by 127, you get back the number 1 as an 8-bit unsigned integer, with an overflow. The actual full product is 16129. To represent 16129, you typically use 16…
The bootstrapper's EU stack for under €10 per month
How close to zero can you get on a 100% EU infrastructure stack? A bootstrapper-friendly setup for under €10 a month.