Evil Martians make the impossible possible: launching a new Aptos Network website in just one month!

Research-backed insights into what AI-assisted developers can actually deliver: when productivity improves, when it doesn’t, and how setting realistic expectations means best quality results.

AI can generate a full dashboard in an afternoon. But speed doesn't eliminate design decisions—and building the wrong thing faster is an expensive mistake. Here's what design work looks like across four stages, from framing problems to learning after launch.

Build optimistic UI in Rails with Inertia by updating props before requests complete. Learn the replaceProp pattern, automatic reconciliation, rollback behavior, and history caveats using a kanban board example.

Prefer SVG over PNG, trust browsers to downscale, drop obscure formats—the ultimate, exhaustive guide to favicons for modern web. Includes steps for static HTML and Webpack.

Read how to accurately detect Safari and iOS versions using WebKit feature checks, behavioral tests, and selective UA hints to gate features safely and avoid breaking UX.

I spent the year in San Francisco, having face-to-face conversations with founders you already follow. Here are the top pieces of advice to take into 2026.

Developers don’t need more tools—they need fewer, higher-quality ones. This piece distills research and field evidence into six principles for faster, calmer, trustworthy devtools in 2026.

We want to make Inertia.js more popular within the Rails community, so we've started a project to provide both Rails-specific community docs and a set of tools for simply integrating Inertia into Rails apps.

Real-time LLM streaming sounds simple—until it isn’t. A Rails-focused look at hidden pitfalls, AnyCable’s fixes, and the future of durable streams.

After seeing dozens of vibe-coded projects up close, it's clear the gap between a prototype and product ready for users is wider than most expect. Eventually, progress becomes impossible without an experienced developer. We explore why that is, and I share a real-world case study to illustrate.

Here's why we scaled up our San Francisco Ruby meetups to SF Ruby Conf with 400+ Ruby engineers for three days of technical talks and startup demos.

We've experimented putting the essence of a Martian engineer into AGENTS.md. Learn the story, then try it on your next AI-assisted project!

A deep dive into Ruby parsing and unparsing, the transition from the Parser gem to Prism for Ruby 3.4 support, and how tooling adapts to evolving AST formats using Unparser as a case study.

Tools like Bolt, Dazl, and Lovable are leading the charge with AI coding tools. But there are design patterns that can improve vibecoding workflows even more, reducing friction, and elevating AI tool UX into intuitive, creator-friendly experiences that appeal to users on a massive scale.

Learn how to optimize payment forms with frontend techniques to reduce abandonment and increase conversion rates. From performance optimization to UX design, discover actionable strategies that can boost conversion by up to 35%.

React dominates with 88.6% of startup funding, but 85% of these projects are dead. We analyze funding patterns, GitHub activity, and ecosystem health across React, Vue, Angular, and Svelte.

How we helped Whop migrate their high-traffic Rails app from PostgreSQL to PlanetScale MySQL without downtime or development pauses. Read about dual-database setups, schema quirks, and the lessons learned.

SF Ruby and Evil Martians are excited to invite you to our premier event: the San Francisco Ruby Conference.

Debugging AI agents shouldn’t mean digging through endless JSON. AgentPrism, an open-source React component library by Evil Martians, turns raw OpenTelemetry traces into clear visualizations, instantly revealing loops, errors, and cost spikes. Built with Quotient AI, AgentPrism helps teams ship reliable, production-grade AI systems up to 5x faster.