A KVM Setup That Just Works

NPCs Powered by Local LLMs

Career Pivots and Monte Carlo Simulations

PC Build 2025

Quick MITM Proxy Toggle for macOS

Teaching Claude How to Make Memes with MCP

An Overengineered View Counter

CT Logs, Subdomain Enumeration, and Every Vercel App

Running the Numbers: Data on my First Marathon

Software Multiple Compression

Concurrency and Multithreading in Python

Hacking the Fallout Terminal Minigame

How to Get Live Crypto Prices in Google Sheets

Building with the FRED API and US Treasury Data

A Greedy Approach to Solving Wordle

Solving Wordscapes with Brute Force

Dice Sum Probability from Catan to Craps

Hello World

Application Load Balancer announces integration with Amazon VPC IPAM

grantslatton.com

walzr.com

Calling All Hackers

joshpigford.com/projects

i tried to sell dingboard to X. they gave me a job instead

A clever hack of the Firefox address bar that demonstrates just how much functionality you can pack into this seemingly simple browser feature.

Wordle In The Firefox Address Bar

AWS announces Amazon-provided contiguous IPv4 blocks

Cognitive load is the amount of thinking a developer needs to do to solve a problem. This article makes the strong case for development practices that minimize cognitive load. This concept resonates with my strongly, and is a good reminder to keep things simple.

Cognitive load is what matters

On becoming competitive when joining a new company

Amazon VPC IP Address Manager now allows you to Bring Your Own ASN to AWS

Hackers and Painters

Well, how did I get here?

New – AWS Public IPv4 Address Charge + Public IP Insights

blog.jonlu.ca

This was my glimpse of GPT-3. I remember how mind blowing this was in 2020 to see natural language to code for the first time. Looking back on this half a decade later, what seemed magical then has become trivial with today's tools. I experimented with my own early GPT apps in college ( [infrastructure-as-text](https://infrastructure-as-text.vercel.app) and [Versify](https://versify.vercel.app), both inactive) but underestimated just how rapidly AI would improve.

A button that looks like a watermelon

The Economics of Being a Founding Engineer

jlowin.dev

The length of task that AI agents can complete independently has been doubling roughly every seven months since 2019, and more recently doubling every four months. If this trend holds, then by the end of the decade, AI agents will be able to complete tasks that would take human developers weeks of work. This is one of the most important trends in technology and something to monitor closely over the coming years.