Week 22 · 2025-05-26 → 2025-06-01 · 16 newsletters
Experts Meet Their Tools
ai-and-the-expert · infrastructure-and-craft · questions-over-answers · voice-and-handwriting
The week from Memorial Day through the first of June. Seventeen emails total, which is exactly what a holiday-shortened week with a long weekend up front should produce. No dominant news story, no breaking macro thread, but a surprisingly coherent through-line emerged: a handful of writers, all of them experts at something, sitting with the question of what their expertise is worth in a year when the tools are catching up.
AI and the Expert: The Adaptation Problem
The strongest cluster of the week was a set of pieces by working practitioners thinking out loud about how AI is reshaping the value of mastery itself. Julie Zhuo at The Looking Glass opened it with "The Thing You Are Expert at Will Be Your Career Downfall," arguing that the people most attached to their craft are the ones most likely to be disrupted by AI, because the habituated neural pathways that make expertise feel automatic are the same ones that make adaptation feel costly. She named three reasons experts struggle to adapt: automatic comfort, identity attachment, and the sunk cost of years spent perfecting workflows the new tools are now doing in seconds.
Jacob Voytko at Client Server ran the practitioner's counterpoint in "My '30-minute rule' for LLM coding agents," racing Google's Jules agent on a deprecation task during his kid's naptime and finishing in 20 minutes while the agent was still spinning. His rule is the right one: coding agents need to save him at least 30 minutes per task over Cursor to justify breaking flow state. He tried it on two real changes to his Discord chatbot. One worked. One did not. The honesty about underspecifying the task on purpose, and watching the agent comment everything out rather than delete it, is the kind of field report the AI-tooling discourse is mostly missing.
Ethan Mollick at One Useful Thing gave the same conversation its visual benchmark in "The recent history of AI in 32 otters," using two years of "otter on a plane using wifi" prompts to track the progress of diffusion models from VQGAN-era noise to current Midjourney coherence. The accidental benchmark works because it stays constant while the tools move underneath it. Sahar Mor at AI Tidbits ran the operator-level version with "The Open-Source Toolkit for Building AI Agents v2," a curated map across eight categories from orchestration to voice to vertical agents, written for builders rather than market-makers.
Two adjacent pieces extended the thread into the design and venture corners. Carly Ayres at Good Graf wrote her May extremely-online report around what she called "the great design thirst of 2025," reading OpenAI's $6.5 billion acquisition of Jony Ive's hardware startup as a bet on design as strategy rather than style. The line that lands: "as AI commoditizes execution, the premium shifts to intention." Shruti Gandhi at Array Ventures announced Array AI Labs, where the headline experiment is an AI-generated podcast that recreates interviews with Ashton Kutcher, Garry Tan, Bill Gurley, Reid Hoffman, and Patrick Collison from publicly available content, hosted at array.vc/array-ai-lab.
The take: Zhuo, Voytko, and Mollick are running the same argument from different vantage points. The expert's problem is not whether AI works. It is whether the cost of rerouting decades of muscle memory exceeds the time the tool actually saves. Voytko's 30-minute rule is the cleanest practitioner heuristic I read this week. Ayres' observation that the premium is shifting to intention is the cleanest strategic one. Hold both.
Infrastructure and Craft: Quiet Pieces in Loud Categories
Three pieces this week were the kind of careful, narrowly-scoped writing that the noisy categories around them rarely produce. Aditya Bhargava at Ducktyped published "An illustrated guide to AWS Security Groups," another entry in his AWS networking series that walks through the difference between security groups (firewall level) and NACLs (subnet level) with terraform examples and the right amount of diagram. The piece is what good infrastructure writing looks like: scope tight, examples real, no marketing language.
Vishnu Rajamanickam at The Logistics Report ran a podcast episode with Colby Potter at Sayari on supply chain compliance under UFLPA, walking through how outside-in trade data from 250-plus jurisdictions surfaces concealed ownership chains and forced labor risks in places like Xinjiang. The real-world example, a North American energy company unknowingly linked to a Chinese state-owned firm flagged for forced labor, is the kind of specificity the compliance discourse mostly avoids.
Kerman Kohli made the macro argument with "You're Watching the End of Fiat," tracing the bond market's response to Liberation Day tariffs and arguing that no path through tariff escalation produces a winner. The bond market gets the last word regardless of who is in the White House, and that is the part of the argument worth keeping. He linked back to his earlier piece on the USDJPY chart as the precondition for the current setup.
The take: the writers who picked a narrow lane and stayed in it produced the most readable work of the week. Bhargava on security groups, Rajamanickam on UFLPA compliance, Kohli on bonds and tariffs. None of them tried to cover everything. All three left me with a frame I will reuse.
Questions Over Answers: Three Writers, One Posture
A small but coherent thread this week was three writers all making the same case for sitting with uncertainty rather than rushing to resolve it. Steven Schlafman at Where the Road Bends wrote "Living the Questions," opening with Rilke's "love the questions themselves as if they were locked rooms" passage and weaving it into Josh Waitzkin's Most Important Question framework from a recent Huberman Lab episode. The argument: in a culture that rewards speed of answer, the discipline of letting a question shape your decisions for months or years is a competitive advantage hiding in plain sight.
Abby Falik at Taking Flight ran the educational version in "The Future of School Won't Look Like School," arguing that the AI conversation in education is asking the wrong question. The right one is not "how do we retrofit school around AI?" It is "what makes us human?" Her Flight School cohort reports 100% transformative experience against the Gallup baseline of 3% for college grads, which is the kind of asymmetry that should make institutional educators uncomfortable. Her New York Times letter this week reframed it as: train better robots, or train better humans?
Maalvika made the smallest and sharpest version with "why are we lying to young people about work?", a one-line gut punch against the "do what you love" gospel that most parents and teachers still recite by reflex.
The take: Schlafman is writing for executives in transition, Falik is writing for educators and parents, Maalvika is writing for the kids on the receiving end of the bad advice. All three are pushing on the same posture: the answers we hand out reflexively are the ones least worth handing out at all. The questions are the asset.
Voice and Handwriting: The Quiet Pieces
Three pieces this week were the kind of writing that justifies the subscribe button on a quiet week. Folu at unsnackable wrote a meditation on her own handwriting as "singular, haunted, and mostly useless but a source of superiority nonetheless," drafted part of the newsletter by hand to test whether she could incorporate the image, and laid out her roster of late-in-life pivot fantasies (Japanese snack gashapon virtuoso, in-house candy strategist, duo-chrome cosmetic chemist). It is the rare voice piece that earns every digression.
Jacob Voytko's second piece of the week, "You have the time to run a technical newsletter," marks his six-month anniversary at 267 subscribers and shares the playbook for building a publication on an hour or two of free time per day with a toddler in the house. The honesty about the constraint (one to two posts per week, no social media following, no time to argue with strangers) makes the playbook usable rather than aspirational. Ben Kassoy sent the FUTURE CASTLES creative writing workshop announcement for June 10 along with a photo of his niece Tillie holding a birthday card. The community-building note ("$15-35 sliding scale, no one turned away") is the right register.
Conscious Talent launched as a new executive talent network from Daniel Reid Cahn and Zaharo Tsekouras, built around the thesis that work can be a vehicle for consciousness evolution rather than a thing you compartmentalize away from your inner life. The launch note is honest about the seven-year arc from spiritual awakening at Troops to building the service Cahn wished had existed. The Stonkstack closed out a deep value portfolio update on Zytronic's failed sale and partial liquidation, with the share price drop from 48-49p to 40-42p as the missed re-entry point.
The take: the quiet writing was the writing that held up best this week. Folu on handwriting, Voytko on building a newsletter in a toddler's nap window, Kassoy on workshop pricing and birthday crowns. The sparse-week reward is that the small pieces get room to breathe.
Three Takeaways from the Week
The expert's adaptation problem is the real AI story of 2025, and it is not the one the headlines are covering. Julie Zhuo, Jacob Voytko, and Ethan Mollick all wrote about it from different angles this week, and the convergence is not coincidence. The cost of rerouting decades of expertise is the friction nobody is pricing into the productivity projections, and the practitioners who admit that friction out loud are the ones worth reading.
The voice this week belonged to the writers who picked a small subject and stayed inside it. Aditya Bhargava on security groups, Folu on handwriting, Steven Schlafman on Rilke. Holiday weeks reward this kind of writing because there is no news pressure to chase. The lesson for newsletter operators is that the constraint of a quiet week is also a permission slip.
If you only revisit three pieces from the week, I would suggest Julie Zhuo on expertise as career downfall for the cleanest frame on the AI adaptation problem, Jacob Voytko's 30-minute rule for LLM coding agents for the cleanest practitioner heuristic, and Steven Schlafman on living the questions for the posture worth carrying into the summer. Three writers, three registers, one through-line: the tools are catching up, and the people who win will be the ones who keep asking better questions about what to do with them.