About AI Vagabond

Real experiments. Honest results. No hype.

The Short Version

I spent the last year building AI agents—some that worked brilliantly, others that failed spectacularly. AI Vagabond is where I share what I learned while wandering through this space, so you don't have to make the same mistakes.

Why This Exists

The AI space is full of two kinds of content: academic papers nobody reads, and breathless Twitter threads that never tell you how to actually do anything.

I wanted something different. Practical guides based on real testing. Honest assessments of what works and what doesn't. Code you can actually run.

Every article here is based on something I built, broke, or debugged at 2 AM. When I compare models, I run actual experiments. When I write tutorials, I follow my own steps to make sure they work.

What I Cover

🤖 AI Agent Deep Dives

How agents actually work, from architecture to production gotchas.

⚖️ Model Comparisons

Real benchmarks on real tasks. Not marketing claims or synthetic benchmarks.

🔧 Build Tutorials

Step-by-step guides that result in working code, not broken demos.

đź’ˇ Honest Takes

When something doesn't work or isn't worth the hype, I'll say so.

My Background

I've been building software for 10+ years. Started taking AI agents seriously in early 2024, when they stopped being demos and started being useful.

I've built agents for email automation, research, code review, and various client projects. Some worked great. Some cost me a $500 API bill and a weekend of debugging.

Those experiences—especially the failures—are what I write about here.

My Approach

  • Test everything. Claims without evidence are just opinions.
  • Show the failures. Knowing what doesn't work saves time.
  • Be specific. Vague advice is useless advice.
  • Update when wrong. The field moves fast. Articles get updated.

Want to Build Something?

Start with an article that matches your goal, or browse everything.

Read the guides

Get in Touch

Got feedback? Found an error? Have a topic you'd like me to cover? I'd love to hear from you.