KAT:
"I just felt a calm and a peace that had to be come from the woman above (his mom). I felt like a kid. It was just fun out here. It felt like I was a kid getting ready to go play my Saturday AAU games and Sunday AAU games. In a way I felt like I was seeing her in the stands"
Phase 1 ✓: pump, solenoid, filter, drum. Lucky for me I have my dad supervising the operation. Fact about my dad, he was diagnosed with #parkinsons disease over a decade ago. The disease has taken some things from him physically. But it hasn’t taken his judgment and system thinking (and his sense of humor too. He said this is the dumbest smart thing he’s ever seen me build). Watching him use @NotionHQ voice + @WhisperFlow to stay in the build, to apply that craft—that’s the real win. These tools restore agency to people who’ve lost it. That narrative is almost entirely missing from AI discourse.
Next phase: the cloud brain. A Notion Worker checks weather 3x/day and enqueues spray commands. A Pi claims them, fires the solenoid, logs every run. Full autonomy, so we don’t have to think about it. My dad’s going to take the wheel to do some building of our digital system. More to come!
This weekend I’m putting @NotionDevs to build a smart mosquito and gnat repellant system in order to solve a problem plaguing suburban dads:
Dad Sprawl — the slow accumulation of subscriptions in the pursuit of home excellence.
I live in the DC area and the gnats and mosquitos are out of control.
I pay 200/mo for a mosquito spraying service. It rains, it washes away, I pay again. I was quoted to upgrade to an "Evergreen” misting system quote: $8,600 + $200/mo + a $25 app.
Respectfully: no.
So I’m building the Mosquito & Gnat Destroyer System:
Notion DB = the source of truth (schedule, rules, zones) + a “Spray Commands” queue + run logs
Notion Worker (runs on a schedule) = reads the DB, checks weather/time, decides “spray vs skip,” and queues a command (zone + seconds)
Raspberry Pi = watches the command queue (or receives a request), flips a GPIO pin to trigger the relay/pump/valve, then marks the command/log as Done
All-in, I’ve spent about $500 on equipment and supplies. Next I’ll cancel my $200/month service. Ongoing costs should be about $45/month for concentrate refills, plus roughly $1–$2/month in tokens to orchestrate everything in @NotionHQ (agent + worker). Follow along the journey if interested!
This was really special for me to watch. Two CEOs that I had (and have) the chance to work with and really admire. So much craft and wisdom packed into podcast.
Some tips for founders navigating the shift from software to AI.
Brian and I also talked about "jazz mode," the Catholic Church, brewing beer, and the Grateful Dead.
@nathanclark_ Your tweets are basically my Notion happiness index.
If you’re happy, the masses are probably happy too. You seem really happy. This is a good omen!
Spot on @bhalligan.
Seeing this play out on the front lines right now and it's why our notion consultant category is growing so fast.
Strategic illegibility is hard and the demand for that skill-set is real.
what's in your head → how your systems support the workflow today → what to make legible to AI → how to do that → what not to build when you can build anything.
That last one is where the a lot of the value lives. and there's no playbook for it yet.
Anyone who’s worked in partnerships knows how heavy the work can be.
Constant air traffic control between partners, customers, and internal teams. Messy. Slow. Easy to become the bottleneck.
I built a custom agent in @NotionHQ that removes me entirely from that equation. It is called the Notion Consultant Matcher. It sits on top of an expertise graph of our consultant ecosystem with hundreds of signals: how they've used Notion, workspaces they've built, industries they've served, integrations they're familiar with, client sizes they've worked with.
Now AEs can match the right consultant to their client's use case directly. No slacking me and waiting for me to reply. They're confident in the recommendations because they're matched with partners who have actually done the work the client is trying to achieve.
That confidence carries through to the client. The deals close faster. The client's are set up to get the results they need from Notion. And I get to spend time on more valuable work: ensuring our consultants are driving exceptional outcomes for our clients.
That's what it all comes down to.
Andrej Karpathy: "To get the most out of the tools that have become available now, you have to remove yourself as the bottleneck.
You cannot be there to prompt the next thing. You need to take yourself outside the loop. You have to arrange things such that they are completely autonomous.
The more you can maximize your token throughput and not be in the loop, the better. This is the goal. So, I kind of mentioned that the name of the game now is to increase your leverage. I put in very few tokens just once in a while, and a huge amount of stuff happens on my behalf."
---
From @NoPriorsPod YT channel (link in comment)
AI is raising the ceiling, and with it, the intensity. Faster pace, higher expectations. Bigger goals. It’s easy to get stuck in the echo chamber of throughput + metrics and miss the joy.
Then you see the human impact of the work you’re doing.
In my case, it’s the @NotionHQ Consultants Ecosystem: a partner sharing they’re already ~70% tracked toward their total 2025 revenue by the end of Q1 '26. Another celebrating their first six‑figure month. Another growing so fast they’re hiring 8–9 people in the next two months because the work is real and the demand is real.
Then I had a call with a seasoned partner and we weren’t talking about AI. We went back to the basics: identity, positioning, what they’re building, and why it matters to them.
These are the moments that makes the pace worth it. The ecosystem is people, consultants gaining confidence, hitting goals, hiring teams, and delivering work they’re proud of because it solves real customer problems.
The intensity is real, but so is the joy. And it’s reminded me why partnerships has always been the work I come back to.
I agree with my friend Brian most of the time.
This is one of those times.
Start making your internal company data legible. Not just for AI, but in general.