Blog

Today we’ll talk about AI broadly and within the context of my organization, books I finished this week, movie reviews, and a baseball primer.

Tech

One of the bigger headlines this week: Anthropic, a leading frontier AI lab, accidentally exposed the entire source code of Claude Code — its popular tool that helps developers write and manage code using AI — to the public. Think of source code as the original blueprint of a piece of software: the human-readable instructions that tell the program what to do. Companies guard this closely because it reveals exactly how their product is built. 512,000 lines of it, just out in the open. Not through a hack or a breach — someone on Anthropic’s team made a packaging mistake when publishing an update, accidentally bundling the blueprint alongside the finished product. Like shipping a piece of furniture with the factory schematics taped to the back. Anthropic confirmed it was human error and nothing more. Oh, and this was actually the second time this happened — a nearly identical mistake occurred with an earlier version in February 2025. To make matters worse, the timing coincided with a separate malicious attack on a common software tool many developers use, meaning some people who updated Claude Code that day may have unknowingly downloaded compromised software alongside it.

Inopportune timing for a company moving toward a likely IPO. But beyond the competitive damage — and competitors now have a literal blueprint for Claude Code’s architecture — this moment captures something essential about this phase of AI development. The pace of shipping has outrun the discipline of shipping. That’s not a knock specific to Anthropic; it’s the operating condition of the entire industry. And it cuts to a real tension: tools deployed faster than they’re understood will always create surface area for bad actors. The irony is that the only way through this messy middle is not to slow down, but to build more responsibly while moving at speed. Easier said than done when your competitor announced something new yesterday.

This headline is attention-grabbing, but the underlying story hasn’t changed. If anything, what I experienced this week at work only reinforced it. Here’s how I’ve come to understand the arc: After the Covid boom and the hiring spree that followed, the interest rate environment shifted. The market transformed from one that rewarded growth-at-all-costs to one demanding responsible, profitable growth. When rates are low, capital is cheap — companies borrow freely, speculate broadly, hire liberally into product areas outside their core. When rates rise, the cost of borrowing climbs, investors get a better safe rate of return, and equity investments have to deliver actual profitability. The golden metric becomes revenue per FTE: how much can you produce with the fewest people?

Then came the AI moment in 2022. And unlike traditional software — build once, serve every additional user at near-zero marginal cost, high margins — AI carries real fixed costs. Data centers. Energy. Chips. Inference at scale. In an environment demanding both AI-powered growth and investor-grade profitability, the immediate lever was obvious: layoffs. One big round, AI as the rationale, costs restructured in a single announcement rather than managed incrementally. Easier to do it all at once than to bleed it out slowly.

Now the question shifts. You have the cost structure investors wanted — but you still have to deliver the same results with significantly fewer people. That makes AI adoption at work not a nice-to-have but an imperative. Boards and executives across every company are asking the same question: we’re paying for all these tools and licenses — show us the return. Not just “my team ships faster,” but why that speed is translating into customers staying longer, expanding their contracts, or new sales. The bar has moved from efficiency gains to revenue attribution.

This is the internal dynamic playing out inside companies right now. We’ve moved from peacetime to wartime. Competitors are shipping, new entrants are appearing, and the pressure to keep everyone aligned and moving at speed is immense. In that environment, culture becomes the actual competitive weapon — more than the roadmap, more than the tools.

Here’s an idea I’ve been sitting with: the most important thing any company can do right now is find its signal. The deep, specific insight about what your company understands that competitors can’t easily replicate — the thing that anchors your value proposition, your customer capture, your ability to compound. And the employees most likely to surface that signal are the ones on the ground: running experiments, talking to customers daily, testing hypotheses in the real product environment. Leadership can articulate the vision, but they can’t be everywhere. The signal has to come from the people closest to the work.

Here’s the uncomfortable part. Once that signal is found and codified — turned into something an agent can execute and iterate on autonomously — the value doesn’t distribute evenly. It flows upward. The executives who own the strategy capture the compounding returns. The employees who did the hard discovery work of finding the signal in the first place become, in the calculus of a cost-conscious board, more replaceable. This is the principal-agent problem of the AI era: the people best positioned to find the signal have diminishing incentive to surface it honestly once they understand where the rewards actually go. Culture that creates genuine psychological safety — and compensation structures that share in what the signal produces — aren’t soft amenities. They’re the conditions under which the signal gets found at all. Without them, you get the other thing: people optimizing not for the right decision, but for the safe one.

Which brings me to where I am right now — and I’ll be direct about it, because this is as much a journal as it is a blog. This is the dynamic I’m navigating at work. I’m on the side that wants to push the boundaries, and the hardest part of operating in a wartime environment inside an organization that hasn’t fully internalized that it’s wartime is the gap between what’s possible and what feels permissible. The instinct to protect, to over-document, to require exhaustive justification before any risk is taken — these aren’t irrational behaviors. They’re rational responses to a culture where people don’t feel safe failing. But they’re lethal in a market where your competitors aren’t waiting for your roadmap to get approved. We produce quarterly roadmaps for a market reinventing itself monthly. We spend energy defending decisions internally that we should be spending on customers. We get in our own way. (More on some of the specifics at the right time.)

So how do I stay motivated inside that? I found a conversation between Derek Thompson and Brad Stulberg recently clarifying in a way I didn’t expect. The distinction they draw is between success and excellence. Success is a moment — a fixed point you can point to. Excellence is a continuous practice. Like physical fitness, it atrophies the moment you stop maintaining it. Treat your peaks and valleys with equal brevity: celebrate wins shortly, let setbacks settle and process them, then get back on the saddle. The goal isn’t arrival — it’s the pursuit itself. Tennyson put it better than anyone in “Ulysses,” his 1833 poem — “that all experience is an arch wherethrough gleams that untravelled world whose margin fades forever and forever when I move.” Experience doesn’t make you feel complete. It reveals how much further you still have to travel. That’s not a curse. That’s the whole point. As Kobe Bryant — the late basketball legend and five-time NBA champion — said when asked whether he plays to win or not to lose: I play to learn and figure things out.

Books

This week I finished Paul Kalanithi’s When Breath Becomes Air — a memoir of a neurosurgeon diagnosed with terminal lung cancer at 36 who uses his remaining time to reckon with the questions he spent his career observing from the other side: what makes a life meaningful, and how to live fully in the face of certain death.

I was drawn in before I even opened it — by the title alone. Four words that stopped me cold: When Breath Becomes Air. Breath is involuntary, continuous, mostly unconscious — the body keeping itself alive without your participation. Air is the raw material, inert and undifferentiated, just the medium. My first reading was loss: breath returning to air, the living decomposing back into mere matter, presence receding into absence. But as I moved through Kalanithi’s beautifully brave and unflinching account — written as he was dying and completed posthumously — it became clear the title runs in both directions. Air becoming breath is birth, the first inhale, the inanimate becoming animate. The title sits at both thresholds simultaneously: the moment of coming into being and the moment of leaving it.

The book is less a cancer memoir than a philosopher’s attempt to find language for that threshold. One passage that stayed with me: when Kalanithi consults with families facing end-of-life decisions after tragic accidents, he describes the gap between how families see the patient — for who they were — and how he sees them — for who they will never be again. I find myself genuinely wrestling with that. Would I want to continue living when I can no longer express and be who I am to the people I love? Would that count as living? There’s something brave in what Kalanithi models — the acceptance that releasing the people who love you to grieve and eventually heal may itself be a form of love.

The book also asks a harder question I haven’t stopped turning over: when does a life accumulate enough form that it can be said to have meant something? Not just been lived — but meant. Kalanithi doesn’t answer that cleanly, and I think that’s exactly right. You can’t wait to know before you begin.

Reading it has validated this project — this website, the habit of writing weekly, the attempt to build something I’m proud of and that produces meaning for at least myself.

Rating: 4.5 / A+ — Excellent

Movies

Before getting into this week’s films, I want to share the rating system I’ll use as I work through the full slate of 2026 Best Picture nominees. This is a project I’m starting now and intend to see through — beginning with this year’s class and working backward over time. My goal is straightforward: sit with the art form. There’s something I genuinely respect about movies that feels increasingly rare. A film forces a perspective and commits to an ending. For two hours, you submit to a director’s complete vision — no algorithm deciding what keeps you watching, no variable runtime based on your engagement. It doesn’t optimize for your attention the way TikTok, YouTube, or even most television does. It optimizes for art. That’s worth protecting as a habit.

The Oscars are not the arbiter of quality — they’ve been historically exclusive, commercially influenced, and sometimes simply wrong. But they mark what mattered in a given cultural moment, and that’s worth engaging with seriously. Here’s the scale:

Score	Grade	Descriptor
5.0	S	Masterwork
4.5	A+	Excellent
4.0	A	Strong
3.5	B+	Good
3.0	B	Solid
2.5	C+	Adequate
2.0	C	Below par
1.5	D+	Weak
1.0	D	Poor
0.5	F+	Very poor
0.0	F	Not done

Frankenstein

Guillermo del Toro’s adaptation of Mary Shelley’s 1818 novel arrives with the weight of something that has never really left American culture. Ask anyone to picture Frankenstein and an image appears immediately — the flat head, the bolts, the green skin, reliably present every Halloween like an old obligation. What del Toro does is strip that familiarity down and give you back the original terror: not the monster, but the man who made him.

The film follows Victor Frankenstein, a brilliant and catastrophically egotistical scientist whose act of creation destroys both himself and the creature he brings into being. Del Toro structures the narrative in chapters, moving between Victor’s perspective and the Creature’s own — a framing that forces you to hold both accounts simultaneously and trust neither completely. I’ll admit I haven’t read the source material yet (it’s on the list), but from what I understand this is a largely faithful adaptation, and del Toro’s thematic instincts feel right.

The film’s central argument, consistent across del Toro’s career, is that the monster is never the real monster. Here, you don’t just sympathize with the Creature intellectually — you feel it. He didn’t choose to exist. He didn’t choose his inability to die. He didn’t choose the revulsion with which the world receives him. The character of Elizabeth does important work in humanizing him — helping us see the Creature not as a freak but as a wound, the consequence of one man’s hubris made flesh. I couldn’t help thinking about AI: the tools we deploy before we understand them, the creations that outlast our intentions.

Jacob Elordi delivers a performance that earned every bit of his Best Supporting Actor nomination. The closing scene, in which Victor finally names the Creature as his son and asks for forgiveness, initially struck me as rushed. But sitting with it, I think that speed is the point. The Creature was never really driven by rage — he just wanted to be seen. As a son. As a human. The forgiveness comes quickly because it was always ready to be given.

Rating: 4.5 / A+ — Excellent

Hamnet

Chloé Zhao’s films have a particular quality: they make peace with difficult realities quietly, without forcing resolution. Hamnet is where that sensibility finds its most demanding subject yet.

The film follows Agnes and William Shakespeare from the sparks of their love through marriage, three children, Shakespeare’s growing absence as he pursues his career in the London theater, and the devastating death of their son Hamnet from the bubonic plague — which becomes, eventually, the grief that Shakespeare transforms into Hamlet, one of the most celebrated plays ever written.

What the film gets absolutely right is where it chooses to orbit. This is Agnes’s story. Buckley plays a woman of wildness and instinct — a witch girl who builds a family out of a longing for belonging, only to have it dismantled. The emotional architecture of the film is built around her, and the performance earns the weight it’s given.

My one pointed critique is how the film handles grief itself. It anchors on the contrast between the quiet domestic world, and episodes of raw, physical crying. At times it feels much, even if I understand the pain. Though there, I would have liked the numbness, the strange bureaucracy of loss, the dark humor that constitutes how people survive it to be played up more loudly compared to the physical pain and anguish. You feel its force without understanding its texture, and that flattens what should be the film’s most resonant emotional register.

The final third recovers beautifully. Paul Mescal’s Shakespeare, after losing his son and writing Hamlet as an act of mourning, watches his actors deliver the lines flatly and cannot bear it. He steps in and performs it himself — with everything it was meant to carry. It’s the most moving scene in the film, and it’s worth the wait. The play becomes the elegy. The art outlives the grief. That’s the healing the film has been building toward, and it lands.

Rating: 3.5 / B+ — Good

Sports

Baseball is back and I’ll always welcome it the way I do every spring. Growing up, April meant baseball: cold early games where catching the ball actually hurt your hand, the season stretching warm into August. I’ll keep this brief — we’ve covered a lot of ground this week already — but I at least have to get my predictions on record. I’ve become less stats-driven as work and adult life have taken over, but I’ll lean on my gut when it comes to sports. You may not, but I will.

American League division winners: Yankees, Tigers, Mariners. I like the Yankees’ offense and the return of Gerrit Cole. The Tigers have a strong rotation with real prospects arriving this year. The Mariners feel like the most complete team in baseball to me — arguably the best pitching rotation in the league paired with a capable offense.

National League division winners: Mets, Brewers, Dodgers. The NL East is genuinely the hardest division to call — honestly the East in general, AL or NL. But I like what the Mets have assembled. Soto and the crew rebound this year. The Brewers are the Rays of the NL — steady, shrewd, chronically underestimated. And the Dodgers are, well, the Dodgers.

World Series: Yankees vs. Dodgers. Powerhouse versus powerhouse. Baseball back on the ascent. The Dodgers three-peat against the franchise that last three-peated. It’s a good story — and I am, after all, a storyteller.