I came across an article this week titled “Why You Should Start Blogging (Even If Nobody Will Read It)” and without even diving into it, I both knew what it would cover and felt validated in finally getting this one up. Writing regularly has been a long goal of mine. Every year it made the list, and every December I’d reflect back with that quiet disappointment that it didn’t happen.
I’ve always heard about the benefits of writing to better articulate your thoughts and create clarity in your thinking. I saw it modeled early — my mom, despite a full and hard day of work and raising my sister and me, would still sit down and write in her journal. In a time of such major change, I think it’s even more important to do the same.
I don’t fully know what the future looks like, but even if it hasn’t fully materialized yet, I can see the shape it’s arriving in. A world where agents at minimum augment your work. If I’m being honest, this blog is already augmented by AI — and use it as an agent to go into my website and publish the article on my behalf. So in a world where AI and agents are becoming more important, I think it matters to keep writing down what you actually believe — both to avoid being consumed by the hype, and to build something that reflects you rather than just echoes everyone else.
Tech/Product
On that note, I keep thinking about the role of AI and agents. It’s impossible not to when you work in technology. For those who don’t know, I’m a product manager at a digital health company, and one of the most exciting things I’m working on right now is building the next generation platform for one of our core products: healthcare billing.
Healthcare billing is complicated — possibly by design — and accounts for a massive chunk of waste in the US healthcare system. Here’s an analogy. I go to a local CrossFit gym three times a week. The owner rents the space, pays coaches, covers operational costs, and when people like me show up for the community and the workout, she gets paid for her service. Clean and simple. Now take a provider running their own independent practice. Same basic setup: overhead, staff, a service rendered. But the analogy stops the moment the appointment ends. Once you’ve seen the patient, you have to document the visit, map it to billing codes for your diagnosis and the procedure performed, submit it to the payer, wait for them to adjudicate it — usually downward — and then chase patients for whatever balance remains. Reminder after reminder. It’s a slog. The whole process is a massive cost to US healthcare, but also a massive employer within it.
What makes this moment interesting isn’t just that AI exists — it’s that the brokenness of healthcare billing is exactly the kind of problem AI agents are well suited to absorb. The friction isn’t clinical. It’s administrative. It’s the gap between the care that was provided and the payment that should follow. That gap is made of documentation, codes, rules, and follow-up — the kind of repetitive, reasoning-heavy work that agents can take on in ways that pure automation never could.
I give that context because the company isn’t simply asking me to build a better version of the current product with new features or a cleaner UI. They want a new paradigm. A lot of companies in tech are facing this same hard decision right now. There’s a market-wide acceptance that even though the final shape of AI isn’t clear, it is clearly arriving. The winners will be the ones who understand — or at least can accommodate — that shape, and who have nailed the experience, maintained loyalty, and built the distribution to carry it forward. That’s what I think about constantly.
The ultimate design problem is building for a world that doesn’t fully exist yet but is clearly on its way. My thinking is evolving fast, but here’s where I’ve landed: AI will be agent-based. And that changes what a product needs to be. I no longer want to just click a button to submit a claim. I want to wake up every morning knowing how much I got paid and what needs my attention. The system should tell the agent why certain claims couldn’t be submitted and give it the ability to reason and act on the user’s behalf to get the provider paid.
My design philosophy right now: whether the user is a human or an agent, the product needs to surface insights, trends, and actions — not hand them a blank canvas and say good luck. The goal isn’t to automate the existing workflow. It’s to make the workflow worthy of the care that preceded it.
Food
This past week, my partner and I dogsat for a friend. We figured it’d be low effort and an easy way to make some money. It was both. Shoutout to Luna for tolerating our little guy Koda. And naturally, once we got paid, we immediately blew it on one of our favorite Korean BBQ spots: Seoul Grill. Fourth visit. Consistent as ever. Solid 4.5/5 — great variety of food, but the real story is the service.
Specifically, it’s Ethan.
He’s our server every time. I don’t remember his face — I never do — but the second he opens his mouth I know exactly who’s serving us. I don’t even know how to fully explain how this happens. I just get transported back to every previous visit and immediately confirmed: it’s him.
Now, I have to tell you about the makeup moment.
We sit down, he takes our order, looks at my partner and says — calm as anything — “your makeup looks great, my girlfriend does it the same way.”
Here’s what happens in my head in real time: I clock the compliment. I watch my partner light up. And for half a second I’m sitting there thinking — okay, is a man really complimenting my partner’s appearance right in front of me? Is this happening? And then the second half of the sentence lands. My girlfriend. Done. Diffused. The whole thing neutralized before I even had time to fully feel weird about it. That one-two punch — lead with the genuine compliment, drop the diffuser before it gets awkward — was so cleanly executed I found myself just genuinely impressed. The man is exceptional at his job.
And that job, it turns out, is rooting for you.
At Seoul Grill, each pricing tier unlocks more premium cuts of meat, each with a per-person limit. Most servers at all-you-can-eat BBQ places sit somewhere on a spectrum between indifferent and quietly hoping you forget to order your last round — less food out means less work for them. There’s no incentive to make sure you get your money’s worth. Ethan is the complete opposite. He tracks where you are in the meal, knows what you haven’t hit yet, and advocates for you like he has a personal stake in it. This visit, as we were slowing down and losing steam, he leaned in and told us to max out that prime short rib. The man wanted us to beat the system.
Something else I couldn’t shake afterward: why does the same guy serve us every time? Why do we always end up at the same table? And why does the manager make us put our name down on a quiet Monday evening, in a half-empty restaurant, have us wait five minutes, and then walk us to our seat in a sea of open tables like he just did us a favor? At first I found it comical. On the drive home I thought: maybe that’s the whole point. Maybe this restaurant has made a deliberate decision to deploy its servers as customer success managers. You build a relationship with Ethan, he knows your tendencies, and he becomes the experience — not just a vehicle for the food. The manufactured wait, the consistent table, the same server every time — it’s a retention strategy dressed up as hospitality. And it works, because the relationship is real even if the system behind it is engineered.
This all leads to their gratuity setup, which is equal parts frustrating and smart. They automatically add 18% gratuity — even for a party of two — and still hand you a receipt with a line for an additional tip. I have complicated feelings about tipping culture (topic for another time). But every visit, even knowing I’ve already paid gratuity, I fill in that line. Ethan earns it.
Books
My reading philosophy is a barbell. One end is staying current — macro news, tech, healthcare, what’s happening. The other end is for more compounding reads — classics, great fiction and nonfiction, ideas that have aged well. The goal is to stay informed without losing the ability to think across longer time horizons.
This week I listened to A Day in the Life of Abed Salama by Nathan Thrall, which won the Pulitzer Prize for General Nonfiction in 2024. The book centers on a devastating crash involving Palestinian teachers and children in the West Bank. I was drawn to it because I was curious about the human element inside a long conflict. I’m not an expert on the Israeli-Palestinian situation by any stretch, but where this book shines is in arguing how the conflict — and the systematic fractures that come with it — shapes the daily lives of the people inside it.
The story follows Abed, a Palestinian father searching for his young son Milad after the crash. Thrall reveals the obstacles layered in front of him: limited freedom of movement, long delays in emergency care, and the specific bureaucratic cruelty of the ID card system. Palestinians in the occupied West Bank carry green IDs; those in East Jerusalem carry blue ones. That card determines where you can go and how fast you can get there. Abed couldn’t reach his dying son directly because he had the wrong one.
What I appreciated — and what I think is a deliberate authorial choice — is that Thrall doesn’t make Abed a saint. He’s a flawed man, an imperfect father. There’s no strong bond established between him and Milad on the page. And I think that’s the point: Thrall isn’t asking you to sympathize because the victim was exceptional. He’s arguing that the injustice is structural, and doesn’t depend on the victim’s virtue to be real. That’s a harder argument to make, and a more important one.
But that choice costs something emotionally. Milad barely exists on the page. We know he liked chips and chocolate. That’s almost all we get. Five years old, and we never really meet him. The kids who died were potential, extinguished before it became anything. And Thrall’s most quietly devastating insight is that it’s the living who are transformed — not the dead. Milad is fixed. Abed is changed forever. The grief belongs to the survivors still inside the system, still navigating it, still carrying their green ID cards.
Where I think the book falls short is that Thrall never fully resolves the tension he creates. He wants the structural argument and the human story, but he lets the structural argument win. The result is a book that is more intellectually convincing than emotionally devastating. For a story about a father who can’t reach his dying son, I wanted it to break me open in a way that’s almost unbearable. There are moments where it gets close — and those moments are extraordinary. But it never quite gets all the way there.
4/5. Important. Deeply researched. A powerful witness to an entire political reality. Read it.
