Working in a focused way with AI is… surprisingly hard.

Not because the tools are bad—quite the opposite. We’re living in a moment where AI tools are genuinely powerful, and there are many areas where they already work extremely well. The challenge is navigating this abundance: what problem are you actually solving, which tool should you use, and how do you avoid drifting along the way?

Exploration Is the Work (Up to a Point)

Today was very much an exploration day.

I spent time looking into Notebook LM, and it’s genuinely interesting. If you’re curious about how it works and what it’s good at, this video is a solid introduction and well worth watching:

👉 https://www.youtube.com/watch?v=OdCmZvPdr4s

This kind of exploration is valuable. It builds intuition. You start to understand what different tools are good at—and just as importantly, what they’re not good at.

The tricky part is knowing when exploration quietly turns into distraction.

AI and the Travel Planning Rabbit Hole

I also started planning a trip. Travel planning is one of those areas where it’s hard to even know which questions to ask upfront. Traditional online research has always felt fragmented and inefficient.

This time, I approached it with AI:

• I kicked off a deep research session in ChatGPT
• Asked follow-up questions to everyone joining the trip
• Let both ChatGPT and Gemini generate their own infographics

It was fun.
It was creative.
It looked great.

But in terms of actual decision-making? Minimal value.

That’s an important lesson when working with AI. It’s incredibly easy to generate artifacts—documents, visuals, summaries—without moving any closer to a real outcome.

A Very Real Problem: Laundry

Then there’s the most grounded problem of all: laundry.

Our current laundry setup no longer scales. When we were two people, it worked fine. Now we’re four. Everyone is active. Everyone trains. Laundry appears out of thin air.

So I asked ChatGPT to do a deep research on laundry basket solutions. I also dropped an image into a folder on my computer and sent the same prompt to Claude Code.

And honestly—I’m still impressed by how confidently Claude Code operates:

• moving around the computer
• navigating files
• browsing the web

And here’s the part that made me laugh:

I genuinely did not think I’d be shopping from a terminal in 2026.
Yet here I am, evaluating laundry baskets through AI agents and command-line workflows like it’s completely normal.

A Lighter Takeaway

Working with AI today is less about rigid plans and more about learning the landscape. Some days are about producing clear outcomes. Other days are about understanding tools, testing boundaries, and discovering unexpected workflows.

Even if that occasionally means ending up in a terminal… buying a laundry basket.

And honestly? That’s kind of fun.