Johanna Larsson argues that the push to maximize LLM agent productivity in software development is a trap, as startups win by thinking differently and producing value, not by outproducing incumbents w…
Johanna Larsson argues that users, not AI tools, are accountable for outputs generated by large language models, comparing LLMs to toasters or compilers. Larsson warns against offloading responsibilit…
Developer Johanna Larsson published a workaround for Claude Code users frustrated by the AI's repetitive use of phrases like "load-bearing" and "honest take," using a custom script and the MessageDisp…
Developers are adopting a new approach to using large language models for coding, treating them as on-demand programming books rather than autonomous coding assistants. The method involves having LLMs…
The widespread use of large language models to generate written content is violating an implicit social contract between readers and writers, according to a new analysis from Oxide Computer. The compa…
The article compares two consistent hashing approaches in Elixir: Discord's ExHashRing, which requires managing stateful processes, and Rendezvous (Highest Random Weight) hashing, which is a stateless…
**Summary:** The article introduces *bunnyx*, an open-source Elixir client library for interacting with the bunny.net CDN API. Developed by the author for use in their uptime monitoring project Larm…
The author, Johanna Larsson, describes her unconventional path into programming, which began with creating a simple website for her father's business using HTML and CSS, sparking a lasting passion for…
The article describes the author's successful setup for running local AI models on an M4 Mac with 24GB of memory, specifically highlighting Qwen 3.5-9B (Q4 quantized) as the best performing model at ~…
The article is a satirical guide by Johanna Larsson on how to deliberately exhaust one's Claude AI weekly usage limit through wasteful practices like running multiple redundant sub-agents, installing …
The article describes the author's decision to move their blog from Cloudflare to bunny.net, citing concerns about over-reliance on a single US corporation and a desire to support a European alternati…
This article summarizes how the author built a blog using Elixir and the Phoenix framework, with blog posts generated from Markdown files via NimblePublisher and compiled at build time for performance…
The article describes a two-mode workflow for using Claude Code: planning and execution. Users must first build shared context by loading relevant documents and code, then research the task thoroughly…
The article argues that project failures often stem from an inability to identify and ship the smallest possible feature that delivers user value. It advocates for "ruthless prioritization," emphasizi…
The article argues that software estimates should be viewed as tools for fostering communication and transparency rather than as rigid deadlines or contracts. It emphasizes that good estimation create…
The article argues that software engineers often mistakenly prioritize uninterrupted coding time over their true job of creating impact and delivering value. It emphasizes that the most valuable work—…
The article argues that defensive programming, which uses excessive conditionals and error handling to prevent failures, actually undermines software resiliency by masking errors and creating systems …
The article argues that developers should aim to merge at least one pull request (PR) per day by breaking work into the smallest possible valuable pieces, rather than working longer hours. It explains…
This article by Johanna Larsson on the AppSignal blog explains how to build a distributed rate limiter in Elixir using the HashRing library. It covers the implementation of consistent hashing to evenl…
The article describes the development of Hex Diff, an open-source tool for comparing Hex package versions. It explains how the project uses LiveView for a responsive search interface, reads package da…