There's a specific kind of anxiety that comes with being in AI right now.

It's not the fear of being left behind. It's the accumulation of "I should really learn that"... the Substack you flagged, the YouTube video someone texted you, the X thread with 200 likes you saved and never opened.

For a while, my "learning system" was a graveyard of browser tabs and starred emails. I knew things were there. I just couldn't find them. And the more they piled up, the less I actually learned because opening that pile felt like work before the work.

So I built a queue.

Here's what's in it right now: how to set up Claude correctly, Claude Cowork, a 20-minute YouTube video on Code and Paperclip, a 9-minute one on Claude and iMessage, Perplexity's new computer mode, two Substack pieces on delegators and agentic tools, and the Lenny newsletter on Openclaw.

I'm familiar with all of these tools; I use several of them daily. But knowing how something works and knowing how to teach it are two different skills. I'm not learning these for me. I'm learning them so I can figure out the simplest, most compelling way to bring a room of skeptical executives along with me. I take insights from founders, engineers and practitioners.

When I have 12 free minutes between calls, I'm not scrolling for something to read. I'm filtering my database by "Video" and picking the 9-minute one. When I'm at my desk with focus time, I go deeper. The queue shapes what I learn when, not just whether.

The system has three rules:

1. If it doesn't go in the queue, it doesn't exist. No more starred emails, saved tweets, or "I'll remember this" promises.
2. Nothing lives in the queue for more than two weeks. If I don't care enough to read it by then, I didn't care enough to add it.
3. The queue caps at 15 items. Adding something new means asking if something old still belongs there.

That cap is the uncomfortable part. It's also the most useful. It forces prioritization at the moment of capture, when your interest is highest and your judgment is clearest.

Inbox zero for learning isn't about reading everything. It's about never losing track of what you decided was worth reading and being honest when that decision has expired.

My queue right now tells me something: I'm deep in Claude tools and agentic infrastructure. That's a reflection of where my client work is heading. The queue has become a mirror.

What does yours look like?

Alex

P.S. I built this in Notion, but a simple spreadsheet works just as well.