Developer and writer A.N. Childress recently published a candid, humor-driven account on DEV.to detailing their weeklong ordeal redesigning a personal portfolio site with help from multiple AI coding assistants—and the results were less than polished. The post, titled 'I Asked for a Portfolio but Got a Filing Cabinet,' has resonated with developers who recognize the all-too-common experience of asking an AI for one thing and getting something completely unrelated back.

When Your AI Friend Doesn't Listen

Childress describes spending nearly seven days locked in arguments with various AI tools, each promising to help modernize their portfolio but delivering frustration instead. The developer notes that most of this time was consumed by troubleshooting why implementations weren't working as expected—a scenario that will sound painfully familiar to anyone who's spent hours debugging code they didn't write from scratch. The post frames these interactions almost like dealing with stubborn colleagues who insist they're right despite overwhelming evidence otherwise.

The Portfolio Problem Nobody Talks About

Building a personal portfolio site sounds straightforward in theory: showcase your work, keep it clean, make it fast. But as Childress discovered, even 'simple' tasks become complex when you're working with AI tools that interpret requests differently than intended. The developer acknowledges they were genuinely trying to get back into writing, using this project as motivation—but the technical hurdles nearly derailed the whole endeavor entirely.

Why This Story Matters for Dev Tool Users

This isn't just a funny anecdote; it highlights real friction points in the developer experience with AI-assisted coding. When tools promise productivity gains but deliver unexpected behavior, developers end up spending extra time debugging the tool's output rather than writing their own code. For teams evaluating AI coding assistants, stories like this serve as important reality checks about where these tools excel and where they still struggle.

Key Takeaways

  • AI coding assistants can interpret requests in unexpected ways that don't match developer intent
  • Even portfolio-sized projects can become multi-day debugging sessions when working with AI tools
  • The gap between 'this should work' and 'this actually works' remains significant in current AI offerings
  • Personal anecdotes from developers provide valuable signal beyond vendor marketing claims

The Bottom Line

The developer experience with AI coding assistants is still maturing. Childress's story isn't an indictment of the technology—it's a reminder that these tools require careful steering, realistic expectations, and sometimes a lot more patience than advertised. If you're planning to lean heavily on AI for your next project, budget extra time for the inevitable 'wait, that's not what I asked for' moments.