Tagged
8 posts
Code has become cheap, but having an agent generate an app doesn't make it a product. A new series reviewing SaaS apps that launched and failed to convert, or sit abandoned — and what went wrong.
How to use LLMs as part of your writing workflow.
A decade of semantic versioning stability means AI agents trained on 2020 code still produce working 2026 code
Code patterns that maximize AI agent success—and the anti-patterns that confuse them
Immutability, pattern matching, and pure functions make Elixir ideal for AI coding agents—the data proves it
Using Instructor for structured LLM output and building AI features in Elixir
How project structure, tooling, and practices evolve when AI agents are central to development