Week 5 · 2025-01-27 → 2025-02-02 · 9 newsletters
The DeepSeek Week
deepseek-and-the-ai-trade · attention-and-media · the-craft-of-work · trump-trade-returns · grace-notes
A thin inbox week, eight emails across seven days. One genuine market-moving story (DeepSeek's R1 release rattling the NVIDIA trade), two pieces about attention in different registers, two on the craft of work, one on Trump's reentry into trade policy, and two grace notes. No dense newsletter ecosystem yet, so the signal that landed had to carry itself.
DeepSeek and the AI Trade: A Moment of Disbelief
The single live industry piece of the week was The Last Bear Standing on NVIDIA's late-January flush. The setup was the news: two Chinese researchers had released Deepseek V3 and R1, large language models that mirrored the capabilities of the American AI frontier but at a fraction of the headline cost, with just $6 million in pre-training spend and full open-source transparency. The market reacted accordingly. NVIDIA dropped, the AI trade wobbled, and the narrative scaffolding underneath two years of US equity outperformance got its first real stress test.
The Bear's reading was the calibrated one. Yes, the $6 million figure is the pre-training number only and the true buildout cost is closer to $500 million in GPUs, some acquired before export restrictions hit. Yes, Deepseek is not a hobbyist outfit but a serious research shop that insiders have tracked for a while. But the models are real, the methods are traceable, and the implication is unavoidable: the scaling-laws thesis that has anchored American AI investment since the OpenAI paper of January 2020 just got a credible counter-example. Loss may scale as a power-law with model size, dataset size, and compute, but power-laws cut both ways once efficiency breakthroughs arrive.
The take: this is the most important week of the year so far, and the only newsletter in the inbox that treated it that way was The Bear. The market reaction got the headlines, but the deeper story is the one The Bear surfaced: the American AI thesis is no longer the only credible thesis, and the cost structure that justified the buildout is now contestable. That changes the math on every dependent trade.
Attention and Media: Two Crises, One Diagnosis
Jenny G. Zhang returned to her newsletter after a five-year pause with a piece on media's unspoken "talent" hierarchy: the names anointed as the next big thing, who get the A-list profiles and the glossy assignments, versus everyone else doing the actual work of keeping outlets running. Her framing came from inside the contraction: she lost a job when Bryan Goldberg cut Gawker 2.0 loose, found another at a publication she had worked at before, and watched newsrooms shrink, text lose ground to video, and Twitter become what she called "a shrunken, bot-littered, fuck-ass husk." The piece is less a "future of media" essay than an honest accounting of what it feels like to keep working inside an industry whose downward slope has been continuous for a decade. The talent-hierarchy frame is the right one because it explains who gets the lifeboats and who does not.
Abby Falik at Taking Flight ran the companion piece from the consumer side. Her "Power of the Pause" was a reflection on attention as the crisis underneath all the other crises: democracy, climate, inequality. The trigger was a conversation she heard between Ezra Klein and Chris Hayes on "attention capitalism," and the line that stuck was the smartphone-as-cigarette parallel, including the historical note that doctors once endorsed preferred brands. Her own discipline (morning meditations, digital sabbaths, news-consumption rules) had cracked through in January under the weight of apocalyptic fires and executive-order ping-pong. The image that stayed: her boys having to practically scream for her attention because she was lost in a scroll, and her friend's rule of "people before phones."
The take: Zhang is writing about supply-side collapse and Falik is writing about demand-side collapse, but they are pointing at the same thing. The media industry cannot pay for serious work because attention has been arbitraged away from any context that would sustain it. The collapse is mutual. The interesting question, which neither piece quite asks, is whether the post-collapse equilibrium produces something better or just smaller.
The Craft of Work: Simplicity and the Midwit Curve
Two operator pieces hit the inbox within hours of each other on January 30. Ami Vora at The Hard Parts of Growth wrote on simplicity as a product-design principle, with WhatsApp as her case study. Her argument: simple products feel immediately familiar because they borrow patterns the user already knows, not because they invent new ones. The WhatsApp move was to match the operating system's conventions (Android's floating action button in the bottom right, iOS's tab bar) so a billion-plus users could open the app and start messaging without learning a new vocabulary. The harder discipline, she said, is resisting the urge to put your unique mark on every feature when borrowing the familiar pattern would serve the user better.
Julie Zhuo at The Looking Glass ran the management version of the same instinct with "The Valuable Employee Paradox." Her observation was clean: every great manager she knows says the reports they value most are the ones who convince them to do things differently, but most reports believe they are most valuable when they do what their manager wants. The reconciling frame was the midwit curve. The worst reports do what they want against the team's interest. The average reports do what their manager says without pushback. The best reports do what is right for the team even when that means disagreeing with the manager, and the manager rewards them precisely because the pushback is what made the decision better.
The take: Vora and Zhuo are both arguing for the same kind of restraint, just applied to different domains. Vora's "borrow the familiar pattern" and Zhuo's "push back when the manager is wrong" are both arguments against the cheap version of differentiation. The cheap version is novelty for novelty's sake or agreement for agreement's sake. The harder version is judgment about when to lean on the existing pattern and when to break it. Both pieces are worth saving as templates for that distinction.
Trump Trade Returns: The First Tariff Read
Vishnu Rajamanickam at The Logistics Report ran the week's only macro-policy read, an In Transit podcast with Jono Liu of Dimerco Express on what Trump's tariff push means for global trade flows. The substance was what you would expect a week into the second administration: importers worried, exporters bracing for tit-for-tat retaliation, the China+1 strategy gaining momentum as businesses hedge against geopolitical risk, the Chinese New Year slowdown layered on top of the de minimis ruling controversy that lets Chinese e-commerce giants sidestep US customs.
The take: the piece is the first of what will be many on this thread, and the framing (importer pain plus exporter pain plus de minimis breakdown) is the right one to carry forward. The de minimis question in particular is the load-bearing one for the Shein and Temu side of the consumer story. Worth watching whether it moves from podcast chatter to policy in the next quarter.
Grace Notes
Gabby Lord at omg lord marked David Lynch's death at 78 with a short Lynch link roundup, opening with the line "that dude sure loved two things: cinema and cigarettes." The Red Room image from Twin Peaks (1990) carried the post. A small dispatch, the right register.
Ben Kassoy at A Strawberry Spinning Like a Dreidel checked in from LA after evacuating from the January fires (he and Kristen live close enough to have left for a week with friends) and announced his solo show "The Funny Thing About A Panic Attack" returning to New York on March 6-7, with the first-ever livestream option for those out of town. The piece doubled as a freelance ask: copywriting and content strategy roles in nonprofits, mental health, education, arts and culture. The blend of fire-update, work-update, and tour-announcement was the honest mix of a working creative in late January.
Three Takeaways from the Week
DeepSeek is the one story that will matter in six months. The Last Bear Standing was the only newsletter in the inbox that engaged with it seriously, and the engagement was the correct one: take the model results seriously, discount the headline cost figure, and recognize that the scaling-laws thesis underneath the American AI trade just got its first credible counter. The market reaction was the loud part. The thesis revision is the quiet part. Track the second one.
The attention and media pieces are the same piece written from opposite ends. Zhang's account of the talent hierarchy inside collapsing newsrooms and Falik's account of attention slipping through her own carefully constructed guardrails are both diagnoses of an ecosystem that no longer pays for sustained focus, on either side of the transaction. If you are building anything that asks for reader attention in 2025, you are building inside a market that is contracting on both supply and demand. Plan accordingly.
If you only revisit three pieces from the week, I would suggest The Last Bear Standing on the DeepSeek moment for the cleanest read on the week's one market-moving story, Julie Zhuo on the Valuable Employee Paradox for the cleanest management frame, and Abby Falik on the Power of the Pause for the piece most worth reading slowly. Eight emails is not a lot. The three that earned the slot earned it twice over.