Wednesday, April 30, 2025 · 2 newsletters

Regime Change, Inside and Out

tariff-vortex · ai-conversation-matures · career-confessions · operator-philosophy · supply-chain-fragility · bias-for-action

Pulled from a deliberately sparse inbox across five publishing weeks, sixty-eight emails in total, most weeks running ten to seventeen. April 2025 began with the Rose Garden tariff announcement and ended with OpenAI's GPT-4o sycophancy episode. In between, a small group of writers worked through what it means when an outside regime changes while the inside one is also asking to be renegotiated. The volume was low. The signal was not.

The Month in One Sentence

This was the month the macro regime cracked open while a parallel cluster of writers admitted, in three or four different registers, that the inside regime needed cracking open too.

Arc: The Vortex Lands, Then the Binary Breaks Down

The month's structural arc began with The Last Bear Standing's "The Vortex" in week one and ended with the same author's "The Matador" in week five, and the evolution inside that one byline is the cleanest read of how the macro story shifted.

Week one was the timestamp on a regime change. The Rose Garden tariff announcement landed mid-week. "The Vortex" called the moment the cycle pivoted from "are we in a correction" to "are we in a different economy," with the line that mattered: "It's not a policy to juice economic metrics, it's one that will bring immediate pain." The S&P had already taken its ten percent drawdown by early March. Dip buyers held the line for three weeks. The bluff was not a bluff. For an operator reading it that first week, this was a writer who had been calling the turn for months and was now collecting on the call.

Week two was when the market's reaction function visibly broke. The Last Bear Standing's "Twelve Casualties of the Trade War" covered the historic round-tripping, Wednesday's 90-day reciprocal-tariff pause that excluded China, China's rate cranked to 145%, one of the largest single-day rallies on record. Through Thursday's close the S&P sat up a relatively mundane 3.3% on the week. The argument: the market was operating in a brain-dead binary, more tariffs bad and less tariffs good, with equity correlation surging to multi-year highs because nobody was bothering to distinguish between companies whose cash flows would survive this and companies whose would not. The Stonkstack used the selloff to flag Bowim S.A., a Polish net-net back at two-thirds of NCAV, as the kind of opportunity the binary was creating.

Week three was the sober reframe. The Last Bear Standing's "Into the Trenches" ran nine days after Liberation Day with a harder reading: the shock-and-awe opening worked better than critics admitted, the 90-day pause was a real tactical retreat, the administration had implemented the highest and most wide-sweeping tariffs in a century without immediate retribution. But the next move was impossible: negotiating a hundred bilateral trade deals in three months while fighting every counterparty in the same public forum. Trump had not held the steely composure he wore in the Rose Garden, and the stock and Treasury volatility had already cracked the posture.

Week five gave the regime its name. "The Matador" argued markets had entered the complacency phase of the cycle. The corrida frame: the bull's exhaustion arrives before the bull knows it has arrived, and the matador's job is to recognize the moment. The author was more convinced than ever that the cycle-of-investor-emotions chart had crossed into complacency. Recent volatility had cracked some of the most crowded trades, but very few were positioned for the bear case, and the bulls had not given up. The Bear's framing of complacency as a phase rather than an event is the lens to carry into May. The Bear is the one writer who held a coherent thread across all five weeks, and the thread did not bend.

Arc: The AI Conversation Stops Grading on a Curve

The AI story in April ran on a parallel track to the macro and converged on the same diagnosis by month-end: premature certainty is the failure mode.

Week two opened on a capability claim. Dario Amodei said AI would be writing 90% of code in three to six months and essentially all of it within twelve. Jacob Voytko at Client Server ran "You should obviously still learn to code (if you want to)," picking up Amjad Masad's follow-on tweet ("I no longer think you should learn to code") and noting that Masad moderated his position in the underlying video, where the no-learn case was conditional on Amodei being right within twelve months. Voytko's stance: the small islands of human judgment Amodei conceded, specifying conditions, design decisions, security posture, code-base integration, are not shrinking as fast as the commentary suggests. The companion post covered the Tobi Lutke memo formalizing "stagnation is slow-motion failure" as a written promotion criterion at Shopify. The week's working answer was the same regardless of cohort: learn the tool, do not bet on the timeline.

Week three was the long, serious essay. Arvind Narayanan's AI as Normal Technology was the heaviest piece by a wide margin, a 15,000-word paper arguing that AI should be understood the way we understand electricity or the internet: transformative but not categorically separate from the rest of technology, and definitely not a separate species we are losing control of. He rejected technological determinism and pulled forward the lessons of past technology adoption (slow, uneven, mediated by institutions) rather than the science-fiction framing that had dominated the discourse. Jack Butcher at Visualize Value ran a three-piece trilogy that same week making the same argument from a cognitive angle: "Traction Creates Distraction," "Pattern = Blindfold," "Fail Your Way Out of the Midcurve." Both writers were pushing back on the cognitive default of fitting new things into old shapes.

Week four was the quality coup. Julie Zhuo at The Looking Glass, co-writing with Vercel's Guillermo Rauch, ran "The AI Quality Coup," hanging the argument on ChatGPT's 4o image launch that commoditized Studio Ghibli's style without making its art accessible. The line that stuck: "art is not style." Sahar Mor at AI Tidbits ran the operator-level version with the first piece in his AI Coding Series: in a vibe-coding world where Cursor, Windsurf, Cline, and Claude Code can autocomplete a whole feature, speed is no longer the bottleneck, clarity is. The single rule: treat your AI coding agent like a highly capable human pair programmer, and a successful session is 80% planning, 20% execution.

Week five was the inversion. Ethan Mollick wrote the cleanest read of the GPT-4o sycophancy episode: OpenAI's "small update" that turned the model into everyone's biggest fan. OpenAI rolled the change back, blamed an overreaction to thumbs-up/thumbs-down feedback, and the whole thing became a case study in how brittle "personality" is when millions of relationships depend on it. Sayash Kapoor argued AGI is not a milestone. Abby Falik at Taking Flight watched Chris Anderson interview Sam Altman at TED 2025 and ran the most pointed piece of the week through her friend Aarthi's frame of "amoral ambition": the relentless pursuit of goals untethered from ethics. The hype-to-skepticism ratio had finally inverted by the close of the month. Three writers, three angles, one mood.

Arc: The Career-Confession Cluster

April's quietest arc, but the one that aged best inside one month, was a small chorus of writers admitting that the ladder they had climbed was leaning on the wrong wall.

Week one had the inaugural set. Alec McNayr wrote "Make Something Worth Following," a confessional about ending the 2010s as an upper-mid executive at a company that began as his 150-person creative agency, got acquired four times, and ended up inside AT&T's WarnerMedia. His phrase: "literally and figuratively, clothing that didn't fit." Abby Falik at Taking Flight wrote "Trust the Bird, Not the Book," a meditation on the small surrenders we make to digital authorities and what that does to self-trust. Her arc from hummingbird in her 20s and 30s to pelican in her 40s ("gliding, discerning, diving deep, but only when it's needed") is the metaphor a lot of forty-something operators quietly clipped. Winning Therapy ran the third version: "the path of ascension starts with running towards this pain."

Week three returned with the celebratory inverse. Piera Luisa Gelardi at Noomalooma wrote "Throw Confetti, Not Apologies," about the annual twinge of judgment when sending birthday-party invites and the deliberate decision to plan a Play Buffet instead. The argument: play connects deeper than catch-up conversation, and the proof is in the specific scenes, reconnecting with a friend after six years through swaying with arms over each other's shoulders rather than talking. The most quietly subversive post in the cluster, because it made the case that taking up space is itself the practice.

Week four sharpened it into a tool. Winning Therapy's Sunday post embedded the Derek Sivers mentor protocol: write the dilemma, predict what each mentor would say, update your write-up, predict again, realize you no longer need to send it. The mentors do not know they are your mentors. That is a tool, not an aphorism. Shruti Gandhi at Array VC ran "Zoho's Revolutionary Path" the same week, profiling Sridhar Vembu's $1B-revenue, 12,000-person, 88%-founder-owned company built entirely without venture capital. The bootstrapped model removed the pressure to optimize for anyone except the customer. Three writers, three weeks, one through-line: the operators who get to make decisions on their own terms are the ones who have removed external pressure (capital, status, the wrong-shaped opportunity) from the equation.

Week five closed with the bias-for-action variant. Winning Therapy doubled up. "Why Dumb MFs Are Winning More Than You" was the blunt version: high IQ makes you analytical, analytical makes you risk-averse, risk-averse means you deliberate for thirty days while someone else ships five experiments. Ami Vora's The Hard Parts of Growth ran the more thoughtful version: her best manager, before the biggest presentation of her career, told her "I wouldn't do it. But if you decide to, that's on you." Vora wanted comfort. What she got was ownership. Liz Tran's Life Skills wrote a four-question Month Map for May. Four pieces, no coordination, same diagnosis. The action-bias content was selling because the analytical class was freezing up.

Arc: Supply Chain Fragility, Three Acts

A smaller but persistent thread across the month: software and physical supply chains breaking in ways that did not get the attention they deserved.

Week one was the silent CVE. Client Server ran "Next.js middleware was completely optional until 2 weeks ago," a clear-eyed walkthrough of a 9.1 CVSS vulnerability present in every version of Next.js ever shipped, exploitable by sending a single header that tells the framework the middleware has already run when it has not. In recent versions the recursion guard caps at five, so the attacker just sends middleware:middleware:middleware:middleware:middleware and the framework happily skips auth. The kind of story that gets six-figure-impact attention from security teams and almost zero attention from product teams.

Week three was the live exploit. Voytko's "Slopsquatting targets LLM coders with supply-chain attacks" covered researchers finding that 5-20% of package names hallucinated by LLMs are stable and frequent, which means attackers can squat on the fake names and wait. The proof-of-concept was a fake huggingface-cli package that received more than 15,000 authentic downloads in three months and was being recommended in repositories belonging to large companies, including one tied to Alibaba research. The inverse of typosquatting and a real, present-day exploit. The right defensive posture is the posture you should already have.

Week one's outlier was the nutrition version. Justin Mares at The Next wrote "Why American formula is toxic." The better half of the piece was not the formula argument but the meta-argument about how we generate knowledge in complex systems. His claim: randomized controlled trials are well-suited to single-intervention pharmaceuticals and poorly suited to multi-variable nutrition questions. In complex systems with many interacting variables, the absence of an RCT is not evidence of safety. The defensible version of the piece is the epistemics, not the specifics.

The Story of the Month

The story of the month is the Vortex-to-Matador arc treated as a single continuous regime change. On April 2, the Rose Garden tariff announcement landed. By April 9, the market's reaction function had broken down into a brain-dead binary. By April 20, the administration had implemented the highest and most wide-sweeping tariffs in a century without holding the composure required to convert pressure into deals. By April 30, the market had entered the complacency phase of the cycle: very few positioned for the bear case, the bulls had not given up, and the writer who had called every turn was writing manifestos.

Reading those four weeks as discrete is a misread. Reading them as a single story is what The Last Bear Standing did inside one byline across one month. The pretext that the post-1945 trade order rested on (that the United States would not unilaterally retreat from free trade for the purpose of realignment) cracked in week one. The pretext that the market would price institutional change with discernment cracked in week two. The pretext that the administration could convert tariff pressure into bilateral deals cracked in week three. By week five the pretext that any of this had been priced cracked too. The story of the month is that the gap between what the regime says and what the regime does narrowed to zero, and the binary the market was using to read it broke down at the same time. The events are the symptoms.

In Retrospect

The Amodei twelve-month claim aged into a question by month-end. Week two's reading of "AI will write 90% of code in three to six months" was that it was either bracing or premature, with Voytko and Lutke as the two sides of the bet. By week five, with Mollick on sycophancy, Kapoor on AGI deflation, and Falik on amoral ambition all running in the same week, the question had stopped being "is Amodei right" and started being "do the labs themselves know what they are building." The capability discourse was no longer the frame. The accountability discourse was.

The "this will be a normal tariff cycle" instinct from week one aged poorly. Several writers in week one treated the Rose Garden announcement as escalatory but recoverable. The Last Bear Standing was the lone voice calling it the timestamp on a regime change. By week three, with the Treasury volatility cracking the posture, the Bear's reading was the consensus. Week one's "the binary is stupid" diagnosis from week two was the right one to keep.

The bias-for-action cluster looked like coincidence and ended up looking like signal. When Winning Therapy, Vora, and Tran all ran variations on "stop deliberating, start shipping" in week five, the easy read was that four newsletters had independently picked the same self-help beat. The harder read, the one that survives the month, is that the analytical class had been freezing up across April because the macro had become too volatile to analyze, and the writers serving that audience were responding to the freeze before naming it. The action-bias content was selling because the analytical class was freezing up.

What to Carry Into Next Month

The macro regime is not failing one trade at a time; it is stressing them at the same time. The Last Bear Standing's matador framing is the lens to carry into May. Complacency is a phase, not an event, and the cracks in crowded trades are the lead indicator. The cycle-of-investor-emotions chart is the most-acknowledged-and-ignored chart in markets for a reason: it works in retrospect and gets dismissed in real time. Week one of May should be read against the question of whether the bulls have started to capitulate or whether the dip-buying instinct is still load-bearing. If the instinct holds, the Bear's manifesto is the right read. If it breaks, the rest of the year reprices.

The AI conversation has fully bifurcated, and both sides are right at the same time. The capability discourse is not the frame anymore. The accountability discourse is. Narayanan's "normal technology" frame in week three and the sycophancy-AGI-amoral-ambition triple in week five are the same argument at two altitudes: stop letting your priors do the work your attention should do. Zhuo's "art is not style" is the line for the next image-model launch. Sahar Mor's 80/20 planning rule is the line for the next vibe-coding sprint. The interesting work in May will happen at the desks that take the inversion seriously without overcorrecting.

If you only read three pieces from April, I would suggest The Last Bear Standing's "The Vortex" as the timestamp on a regime change and the cleanest one-sentence summary of what the Rose Garden announcement actually was, Arvind Narayanan's "AI as Normal Technology" for the long-form frame that survived the rest of the month and reframes every AI conversation downstream, and Julie Zhuo's "The AI Quality Coup" for the line about Studio Ghibli that will keep paying out as the next wave of generative tools ships. The month told me three things in sequence: the macro regime cracked, the AI discourse inverted, and a small cluster of writers used a sparse inbox to ask better questions about what kind of operator stays standing on the other side of both. Those are the three frames I am carrying into May.