I went to catchup with a friend last night, and between some embarrassing shots on the golf simulator and talking ideas of things we could build, I said to him, “You see Dave, this is my problem, I get excited about everything, and I genuinely believe that every idea is awesome. But that doesn't mean I can do every idea.”

That line isn't originally mine—stole it from Tristan de Montebello on Lenny's pod. It's been living in my head since.

Last week as you heard Claude Design launched. Design is my favorite component of product work, so I ran over to play with it. The idea I picked to test it was a personal product I've been sitting on to help me grow my newsletter business.

The idea is it takes just a single URL for the newsletter (www.the-diff.com) and your email address and it spins up a multi-agent driven newsletter sponsorship pipeline. Research the pub, research overlapping letters, finds advertisers already investing in this channel, drafts the pitch, runs the cold outreach in a test-driven way. End to end from two inputs and I just wait to get replies.

It looked great because Claude Design is very good, and that’s my damn weakness if it’s worth mentioning here folks. I’m a true sucker for something that looks and feels good, and when design teases the product to me, I get all giddy and I want.

In this little endeavor to simply learn Claude Design, I felt the urge rise up again, right on cue—like a Labrador, I need to run and fetch the stick! Except, I already have a ball.

The Labrador Problem isn't about the compelling feeling of chasing bad ideas. It's about ideas that feel exciting enough to be the thing you should be working on right now. The build it now or someone else will urgency. That's the pull. And I find it’s always pulling.

A few years ago the line of defense for focusing on one idea was simple: there was a real measurable cost to switching. You had an idea, you picked it, and it took time to see it to some milestone. Shiny objects were less compelling when you couldn’t just take them.

But now I feel like a builder who’s taken Compound-V and has all these new powers to make any idea happen. When ever shiny object can be summoned to existence quickly with low to no cost, that line of defense falls away. When it’s no longer about can you do it, it’s only about should you do it.

And should I is much harder to answer than can I.

I’m stealing the term from economics and bringing it to the world of building things … should I becomes FOMO-flation.

As it gets easier to make things and have working prototypes live in hours, and you hear about more people making things that are overnight successes, the rising fear and insecurity of not doing enough with what’s at our disposal and we’re missing out on the right thing.

Come on Jaryd. You’re just one niche away. One type of agent or cracked workflow that gets it right for you. One shift in focus and you might be on the money. Or maybe you just doing it wrong with just a MacBook Jaryd…why no Mac Mini?!

Because knowing I can do the thing, I hear the siren-song itching in my ear… “Are you building the right thing?”, “Are you doing enough with AI?”, and “Is this the best return on your focus?”.

This quick email is perhaps a letter to self as I try to protect my focus on the thing I’m building right now, but if you’ve ever felt the feeling of not doing enough, maybe this resonates…

And the pull to build something new…the feeling of not doing enough…it's bigger than whether to start something entirely new.

I’ve found the same thing shows up inside the thing I’m already building too. Add the feature. Add another feature. Ship the Brain Graph. If you don't do it, it's not enough.

When everyone is making things, there’s a feeling of you need to make your thing more to keep it ahead.

V1 of Little Moments had so many features for exactly this reason. AI Storyteller Coach. Rewind. Memory Race. Multiple prompt types stacked on top of each other. Six feature splash screens to show them all off. I was trying to be too much because the cost of adding anything was basically zero — and not adding felt like I was leaving something on the table.

I killed most of it and really rethought it to be simple and focused on one hook. But the feeling that drove me to build it all in the first place—that same feeling and inner conflict shows up every week about what to add next.

It’s that classic battle of the Two Wolves. Which wolf wins? Whichever one you feed.

There is no natural friction anymore. Adding a feature used to cost you a week. Now it costs you an afternoon. The friction used to prioritize for you. It doesn't anymore. You have to do the prioritizing yourself.

And that is harder than it sounds because it requires tuning your product sense, tuning out noise, and being systematic and pragmatic in how you time box and approach validating ideas.

Truly now more than ever, when more is easy, less is so much more.

When the hum and drum to do more and do different gets loud, I have a few things I imperfectly come back to.

1. Focus is the commodity and comparison is the thief of it

Everyone has the same tools now. Claude, Cursor, Expo, Supabase, PostHog, RevenueCat. The tools are not the edge. The edge is what you choose not to build. What you stay with when the next shiny thing launches. What you actively ignore.

If tools are a commodity, attention is the differentiator. And I mean my attention — on Little Moments, on the 28 real people who opened it this week, on the specific problems in front of me this week.

It’s so easy to compare someone elses’ success, idea, or setup to yours and feel inadequate and that inadequacy is what can pull you away from your focus.

2. Use that focus to finish one thing.

A lot of the people posting about their 40-agent setups and their token counters and their "I shipped 12 AI products last quarter" threads have zero users and didn’t really ship anything live. No revenue (me too though). I'd bet the number of people running multi-monitor Mac mini rigs with command center views of all their agents who still haven't shipped a finished useful product to real humans is insanely high.

Tokenmaxxing is the new badge and easily a false proxy for real productivity. Lots of people who sell courses on how to make courses and telling you that you’re behind without commenting “agent” on their post to get sent their paid playbook.

Finishing a thing is still rare. Finishing a thing that 30 people actually use is rarer. Don't confuse motion with progress…this is literally what I need to keep telling myself because the part I love is the motion part.

3. Your focus isn’t committed forever…use the hour glass tactic

I like to do new things and it does give me energy, and I don’t want to lock myself up in a way that stops be from experimenting and trying new things. Never put limiters on your energy and passion because when you have it those are superpowers.

But I want to push that energy and passion into one idea channel vs scatter shot it. So, I’m actively trying to create milestones and time constraints for the one idea at a time I am focusing on.

Like an hour glass. Wide point of view at the top, then hyper-focused and narrow for a limited time on one idea I’ve deemed good, then reassess after. Wide at the bottom again.

Now you turn it around after having learnt some stuff, and decide: do you double down again on the same idea in another milestone sprint, or pick another one if the first milestone didn’t pan out.

This is a tool that doesn’t fully cut myself off from ideas, but cuts me off from being pulled to new shiny ideas while focusing on finishing one thing.

Every product idea I follow now must at least make it’s way to production, and at least get to 50 users. No unpublished repos and local graveyards—it must make it’s way to someone at the bare minimum.

4. When something new drops — learn, try, apply.

You don’t have to build something new to learn a new skill.

Learn. Try. And apply it to the thing you're already on. I didn’t need to, for example, use Claude Design to go and tease myself into temptation with a different idea. I could have used it to design something for Little Moments and stay focused…which I later did and made my product marketing website.

Check it out it’s pretty cool and different: www.getlittlemoments.com

That's the learning loop.

Learn, try, apply. If you're doing that, you're already ahead of ~99% of everyone who owns an Anthropic or OpenAI sub.

The only way I lose at Little Moments right now is if I stop building Little Moments right now. And the most likely reason I'd stop is because the hum and drum convinced me something else was more important.

It might be, but I’m in the thin part of the hourglass right now and this hour is about this:

Then…assess at the bottom. What does the PostHog data tell me with real users not test users? Is this the thing that’s worth turning the hour glass and doing another “hour” on this idea?

Because at the end of the day, the only way to actually be not doing enough is by trying to do too much.

I don’t know…is this just all me? Does anyone else feel like this? Genuinely curious.

I’ll see you soon. Gotta focus.
—Jaryd (view my stack of non-negotiable building tools here)