about

James L. Cowan Jr.

@jameslcowan

Founder of Kleto.

RSS Feed

articles

  • Pydantic: The Open Source Layer Quietly Running the AI Economy OpenAI and Anthropic get the headlines. Pydantic ships the work. With 10 billion downloads and the entire AI stack depending on it, the most important tool in AI is the one nobody is talking about.
  • Karpathy Was Wrong: OpenClaw Still Outruns Its 5 Real Alternatives Karpathy called OpenClaw a "vibe coded monster," yet production work tells a different story. I keep OpenClaw in the lead and show where each challenger actually helps.
  • OpenClaw 2026.2.23: The Agent Browser Upgrade Most "agents" are just chat interfaces that hallucinate action. The latest OpenClaw release gives them actual eyes and hands.
  • The YC S26 Deadline Just Closed. Here's What Separates the 1.5% From Everyone Else. What YC is funding right now, what kills your application, how to show traction without big numbers, and interview tips from founders who made it through.
  • OpenClaw: What It Is and How to Get the Most Out of It A founder focused guide to OpenClaw, why it matters for SMBs, and how to turn chat based control into real operational leverage.
  • Why I Skipped the CMS I evaluated Decap, Directus, Strapi, Sanity, and Contentful before deciding that none of them were worth the cost for a content site run by one person.
  • Documentation That Scales: Constitution, Contracts, and Runbooks How I structured documentation so a team of one (plus AI agents) can make consistent decisions fast, without the docs becoming a liability.
  • Fixing Navigation and Analytics: When Your Data Lies About User Behavior A behind-the-scenes update on fixing shadow-DOM navigation edge cases, restoring back/forward swaps, and making analytics actually reflect what readers do.
  • HAL: Cutting 100-300KB of JavaScript by Moving Routing to Build Time A performance case study showing how build-time link rewriting eliminates the SPA routing tax. Faster pages, less code, same user experience.
  • HAL: Build-Time Link Rewriting That Makes Navigation Feel Instant How a simple build-time pass gives you SPA-speed navigation on a static site - no router library, no framework runtime, no ongoing cost.
  • Web Components + HTMX: A Lean Architecture for Content Sites That Ship How I use native Custom Elements and HTMX to build a site that feels like an app, costs nothing to host, and doesn't require a framework upgrade every 18 months.
  • Zero-Server Analytics: How I Replaced a SaaS Bill with Netlify Functions and GitHub Building a self-running analytics pipeline with Netlify Functions, Blob Store, and GitHub commits - zero monthly fees, full data ownership, and nothing to babysit.
  • Mobile-First Layout That Ships: How PrimaryLayout Solves Real UX Problems How a single custom element handles fixed headers, safe-area insets, recommendation cards, and loading feedback on mobile - without a framework or a UX team.
  • Why I Built My Own Analytics Pipeline (And What It Actually Costs) A lightweight, privacy-first analytics system that runs on Netlify functions and JSON files. No vendor lock-in, no monthly SaaS bill, and it took a weekend to build.
  • The Static Site Playbook: Shipping a Content Product on a Near-Zero Budget How I built a production content site with Markdown, a build script, and a 35KB JS budget - and why the boring choices are the ones that compound.
  • How This Site Works: Architecture for a One-Person Team The technical architecture behind a fast, SEO-friendly content site built by one person. SSG, Web Components, HTMX, and the business logic behind each choice.
  • Progressive Enhancement with HTMX: Ship Fast, Degrade Gracefully How HTMX lets small teams ship dynamic-feeling sites without the cost and complexity of a full JavaScript framework.
  • Web Components as a Business Decision Why Web Components are a pragmatic choice for small teams that need reusable UI without framework lock-in, hiring constraints, or migration risk.
  • Less JavaScript, More Leverage: Why I Ship With a 35KB Budget Cutting JavaScript isn't a purity contest - it's a business decision that compounds in speed, cost, reliability, and team velocity.
  • Static Site Generation: The Business Case for Pre-Rendered HTML Why pre-rendering HTML at build time is the highest-leverage hosting decision for content sites - faster pages, lower costs, and less operational risk.