LLM output is malicious user input
The most common programmatic interaction with Large Language Models (LLMs) and LLM APIs (ChatGPT, Claude) is to give it some natural language instructions and get a shaped, specific output back.…
Using a local LLM to Automate an Android device
While most well known Large Language Models (LLMs) are closed and behind paywalls, there exist open models such as LLaMa and its derivatives, available for free and private use. A…
?, a simple CLI lookup tool
As I spend a lot of time on the CLI, I often need to look up commands, even if I’ve used them before. I like to offload memory elsewhere if…
Using threat modelling to choose a password manager
Common ways of choosing a password manager are to see what everyone else is using, search for what’s popular, or just pick something convenient. I do the same, but also…
.NET’s underrated configuration feature
My favorite kind of features are usually ones that let you start simple and still let you build powerfully on top without being overwhelming. .NET’s ConfigurationBuilder being one, is one…
The unpleasant hackiness of CSS dark mode toggles
There are two ways that websites can offer users a choice between light and dark mode. The first makes use of pure CSS and is managed natively by the browser.…
My Kobo Customizations
I recently switched from a Kindle device to a Kobo Libra 2, and have been playing around with its customization and tweaks. These are the ones I’ve found useful so…