laiso

Web and mobile application developer. I write about AI agents, developer tools, and TypeScript.

What I Do

• Technical writing — AI agents, coding agents, serverless architecture, and developer tools
• Book — Author of a technical book on building AI agents with TypeScript
• Open source — Building niche developer tools and benchmarks

Current Focus: ts-bench

I maintain ts-bench, a benchmark CLI for comparing AI coding agents on TypeScript workloads.

ts-bench v2 evaluates real-world SWE-Lancer tasks from a large TypeScript / React Native codebase. It compares practical agent/model combinations such as Claude Code, Codex CLI, Copilot CLI, Gemini CLI, opencode, and others.

Running these benchmarks requires API credits, subscriptions, CI compute, and ongoing maintenance. Sponsorship helps me keep the leaderboard updated, add more agents and models, and expand the benchmark set.

Sponsorship

Looking for supporters for my technical writing and OSS work, especially ts-bench.

Goal: $50/month — helps cover benchmark runs, CI compute, subscriptions, and research materials.

$1/month

• Access to a private repository with draft articles, book chapters, and early benchmark notes
• Early access to new tools

$5/month

• All previous benefits
• Early access to ts-bench analysis and benchmark result notes
• Priority response to questions and feedback

$10/month

• All previous benefits
• Direct feedback on draft content
• Monthly update newsletter about ts-bench, AI agents, and technical writing

Links

• ts-bench: https://github.com/laiso/ts-bench
• ts-bench v2 article: https://laiso.github.io/ts-bench/articles/ts-bench-v2-coding-agent-benchmark.html
• Blog: https://blog.lai.so/
• X: https://x.com/laiso