A behavioral scientist who can't change her own behavior.

An n=1 adoption problem.

I listen to pods like The AI Daily Brief, Lenny’s Podcast, and (when I’m feeling especially cognitively competent) Lightcone.

I mentally gulp YouTube video after YouTube video where people talk about problems they’ve spotted in their lives and turned into (non-insignificant) business opportunities.

I read Hacker News instead of watching Love Island (you already know this if you’ve seen my Hinge profile).

Claude Code, CLIs, and MCP won’t let me be.

But when it comes to actually using these tools and frameworks, I feel like I’m sitting at the end of the pool with my legs hanging in, waiting for the right moment to jump in, which seemingly hasn’t presented itself.

I’m promptmaxxing from the backseat.

I rarely start projects. When I do though, I have a lot of fun and really enjoy it (clearly, there is both personal and societal importance to my 2026 version of FMK, affectionately reimagined as Date, Mate, Donate, that is currently hosted locally on my machine).

… But, I tend to not start because I don’t know what to work on at any given moment. I usually default to just watching YouTube videos instead, waiting for my automation aha! moment to arrive (as above pool metaphor alludes, it has not).

Every time I sit down to “learn AI,” I’m trying to roll myself forward from a cold start and need to decide what to work on. This decision is cognitively expensive, because I’m usually doing this in the evening from my sofa, when my brain’s gas tank is running low on fuel. YouTube is my autopilot mode; I can feel like I’m getting somewhere, and it just plays.

I feel compelled to work on projects that might have market opportunity, when I know, in actuality, I just need to work on projects out of my own self-interest. I think this has, low key, been one of my biggest blockers in trying new AI tools and tinkering.

I can’t seem to give myself permission to do something unproductive. Everything I invest more than a few minutes in has to earn its place by producing output.

Aspirational YouTube content is 1000% fueling this (sine-wavingly frantic) sense of: Find something to build now. Build it now. Do it now.

YouTube gives me a destination with no map; passive consumption masquerading as learning at its very finest.

If only I were playing instead of spectating.

Yet, I can’t find a product where design choices remove the decision of what to practice today, choices that alleviate the behavioral (hi, it’s me) problem by putting something in front of me the moment I show intent to engage.

Existing AI products (courses, newsletters, cohorts) assume I’m showing up with intent and energy. But usually I’m tired and pseudo-horizontal on my sofa (honesty is the best policy).

So, I’ve been cooking up a project that takes away the initiation decision, where a fleeting moment of wanting to learn more about AI is enough.

The product does the heavy lifting.

Writing while I build, which truthfully just feels like writing while I learn.

Next up, I’m talking about why I often have a hard time finding (all of these so incredibly obvious) AI use cases in my own work and life.

Clickbait-y pattern-matching examples make this seem simple, making me feel silly when I can’t connect the dots. And I kind of want to talk about it.