I’m a generalist with a love for learning about and using new technology.
I write about SEO, business, and how to ship products with AI that have craft, quality, minimalism and simplicity at their core.
Posts
2026
- Cold outbound guide for high DR backlinks
- Delegate computer use from Amp to Codex via Herdr
- One-shot a Prince of Persia level with Opus 5.5
- How to stop Google indexing your trycloudflare tunnel URLs
- Analyze your X posts with an AI agent
- Google's contentEffort score rewards enriched data
- Basic SEO brings 50 visits a day in week one
- Use public carrier data for a freight risk tool
- Clean and enrich data before publishing SEO pages
- Build an interview prep guide with ChatGPT
- Weigh paid backlinks against the risk of a penalty
- Ask Claude to remove mannered prose
- Run Ahrefs reports with Codex and Chrome
- Speed up programmatic SEO with Astro incremental builds
- Turn a transcript into an AgentOS prompt with Grok
- Plan and review with GPT, build with Composer
- Give your personal agent reliable local tools
- Override Codex and Claude Code system prompts
- Plan parallel builds with Claude agent teams
- Scan installed agent skills for security risks
- Limit agent-written files with a LOC linter
- Get a URL as Markdown
- Use Cloudflare workers for secure access
- Find and install agent skills from the CLI
- Fix Codex issues with Opus
- Clone a repo into separate folders for parallel agents
- Steer Codex while it is still working
- Improve AI loop with simple iteration
- Move PHP sites to Cloudflare with Codex
2025
- Keep programmatic pages indexed by making them useful
- Give the next agent context from your session log
- Build one feature at a time with AI
- Ask GPT to finish each step without stopping
- Use a few agents at a time
- What the Codex performance report found
- Force Codex to ask for permission in sandbox
- Review Git changes and security with Droid subagents
- Improve Codex CLI conversations with a global prompt
- Give agents unfinished functions and TODO comments
- Break an AI loop by comparing five refactors