Unqualified Advice

Sean Filipow and Daniel Hatke
  • 4.8 (4)
  • Careers
  • Updated Biweekly

Hello and welcome to Unqualified Advice, an entertaining show for entertainment purposes. Join us as we talk about running our small businesses, what we've been learning, and how we're applying lessons from academia and real life as entrepreneurs and investors.

  1. 4d ago

    The Easiest $250 I've Ever Lost

    Easy Come, Easier Go Hello dear show notes readers! Happy 250th, America! I spent the Fourth in Dallas watching Egypt and Australia go to penalty kicks in front of seventy-some thousand people. Nothing like getting full value for the ticket. In our increasingly small world, I randomly found a Egyptian coworker in the crowd to celebrate with. Then I went back to my friend's parents' place (shout out Hotel Mary), pulled up Kalshi for the first time in my life, and put $250 on the US to win the whole thing. In honor of the 250th birthday, naturally. Four percent odds, a plan to sell into the spike if we beat Belgium, and a genuinely unsettling realization: it took me five minutes to set up an account and place the bet. Meanwhile Dan couldn't figure out which of three Fox apps would actually show him the game. He put it perfectly: you can't easily pay to watch the World Cup, but you can bet on it in seconds. You already know from the title how the trade worked out. Before we got anywhere else, Dan filed a formal appeal with the fact-check department. He'd taken a yellow card in a recent episode for his "54 cents of every Amazon dollar goes to Amazon" claim, and he came armed with an argument about survey scope — his number is from a US-sellers-only dataset. I heard the case, ruled from the bench (see the fact-check below), and took the opportunity to point everyone at our new website, unqualifiedadvicepodcast.com, where the whole back catalog now lives with fact-check and prediction scorecards. We're holding up around 85% on facts. The middle of the episode is the most practically useful AI conversation we've had in a while. With Fable back in our hands for the weekend, Dan laid out his playbook: make it the architect. From there we got into local models, why Dan priced out a home rig for his data-crunching needs and came back with a number north of $225,000, parabolic memory prices, and the strange norm-reversal where the US government gates model access like a state secret while China open-sources everything in sight. Dan floated a policy idea I suspect we'll return to: if the government takes stakes in frontier labs, then twelve-month-old models should go open source by default. The American people paid for them. A postscript from the editing desk, two weeks later: Belgium 4, USA 1. Spain won the whole thing. The $250 is gone, and the episode title wrote itself. Easiest money I've ever lost. Stay frosty, and go build something. Cheers, Sean Tools & Platforms Mentioned Kalshi — the prediction market where the $250 went to die. Disturbingly easy. Claude Fable — Anthropic's frontier model, back in our hands for the weekend; the "architect" in Dan's workflow. Codex — OpenAI's coding agent; Dan's current pick for the actual coding subagents. Qwen — Alibaba's open-source family; Sean runs Qwen 2.5 Coder 14B locally on a Mac Studio. Gemma (Google) & GLM 5.2 (Zhipu AI) — the other open-source models in the local-LLM conversation. Companies Discussed Kalshi, Fox, Telemundo, ESPN, Apple — the bet-vs-broadcast friction story. Anthropic, OpenAI, Alibaba, Google, Zhipu AI — frontier and open-source model landscape. Etched, Samsung, Intel — the chip-frontier segment. Amazon — the appealed yellow card. SpaceX, Porsche — the space-bubble aside and the $225K rig's opportunity cost. Harvard, UC Berkeley, Purdue, Indiana University — the education-and-training-grounds segment. Links & References Unqualified Advice — the website — back catalog, show notes, fact-check and prediction scorecards. Go poke at it. Invest Like the Best, EP.480 — Etched — founders Gavin Uberti and Rob Wachen on building AI hardware from first principles; the episode in Dan's queue. Etched exits stealth: $800M raised, $1B in contracts — the news behind the buzz, days before we recorded. Maxinomics on YouTube — Phil Andrews' channel; source of the elevator-operator jobs framing Dan cites. WEF: Automation has eliminated exactly one career in 60 years — the research (James Bessen) behind the elevator-operator claim. Tom's Hardware: memory prices keep climbing through Q3 2026 — receipts on "parabolic." TrendForce: Samsung restarts its 1.4nm push — the chip-roadmap news from the week we recorded. FIFA: Australia 1-1 Egypt (2-4 pens) — the match Sean attended in Dallas. ESPN: USA 1-4 Belgium — the Round of 16 result that settled the bet. RIP. Unqualified Fact-Check 🔍 We said some things. Here's how we did. 🟢 = Nailed it | 🟡 = Close enough | 🔴 = Whiffed it 🟢 Roughly 71,000 at Egypt–Australia in Dallas Sean said there were "71,000 people roughly in attendance." Official attendance at the sold-out AT&T Stadium match was 70,244, and yes — 1-1 after extra time, Egypt through 4-2 on penalties, their first-ever World Cup knockout win. Nailed the crowd, nailed the result. 🟢 The Kalshi math Sean said the US was ~4% to win it all, that round-of-16 "straight math" is 6.25%, and that the payout "goes something like 60,500." One-in-sixteen is exactly 6.25% ✓, and $250 at 4¢ contracts pays $6,250 — which is what "sixty-two fifty" sounds like to a transcription engine. We're scoring the math, not the microphone: green. 🟢 Memory prices went parabolic Sean said memory prices "did go parabolic." Understatement, if anything: DRAM contract prices rose 50–55% quarter-over-quarter into Q1 2026, some segments spiked 80–90%, Micron declared itself sold out for 2026, and PC makers announced 15–20% price hikes. (Dan's "more room to run" is now in the predictions file — analysts do see prices climbing through Q3 2026.) 🟢 Etched, the first-principles chip kids Dan said Etched rethought AI hardware from first principles, that the founders are "like 24," and that it's in his Invest Like the Best queue. All checks out: Gavin Uberti (Harvard dropout, early-to-mid twenties) and Rob Wachen build Sohu, a transformer-specialized ASIC; they appeared on ILTB EP.480; and they exited stealth on June 30, 2026 with $800M raised and $1B+ in contracts. Loose and goosey, yet correct. 🟡 1.4nm chips, and "current is three nanometer scale" Sean said Samsung and Intel are both working on 1.4nm chips (true — Samsung revived its 1.4nm push days before we recorded, targeting 2029; Intel's 14A is on deck for ~2027) and that they're "basically talking about the size of the transistors themselves." Two dings: the leading edge is already 2nm — TSMC entered volume 2nm production in late 2025 — and node names are marketing labels, not literal transistor dimensions. Right story, one node and one metaphor behind. 🟢 The elevator operator Dan said (via Maxinomics) that "the only time technology completely wiped jobs off the face of the earth was the elevator." This matches economist James Bessen's famous finding: of the 270 occupations in the 1950 US census, exactly one was eliminated by automation — the elevator operator. As paraphrases-from-memory go, this one's a clean lift. 🟡 "The computer revolution actually made more bookkeepers and accountants, not less" Half right. Accountants and auditors grew substantially in the spreadsheet era as the work moved up-stack — but bookkeeping and accounting clerk employment declined as VisiCalc and successors ate the data entry. The expansion story is real; it just didn't include the bookkeepers. 🟢 1929 ran on radio and aeronautics Sean said the 1929 bubble's leading technologies were radio and aeronautics. Correct — RCA was the era's glamour stock and aviation issues boomed after Lindbergh's 1927 flight. Everyone could see the technologies would win; almost nobody picked which companies would. Which was, of course, the point. 📋 The Appeals Court: Dan v. Fact-Check Department Dan appealed his 🟡 from No Movies About the Maintenance Team ("54 cents of every dollar earned on Amazon goes to Amazon"), citing a specific US-sellers-only dataset — a cohort with structurally higher costs than the global-average ~50% figure we checked against. The bench finds the appeal persuasive: upgraded to 🟢, with an asterisk that the number is scoped to US sellers. Per the boss's on-mic instruction, the fact-check department will henceforth check survey scope before scoring. The 85% hold rate ticks up accordingly. BONUS 🗓️ Sean said he attended the match "on Saturday." The Egypt–Australia game was played Friday, July 3. In fairness, it was a long holiday weekend, there were penalty kicks, and Hotel Mary's hospitality is reportedly excellent. Final Score: 6 🟢 / 2 🟡 / 0 🔴 The facts survived the weekend better than the $250 did. Chapters 0:00 — Cold Open: A Server Rack or a Porsche 0:37 — Happy 250th: A World Cup Fourth in Dallas 2:35 — The Easiest $250: Betting on the USA 4:07 — Easier to Bet Than to Watch 6:55 — Dan's Midwest Loop & a Tornado's Path 9:17 — The Appeals Court: Dan Contests a Yellow Card 11:28 — The Website & Our Scorecards 13:39 — A Weekend with Fable: Hire the Architect 17:20 — Local Models: Qwen, Gemma, GLM 18:43 — Reversed Norms: Gated Models vs. Open Source 21:30 — The $225,000 Basement Rig 25:10 — Etched: First-Principles AI Hardware 27:32 — 1.4nm and the Edge of Moore's Law 28:58 — Rethinking CS: Freshman Year in the CEO's Office 32:15 — AI Fear, Elevator Operators & Lost Training Grounds 39:52 — 10x People and the Inequality Narrative 41:55 — Boom or Bubble? Radio, Aeronautics & AI 44:30 — Should Everyone Own Stocks? 46:18 — Exclusivity as the Product 47:56 — Stay Frosty, Go Build Something

  2. Jul 20

    There is No Such Thing as Free Admission

    Talking to God for a Couple of Days Hello dear show notes readers! I came into this one buzzing. I'd just stood in a roaring crowd at the World Cup fan fest in Houston watching Team USA put four past Paraguay, and I wanted to talk about what that kind of communal energy actually costs — who pays for the security (FEMA and Texas, it turns out, which means we already did), who rakes in the revenue (FIFA), and why it's so hard to manufacture that "everyone shoulder to shoulder" feeling in a country built for cars. Dan and I wander from European piazzas to American sprawl to whether self-driving cars will quietly make sprawl worse, and land on a thing we keep circling back to: the third place, and how fragile and expensive it's become to just be around other people. From there it's a proper grab-bag. We get into the great game of mega-event ticket pricing — why you can't repeal supply and demand no matter how unfair $10,000 nosebleeds feel — and Dan points everyone at a Maxinomics video on the Ticketmaster antitrust mess. Then we turn to the SpaceX IPO I missed by four days: tiered lockouts, a two-trillion-dollar price tag at a hundred times sales, and the eternal tension between voting with your heart and weighing with your head. Dan walks through a lockout-timed options play that's paid him before, and we both agree that the people who buy on IPO day, look like dummies through a 90% drawdown, and just hold tend to win in the end. And then the part that gave the episode its name. For about two days, a handful of institutions and government agencies got their hands on Fable — the nerfed-and-released version of Anthropic's Mythos model, the most powerful thing anybody's built — before the government clawed it back. Dan burned something like eighteen thousand dollars of tokens in forty-eight hours and lived to tell the tale. We get into what it means to treat a model like a munition, why the diffusion happens anyway, and why both of us are increasingly tempted to just put a box in the basement and run an open-source model we actually own. The model is the commodity. The hardware is the value. Cheers, and as always — go build something that lets you touch more grass. Sean Books Discussed Reminiscences of a Stock Operator by Edwin Lefèvre — the source of "the first eighth and the last eighth are always the most expensive." Shows / Films Discussed The Wizard of Oz — invoked as the "man behind the curtain" metaphor for how hiding something only makes people want it more. Tools & Platforms Mentioned Claude Code — agent coding tool; the remote-control workflow we get into at the end. Codex — OpenAI's coding agent; also supports remote/mobile control. Qwen — Alibaba's open-source model family; what we'd run locally. NVIDIA DGX Spark — the "put it in the basement" home AI box. WEC (FIA World Endurance Championship) — Dan's subscription for watching the 24 Hours of Le Mans. Maxinomics — Phil Andrews' data-driven economics channel; the Ticketmaster/Live Nation antitrust video. Links & References Maxinomics on YouTube — the Live Nation/Ticketmaster breakdown Dan recommends. 2026 FIFA World Cup — 104 matches across the US, Canada, and Mexico, June 11–July 19, 2026. Triple Crown of Motorsport — Monaco GP, Indy 500, and the 24 Hours of Le Mans. Unqualified Fact-Check 🟡 "It goes through July eleventh, I think… a hundred and four games." — The match count is right: the 2026 World Cup is 104 games. But the tournament runs June 11 through July 19, 2026 (final at MetLife Stadium), so "through July 11" is off by about a week. The "almost a month" instinct is correct — it's closer to five and a half weeks. 🟡 "FIFA's gonna rake in about twelve billion dollars." — In the right ballpark, depending on what you're counting. Tournament-specific revenue estimates land around $9–11B; the full 2023–2026 four-year cycle is projected near $13B. So $12B is a fair splitting-of-the-difference, just worth knowing it blends two different figures. 🟢 "The first eighth and the last eighth are always the most expensive." — Correctly attributed to Reminiscences of a Stock Operator (the Jesse Livermore-inspired classic). The idea: don't try to catch the exact bottom or top; the extremes cost you the most. 🟢 "The Triple Crown is Monaco, Indy, and Le Mans." — Correct. The Triple Crown of Motorsport is the Monaco Grand Prix, the Indianapolis 500, and the 24 Hours of Le Mans. Graham Hill remains the only driver to have won all three. 🟢 "Porsche put the ignition on the left so drivers could start the engine while running to the car." — True. Porsche's left-side ignition is a genuine nod to the old Le Mans running starts, where drivers sprinted across the track, jumped in, and fired up. 🟢 "Mercedes left Le Mans after their car went airborne." — True. After the Mercedes CLR flipped at speed in 1999, Mercedes pulled its factory effort from Le Mans prototypes for a long stretch. 🟡 "Amazon and Facebook had 80–90% drawdowns and you'd be filthy rich if you'd just held from IPO day." — Directionally true and a great point, but the specifics are loose. Amazon did fall ~90%+ in the dot-com bust and recover spectacularly. Facebook/Meta's worst stretches were closer to ~50–75% (2012 post-IPO and 2022), not a clean 90%. The lesson holds; the matching percentages don't quite. 🟡 "Toby Lütke races an LMP2 car on a bronze license." — Lütke (Shopify's founder) is a real and serious amateur racer, so the spirit is right. We're less certain about the exact class and license detail, so call it broadly true pending confirmation. Note: the SpaceX IPO mechanics (tiered lockouts, ~$2T valuation, ~100x sales) and the Fable/Mythos/Project Glasswing storyline are live, fast-moving items we were reacting to in real time — treat the specific numbers as our best read on the day, not gospel. Score: 5 🟢 / 4 🟡 / 0 🔴. A clean week — the misses are date-and-decimal stuff, not blown calls. Mostly we got our facts right and our jokes righter. Chapters 0:00 — World Cup in Houston: USA Stuns Paraguay 4:57 — The Economics of a Mega-Event (FEMA, FIFA's $12B) 10:21 — Ticket Pricing, Maxinomics & the Taylor Swift Problem 13:41 — Third Places, Piazzas & American Sprawl 21:34 — The Trillion-Dollar Light-Rail Joke 24:20 — Why Le Mans? Endurance Racing & the Triple Crown 30:42 — The SpaceX IPO & Tiered Lockouts 35:16 — Playing the Lockout: Dan's DocuSign Put Strategy 38:50 — Voting Machine vs. Weighing Machine 43:47 — Staring at the Tape & the Three-Body Market 45:09 — Talking to God: The Fable Five Nerf 49:14 — Build Your Own LLM in the Basement 54:23 — ITAR, Operation Fury & the TAM Question 57:02 — Alien Documents & the Wizard of Oz 58:13 — Touch Grass, Build Something

  3. Jun 22

    No Movies About the Maintenance Team

    The Averaging Machine Hello dear show notes readers! Dan's wife and kids were off camping leaving him time to watch the whole of Glengarry Glen Ross in a single sitting. Why does nobody make movies like that anymore? Scale homogenizes. Once a film has to sell in China and Germany and everywhere at once, it gets sanded down to the middle. Dan claims that AI is the same kind of averaging machine globalization always was. I say the same corporate dispersion that flattens taste also spreads it. I pulled into a new Chinese chain half a mile from the plant in Texas, sat down, and it tasted exactly like what I ate in Chengdu. You don't get that without a machine pushing it outward. Homogenization and discovery are the same engine running in two directions. The back half turned macro. Dan walked through his Amazon Prime Day strategy: pulling ads to zero while everyone else bids each other into the ground, taking a worse revenue number for a much better profit one. You win by managing the downside, not chasing the upside. Then the paper-cuts of inflation: a little tariff, a little fuel, a little Hormuz, and a tungsten-carbide problem most people will never hear about that's quietly tripling my industry's tooling costs. Each piece feels transitory, but what if the sum is sticky? Which got us to Warsh taking the Fed chair and a genuinely fun thought experiment: if today's inflation is mostly a supply-side problem, does hiking rates actually make it worse? Go build something this week. Cheers, Sean Shows/Films Discussed Glengarry Glen Ross (1992) — dir. James Foley, written by David Mamet; Jack Lemmon, Al Pacino, Alec Baldwin, Ed Harris, Alan Arkin, Kevin Spacey, Jonathan Pryce A Few Good Men (1992) — Dan's "another small-set, big-writing movie" reference The Hunt for Red October (1990) — referenced for Alec Baldwin's prime (he played Jack Ryan) Scent of a Woman (1992) — referenced for Al Pacino's heavy-hitting years Tools & Platforms Mentioned Nano Banana — Google's Gemini image model; my example of the under-funded founder generating 500 versions of a vision for $100 Amazon Seller / Advertising — the Prime Day ads-strategy and B2B-placement discussion Walmart Marketplace — "Amazon five years ago" on ad maturity Companies Discussed Amazon (seller-side ads, fees, B2B placements) Walmart (marketplace maturity) Alphabet / Google (Berkshire's $10B placement) Berkshire Hathaway Anthropic (referenced in the "free call option on AI" riff) Apple / AirPods; Dyson (and the $100 knockoff) YGF Malatang — Katy, TX — the YangGuoFu Malatang location Sean stopped at ("two billion bowls served," tasted like Chengdu) SLB / industrial-consumables suppliers (Sean's input-cost discussion) Links & References Derek Thompson — the journalist whose recent interview with a neuroscientist (on how AI is reshaping how we think, read, and produce) anchored our front-half discussion Dan Wang — Breakneck — callback to our earlier talk about America drifting toward a "lawyerly state" Kevin Warsh — newly confirmed Fed Chair; Dan's referenced long-form Hoover Institution interview Berkshire Hathaway invests an extra $10B in Alphabet (CNBC, June 1, 2026) — the "free call option on AI" The midwit meme (IQ bell curve) — and the Dunning–Kruger effect it inverts Chinese tungsten prices surge in 2025 amid export controls (Fastmarkets) — the carbide/APT cost story Unqualified Fact-Check 🔍 We said some things. Here's how we did. 🟢 = Nailed it | 🟡 = Close enough | 🔴 = Whiffed it 🟢 The Glengarry Glen Ross roll call. Dan rattled off "Lemmon, Baldwin, Arkin, Spacey," and Pacino. All correct. The 1992 film (dir. James Foley, written by David Mamet) starred Jack Lemmon, Al Pacino, Alec Baldwin, Ed Harris, Alan Arkin, Kevin Spacey, and Jonathan Pryce — and yes, it's basically one room and a lot of rain, so the actors were the budget. 🟢 Baldwin "after Hunt for the Red October," Pacino "Scent of a Woman timeframe." Both sharp. The Hunt for Red October (1990) had Baldwin as Jack Ryan, two years before Glengarry — squarely his young prime. And Scent of a Woman (1992) is the same year as Glengarry — the one that won Pacino his Best Actor Oscar. Dead-on era placement on both. 🟡 "54 cents of every dollar you spend on Amazon goes to Amazon." Dan cited a survey of large sellers and flagged he's personally more efficient. The widely-cited figure (Marketplace Pulse) is that Amazon's total take — referral fee, FBA fulfillment and storage, plus advertising — reaches up to ~50% of seller revenue, with ads being the technically-optional piece that pushes it there. 54% is plausible for a specific large-seller sample, just a hair above the headline number. Directionally right; "every dollar" is per-seller-revenue, not per-consumer-dollar across all products. 🟢 China is "playing with the APT market," and carbide tooling costs have spiked. Correct, and Sean was actually conservative. China imposed export controls on tungsten products — including ammonium paratungstate (APT) — on February 4, 2025. Chinese tungsten-product prices surged more than 200% through 2025, and the European APT benchmark was up roughly 557% from its pre-control baseline by March 2026. Sean's "up about 3x since August" understates it; the China-as-cause attribution is spot on. 🟢 Warsh "is the chair, he's been confirmed, he's in the seat now." True. Kevin Warsh was confirmed by the Senate 54–45 on May 13, 2026, and sworn in as Fed Chair replacing Jerome Powell. He's also a Hoover Institution visiting fellow, so the long Hoover interview Dan watched checks out. 🟢 Berkshire put ~$10B into Google, a small slice of its cash. Confirmed. CNBC (June 1, 2026): "Berkshire Hathaway invests extra $10 billion in Alphabet, deepening bet on AI," bringing Berkshire's total Alphabet position to roughly $26.6B — still pocket change against its cash pile. Sean's "$400 in your pocket, you invest $10" analogy is about right, and "a free call option on AI" is a fair read of the structure. 🟡 "Donut theory" of interest rates. Sean floated the idea that inflation can rise if rates are too low or too high, with a sweet spot in the middle — and openly admitted he couldn't remember the source ("read it three years ago"). There's no established economics term called "donut theory," but the underlying intuition isn't crazy: it overlaps with neutral-rate (r-star) reasoning and the fiscal channel, where high rates funnel interest income to wealthy savers who then spend. We're scoring the instinct as defensible and the label as a Sean original. If anyone finds the original piece, send it our way. Final Score: 5 green, 2 yellow, 0 red. A strong showing for a Saturday-morning ramble. The film recall alone deserves a green star on the fridge. Dan, your Amazon math was close enough that we'll let the robots keep the change. Chapters 00:00 — Cold open: "There's no movie about the maintenance team" 00:14 — Welcome / "Are you ready for today?" 00:30 — Dan watches Glengarry Glen Ross in one sitting 01:14 — Homogenized Hollywood: movies as an averaging machine 01:56 — Franchise math: why the small films don't get made 04:17 — Derek Thompson and the Instagramification of everything 06:34 — Dispersion aids discovery ("tastes like Chengdu") 09:00 — "Zero value in the new economy": taste, judgment, synthesis 11:18 — Tool literacy and the technical-language advantage 12:40 — Nano banana and the under-funded founder's lookbook 13:55 — Declining-everything charts; creation as touch grass 15:44 — The midwit meme and the inverse Dunning-Kruger 18:12 — Shiny-object syndrome has a bright side 19:27 — The grind nobody posts 21:00 — No movie about the maintenance team (in context) 21:37 — The pit crew, Monaco, and managing the downside 23:17 — Amazon Prime Day: going for less, not more 26:18 — Conversion rate, deprioritization, and B2B placements 28:55 — Amazon vs. Walmart maturity; 54 cents on the dollar 30:58 — Replacement parts, knockoffs, and landfills 32:03 — "We've calibrated": the world hasn't slowed, but we have 32:45 — The paper-cuts of inflation: consumables and tungsten carbide 35:57 — The T-word, and Warsh takes the chair 36:51 — Donut theory: does hiking fuel supply-side inflation? 38:37 — Walking the chapters: housing and the pre/post-2022 divide 43:46 — Healthcare, childcare, food, energy 45:57 — Boomers, interest income, and where inflation hides 48:29 — "Let's take it down a point" / let her rip 49:18 — The cure for high prices; the risk of the next shock 51:05 — Berkshire's free call option on AI 53:05 — Equity vs. debt, WACC, and uncapped upside 54:35 — Sign-off: go build something / touch grass

  4. Jun 15

    Gone Building

    Hello dear show notes readers! This I ask Dan if he's been building anything lately, and we spend the hour discussing the ins and outs of building new tools and products. Two guys, exploring niches. That's the whole episode. Dan is launching Kabu (kabu research dot substack dot com, link below) — a paid newsletter and model portfolio centered on small and mid-cap Japanese equities, anchored on the post-Shinzo-Abe governance reforms and the roughly 2,000 names nobody is covering in English. We get into tier pricing, what makes a paid Substack actually worth paying for, and the practical compliance hygiene around publishing investment ideas — including Dan reaching for Lowe v. SEC, which the fact-check rewards him for. I'm fill Dan in on the real estate brokerage site I've been making for past guess and current business owner, Lindsay Howard, and a property-tax-allocation audit tool I had a vision for. That kicks us into the bigger question hiding underneath: when one citizen with a laptop can audit a municipal budget, what changes? Dan is bullish on the ground-up, people-doing-stuff version. We're both eyes-open about the pitchfork version. The honest answer is it goes both ways. Carve-outs to close: How to Win a Trade War by Soumaya Keynes and Chad P. Bown (the pirate-ship / warship / merchant-ship taxonomy alone is worth the read); Richard Hamming's The Art of Doing Science and Engineering (1990s prose that reads like it was written this week, especially on AI and human uniqueness); and Patrick Radden Keefe's London Falling and The Snakehead, plus the Wind of Change podcast. Go build something this week. Cheers, Sean Books Discussed How to Win a Trade War: An Optimistic Guide to an Anxious Global Economy by Soumaya Keynes and Chad P. Bown The Art of Doing Science and Engineering: Learning to Learn by Richard Hamming London Falling: A Mysterious Death in a Gilded City and a Family's Search for Truth by Patrick Radden Keefe The Snakehead: An Epic Tale of the Chinatown Underworld and the American Dream by Patrick Radden Keefe Tools & Platforms Mentioned Substack — Dan's distribution platform for kabu research Claude Code — Sean's build tooling for the Lindsay Howard site and the property-tax audit tool Interactive Brokers (IBKR) — referenced as the broker most likely to give US retail access to Japanese small/mid-cap names Charles Schwab — Dan's brokerage; reports no issues purchasing Japanese names GitHub — Sean's planned open-source venue for the property-tax tool Companies Discussed SLB (formerly Schlumberger; rebranded ~2022 to match its NYSE ticker) Halliburton; Baker Hughes (implied as peer-set) DynaDrill (an SLB sub-brand Sean works in) Saba Capital / Blue Owl (referenced obliquely via the private-credit thread carried over from prior episodes) Bloomberg L.P. NPR Planet Money (creators of @Botus / Bot Of The United States in the first Trump term) Wall Street Journal Lindsay Howard's brokerage (Sean's site project; Florida expansion incoming) Links & References Maxinomics on YouTube — Phil Andrews's data-driven economics channel; source of the flatbed-trailer thesis Dan walks through kabu research on Substack — Dan's newly launched Japanese small/mid-cap equity newsletter and portfolio Citrini Research — the independent-research benchmark Sean references for pricing and pre-publish trading discipline The Science of Hitting (TSOH) — Sean's referenced precedent: publish first, let readers trade, then act Wind of Change podcast (Patrick Radden Keefe) — the CIA / Scorpions story Dan mentions Unqualified Fact-Check 🔍 We said some things. Here's how we did. 🟢 = Nailed it | 🟡 = Close enough | 🔴 = Whiffed it 🟢 SLB rebrand was about four years ago. Dan said "we rebranded four years ago" (in response to Sean's joke). Schlumberger officially became SLB in October 2022. Four years from May 2026 lands at May 2022, close enough for radio. Solid. 🟢 Hamming worked on the Manhattan Project. Dan said Hamming "started his life and career at the Manhattan Project." Correct — Richard Hamming provided computing support at Los Alamos in 1945 before joining Bell Labs in 1946. Where he won the Turing Award (1968) for error-correcting codes. Both true. 🟢 Wind of Change is a real Patrick Radden Keefe podcast about the CIA and the Scorpions song. Dan got the show, the author, and the premise right. The Pineapple Street Studios podcast launched in 2020; eight episodes; explores the (still unconfirmed) theory that the CIA had a hand in writing "Wind of Change" as a Cold War psyop. 🟡 5% of US tractor-trailer supply is flatbeds. Dan paraphrased from memory. The precise share varies by data source — ATA / FTR reporting puts the flatbed share of US for-hire trucking capacity somewhere in the 5–8% range depending on how you cut it. Directionally right; the booked-out condition is the real signal, and Maxinomics is a credible source for it. (Bonus 🟡: the channel is Maxinomics, not "Maximomics" as Sean said. Phil Andrews, not "Max.") 🟡 France requires you to bequeath 75% of your estate to your kids. Dan said "if you want to give your money away, you can't. You can't, and if you have two kids and they're both heroin addicts on the street, you still have to give 75% of your estate to them." The French réserve héréditaire is real, but the reserved share scales with number of children: 50% for one child, 66% for two, 75% for three or more. Dan's directional point (Europe has a structural lock on inheritance vs. the US doesn't) is right; the specific 75%-for-two-kids math is a stretch. 🟡 "You can't trade ten days after publishing — Lowe v. SEC." Dan was citing Lowe v. SEC, 472 U.S. 181 (1985), and the case is exactly on point. Lowe is the foundational Supreme Court ruling that bona fide investment newsletters (regularly published, impersonal, generally circulated) are not "investment advisers" under the Investment Advisers Act of 1940 — they qualify for the §202(a)(11)(D) publisher exemption, even when they contain specific buy/sell commentary. This is the shield that lets kabu research, Citrini, TSOH, and every other Substack-shaped equity newsletter operate without SEC adviser registration. So the case is real and correctly recalled. The catch: Lowe itself doesn't impose a "10-day no-trade after publishing" window. That convention is the practical compliance hygiene publishers observe to stay within the Lowe exemption — if you trade around your own publications, you start to look promotional or interested, which forfeits the bona-fide-disinterested-publisher status the exemption requires. The companion anti-scalping doctrine is SEC v. Capital Gains Research Bureau (1963). Dan's instinct (don't pull the rug; pre-publish your exits) is the right one — it's what keeps you protected by Lowe. 🟡 Author's name: Chad Brown. Sean said "Chad Brown" at 42:13 then corrected to "Chad Brown" again — it's actually Chad P. Bown (no E). One letter; the man writes about tariffs for the Peterson Institute and runs the Trade Talks podcast. Sean is forgiven. 🟢 The Snakehead is about Cheng Chui Ping (Sister Ping) and the Chinatown smuggling network. Dan's recollection ("the lady who basically ran it for decades, literally") is correct. Cheng Chui Ping ran the snakehead human-smuggling operation in NYC's Chinatown from 1984–2000, with the 1993 Golden Venture grounding as the inflection point. Patrick Radden Keefe's The Snakehead (Doubleday, 2009) was a finalist for the J. Anthony Lukas Prize. Final Score: 4 green, 4 yellow, 0 red. Solid building-conversation accuracy. The case-law beat was the right case (Lowe v. SEC) — Dan just had to hear "Lowe" through the transcript robot before we got there. Dan, your law school called; they want their Lowe hypo back. Chapters 00:00 — Cold open: "Been building anything lately, Dan?" 00:10 — Welcome / an organic weekend 01:08 — Tariff reshuffling at SLB 02:12 — What marketing looks like at SLB 02:53 — "There's a group that just sells dirt" 03:24 — "See that water feature?" / activists and the case for diversification 05:25 — Order books: next year strong, this year wait-and-see 06:48 — Maxinomics, flatbeds, and the middle of the country 10:19 — Data centers, trades shortage, "at least into 2030" 11:01 — "Next week or August" and Hayekian tacit knowledge 11:50 — Dan Loeb and analysts who fly out 12:49 — Big Short alligator-in-the-pool callback 13:02 — "Been building anything lately, Dan?" (return) 13:12 — Gone Building: the Japanese-equities Substack thesis 15:24 — Why the timing works (foreign exposure, trend-following) 17:54 — Pricing tiers and product design 20:14 — Korean equity access and ADRs 21:26 — Why Japan's culture is changing 21:59 — France's 75% inheritance law and capital mobility 23:57 — Launch this week 24:19 — kabu research dot substack dot com 26:31 — Compliance: Lowe v. SEC and pump-and-dump prevention 27:27 — TSOH precedent: publish, then let readers trade 29:29 — Pre-publishing exits 32:44 — Citrini, $400/year, the independent-research economy 33:41 — "Buying IBM because Trump told me to" / Investor in Chief 34:08 — @Botus and the NPR Planet Money throwback 34:20 — Hidden government vs. flood-the-zone 35:55 — Sean's Lindsay Howard brokerage build 37:32 — The property-tax-allocation audit tool 39:08 — AI-enabled civic transparency, "ground up, people doing stuff" 41:52 — Carve-outs begin 41:56 — How to Win a Trade War (Keynes & Bown) 43:02 — Pirate ship, warship, merchant ship 44:34 — The Art of Doing Science and Engineering (Hamming) 46:50 — London Falling and The Snakehead (Keefe) 48:46 — Wind of Change podcast 49:25 — Sign-off: go build something

  5. Jun 8

    Frameworks in the Atom Economy

    Hello dear show notes readers! This week Dan and I open on the image that everyone seems to have fixated on — Jensen Huang in his black leather jacket on the Anchorage tarmac, allegedly traveling with just a backpack — and from there we slide into the actual substance of what happened at the Trump–Xi summit. The deliverable was, in Dan's word, frameworks rather than specifics: no signed trade deals, no headline numbers, but a tentative agreement to stand up a bicameral US-China trade council to replace what we ended up calling, with a straight face, "trade agreements by tweet." The thing that kept jumping out at me was the quiet admission underneath the framework talk. China and the US sitting down to formalize a direct two-party institution, with Russia visibly excluded. "You're a menace, but you ultimately don't matter," as Dan said. The choke-point math is what makes it real: 20 percent of global oil through Hormuz, half of the global container traffic through the Strait of Taiwan, and the UAE laying pipe to route around the first one. Oil will have a relief valve. Container ships won't.  From there we end up where the show always seems to end up these days — building material for what I keep calling our Promethean candidate. Dan introduces a frame I think we're going to be reaching for again: the atom economy. The bits era is over, atoms are next, and the next CapEx wave is energy grid, sensor networks, smart-cooling infrastructure — the public-infrastructure backbone that everyone plugs into. Tragedy of the commons, basically: nobody's market alone will fund this, and that's why it has to be policy. I revive my six-year single-term presidency idea, except this time the rationale is industrial-policy continuity — you vote for the direction, then you watch it actually get built across years two through five. We wind up in the deep, dark forest of US sales tax administration. Dan's lived experience here is its own argument. Forty-eight filings a month across nested Colorado jurisdictions, by his estimate 85 percent of small businesses out of compliance, and an entire shadow industry of compliance labor that has mostly migrated to the Philippines and El Salvador. I caught myself realizing that the compliance burden is effectively a tax on Americans that funds growth in other countries. That can't be how this is supposed to work.  Thanks for sticking with us. Have a good week, and as always — go build something. We'll see you next time. Cheers, Sean Books Discussed Why Nothing Works by Marc J. Dunkelman — Subtitle: Who Killed Progress—and How to Bring It Back. The progressive-paradox book — we want someone to lead this, but not like that — anchored a big chunk of our top-down-vs-bottoms-up wrestling. Tools & Platforms Mentioned Polymarket — referenced in passing in the broader prediction-markets backdrop Nest — the smart-thermostat example for grid-efficient cooling at scale TurboTax (Intuit) — exhibit A in the tax-compliance rent-seeking layer H&R Block — exhibit B Companies Discussed Nvidia — Jensen Huang's leather-jacket cameo on the Anchorage tarmac Walmart — Dan's initial guess for the largest US employer (it's the largest private employer) Intuit / H&R Block / TurboTax — the income-tax compliance ecosystem Nest (Google) — distributed grid efficiency Rolex / Hermès — luxury-tier illustrations in the consumption-tax thought experiment Links & References Trump–Xi summit readouts (October 2025) — official US readout (search "Trump Xi Anchorage summit") Xi Jinping's four-pillars language — Chinese MFA readout (search "four-pillars cooperation") Strait of Hormuz / EIA petroleum chokepoint analysis — the 20%-of-global-oil reference Strait of Malacca container traffic — UNCTAD context — for the choke-point comparison China's 15th Five-Year Plan — overview — Chinese government information portal US Census Bureau — US Trade in Goods with China — official trade-balance data for the down-28%-YoY and down-31%-YoY claims Tang ping / "lying flat" — background — the Chinese youth movement Dan referenced SBA Office of Advocacy — small business GDP share — for the ~44%-of-US-GDP-from-small-business claim E-Estonia (Estonia digital government) — the federal-tax UX benchmark we kept reaching for Charlie Munger — "show me the incentives" — Wikiquote source for the line Abe Campbell on X (search by handle) — couldn't pull the exact post date on-mic; readers, if you find the original tweet, tag us Unqualified Fact-Check 🔍 We said some things. Here's how we did. 🟢 = Nailed it | 🟡 = Close enough | 🔴 = Whiffed it 🟢 Strait of Hormuz ≈ 20% of global oil Sean (~04:20): "the Strait of Hormuz, which is 20% of the global oil supply." The EIA's most recent chokepoint analysis puts roughly 20% of global petroleum liquids trade through Hormuz. Sean is in the right neighborhood. 🟡 Strait of Taiwan ≈ 50% of global container traffic Sean (~04:20): "50% of the global container traffic passes through the Strait of Taiwan." The Strait of Taiwan does carry a very large share of Asia–US-bound container traffic, and roughly half of the global container fleet passes through it each year. But the more commonly-cited "global chokepoint" headline is the Strait of Malacca (around 25–30% of global trade by value). Sean's figure is defensible for ship-counts but is easy to misread as "50% of all global trade by value." Partial credit, with a footnote. 🟢 Tariff trajectory: 47.5% from a May-2025 peak around 127% Sean (~05:18): "tariffs are down to about 47.5% at the current state. So from 127.2 of May of 2025." This tracks with the Peterson Institute and Tax Foundation effective-rate trackers — peak China-applicable tariffs in May 2025 were in the high-120s, and they've come down through 2025 negotiations into the high-40s. Within rounding. 🟢 2026 is year-1 of the 15th Five-Year Plan Dan, reading Google's AI Overview (~12:21): "2026 is the first year in the running five-year plan. This is the 15th five-year plan." Correct — the 14th plan ran 2021–2025; the 15th plan runs 2026–2030. 🟡 China youth unemployment "like 25%" Dan (~14:08): "youth unemployment there is like 25%, something crazy." The pre-2024 NBS series did reach 21.3% in June 2023 before publication was suspended. Under the revised methodology (excluding students searching for work), the 2025 figures have run roughly 14–17%. The framing is directionally correct — youth unemployment is structurally elevated — but the specific "25%" figure was the pre-revision number; current readings are lower. Partial credit. 🟢 Charlie Munger maxim Dan (~41:09): "Charlie Munger's show me the incentives and I'll show you the outcome." Correctly attributed; the maxim is one of Munger's most widely-quoted lines. 🔴 "Funded through tariffs and property tax" pre-income-tax Sean (~41:25): "Without the income tax, we were funded through tariffs and property tax." Pre-1913, the federal government was funded primarily through tariffs and excise taxes (whiskey, tobacco, etc.) — not property tax. Property tax is and has been primarily a state and local revenue source. The big idea (federal revenue used to come from somewhere other than income tax) is right; the specific second category is wrong. 🟡 "Early 19th century" growth in federal scope Sean (~41:37): "It really grew in size in the early 19th century, particularly through the war." Sean almost certainly meant early 20th century — the 16th Amendment authorizing federal income tax was ratified in 1913, and the major federal-scope expansions were New Deal–era. Verbal slip on the century, but the historical thrust is correct. 🟡 "44% of US GDP is small business" Sean / Dan (~32:05): "44% of USA GDP is small businesses." The most recent SBA Office of Advocacy figure puts small businesses at roughly 43.5% of US GDP. Close enough. 🟢 Federal government as the largest US employer Sean (~29:43): "the federal government, I'm pretty sure" is the largest US employer. Walmart is the largest private employer (~1.6M). The federal government employs roughly 2.1M civilians plus ~1.3M active-duty military — so in aggregate, Sean is right. Final Score: 5🟢 / 4🟡 / 1🔴 The geopolitical math holds up; the tax-history detour wanted one more pass before it left the room. Strong on what mattered. Chapters 00:00 — Cold open: "the entire world wants access to China" 00:21 — Welcome / Saturday hellos 00:45 — Jensen Huang on the Anchorage tarmac 01:14 — The Trump–Xi summit: frameworks, not specifics 04:20 — Chokepoints: Hormuz, Taiwan, Malacca 05:18 — Tariffs at 47.5% and the trade-volume collapse 06:44 — UAE pipelines and the Hormuz toll booth 08:58 — The bicameral trade council 09:28 — Russia gets sidelined: the bipolar admission 10:55 — Xi's four pillars 12:21 — The 15th Five-Year Plan reads through 14:08 — Lie flat: when the S-curve flattens 15:08 — Top-down vs. bottoms-up: the answer is both 16:37 — Why Nothing Works and the progressive paradox 18:49 — Building a Promethean candidate: industrial policy 21:10 — Six-year single-term presidency 22:24 — The atom economy: CapEx as the next regime 24:47 — Texas Triangle / look west for failure 25:30 — Tide goes out, boats on rocks 26:57 — 85% per layer: compounding bureaucratic loss 27:42 — Dan's 48-filings-a-month sales-tax rant 30:16 — Plug into one big pillar: federal sales-tax design 31:27 — Why your tax bill should come pre-filled 35:09 — Compliance as a tax on Americans funding offshore labor 40:12 — Abe Campbell's 100% inheritance / 0% income proposal 41:09 — Charlie Munger and the history of US taxes 44:20 — Tax consumption, not production: the Birkin tier 45:54 — Future-episode bookmark

  6. May 31

    The Canary in the Gig Mine

    Three Weeks to Get Numb Hello dear show notes readers! This week Dan and I sit with a question that's been quietly bothering both of us: have we all just gotten numb? Dan says it only takes three weeks to get numb to anything. I had to agree. Once we saw two data points for the same trend, the rest of the conversation flowed. If the standard warning system has gone quiet, what canaries are still working? We began to build a bellwether basket. Not Netflix (too sticky), not Amazon Prime (it delivers the diapers — that's a bottoming signal, not a stress signal), but the second-tier streaming subs (Paramount, HBO) where the swing voter actually shows up. Dan offered the cleanest signal of the episode: at DoorDash, don't watch order count — watch the delta between pickups and deliveries. When pickups maintain while deliveries fall, that's the belt actually tightening. We talked through coupon-scan volume as a possible two-sided-marketplace business (somebody build this), median rent, dating-app subscriptions, home-security subs. I told Dan about my own confession — I hadn't watched grocery unit prices in twelve years and now I check them every trip. Dan said the same thing started for him about eighteen months ago. We are, apparently, our own data points. Dan plants a thesis for September. Strategic oil reserves are at or near zero in pretty much every country outside the US, the Northern Hemisphere heating season is about to kick on, and farmers in the Philippines and Thailand are under-fertilizing for the 2026 rice cycle. He called it a "triple quadruple whammy." Mark your calendars — that one's specific enough to be falsifiable, and it's going into the predictions file. Then we shifted to the macro question: US debt now exceeds GDP for the first time since 1946. Japan's been over 200% for a long time, so it's not unprecedented, but it does narrow the policy options. Dan and I came down hard on the same side: you don't coupon your way out of this — you grow your way out. Dan's working frame is that AI might be the productivity multiplier that makes that math work. "If the rare unicorn engineer in Silicon Valley was the 10x engineer," he asked, "what if everyone worth 50% IQ and up is now a 10x person?" That's the bet. Pair it with immigration and assimilation (which we still do better than anyone) and the path is at least findable, even if the politics around getting there are nuts. He closes us out by quoting Marco Rubio — yes, I know, that Marco Rubio (and yes, the Sad Marco meme remains my favorite of the year) — but his recent line on national exceptionalism is the kind of thing Dan wants to bottle up and give to everyone: we are exceptional, the color of your skin doesn't matter, your education doesn't matter, all you have to do is get after it. I won't argue with that. Thanks for sticking with us through a slightly more diagnostic episode than usual. The numbness is real, but the signals are still out there if you know where to look — and the cure for high prices, as Dan likes to say, is high price. Creation fills, consumption drains. Go build something everybody. — Sean Tools & Platforms Mentioned FRED — St. Louis Federal Reserve Economic Database — the gas-prices-as-percent-of-take-home-pay data source. The series you want is Federal Debt: Total Public Debt as Percent of Gross Domestic Product (GFDEGDQ188S) for the chart we walked through. SEC EDGAR — the disclosure-data source I named for the bellwether-tracker project I keep promising to ship. David Deutsch's Twitter (@DavidDeutschOxf) — Dan recommends turning notifications on; he's all over what's happening in the UK right now. Companies Discussed Streaming / subscription: Netflix, Paramount, HBO, Amazon (Prime) Food delivery: DoorDash, Uber Eats, McDonald's Retail / consumer: H-E-B (Texas grocery), Walmart, Coca-Cola, Procter & Gamble, Heinz, Tide, Johnson & Johnson Home security: ADT, SimpliSafe Dating / social: Bumble, OkCupid (historical) Luxury / event: LVMH (Hermès Birkin), Live Nation, Ticketmaster, Miami Grand Prix Sports / live: NWSL — Houston Dash, Denver Summit Links & References FRED chart: Federal Debt as % of GDP — the line we walked: 40% in 1942 → ~105% post-WWII → ~23% by 1974 → mid-90s peak around 90% → 2001 local bottom → 100.2% today. IMF Japan debt-to-GDP historical — for the comparison-case figures Dan cited (232–260% range over recent years). Trading Economics — Japan Government Debt to GDP — Dan's specific source for the Japanese numbers. Wall Street Journal — US debt now exceeds the size of the economy — the chart Dan shared; sourced from White House OMB, Treasury, BEA. Howard Marks memos at Oaktree — referenced in spirit (cure-for-high-prices-is-high-price register). Hilbert's hotel / "infinity hotel" — the David Deutsch concept that came up again in passing. Unqualified Fact-Check 🟢 Sean (~03:16): "FRED, the St. Louis Federal Reserve Economic Database… you can go in and find these great data sets that go back 100 years in some cases." Correct. FRED is the canonical free macro-data archive; many series go back to the 1920s–1930s, and gas-price plus median-income data are both available there. The exact "gas price as percent of take-home pay" ratio Sean references isn't a packaged series, but it's straightforwardly computable from FRED inputs and is, directionally, at multi-decade lows. 🟢 Sean (~33:30 → 37:00) US debt-to-GDP walk: "40% in 1942 → about 105% by 1950 → bottoms around 23% in 1974 → climbs to about 90% in early 1990s → local bottom 2001 → 100.2% today." Strong. All figures align with FRED's GFDEGDQ188S series (Federal Debt: Total Public Debt as Percent of Gross Domestic Product). The 23% trough in the mid-1970s and the 100%+ figure today are both correct to within a point. 🟢 Dan (~37:17 → 37:51) Japan debt-to-GDP: "232 to 260% as of 2024 to 2025… high of 258 in 2020." Directionally accurate. IMF's general-government gross-debt-to-GDP for Japan was ~257% in 2020 and is currently estimated in the 234–250% range for 2025. Dan's numbers are within 2–3 points of consensus. 🟡 Dan (~13:43): "Christmas isn't happening" / inflation prediction walk-back. This is an in-episode revision of a prior call from May 2025 (Calling All Mental Nomads, where Dan's "$1,600 Nintendo Switches" prediction was made). Dan revises in real-time: the prediction didn't come true the way he expected. Tracked in the Canon Predictions file as christmas-2025-supply-shock; this episode is the on-mic acknowledgment that the prediction came in softer than forecast. Yellow rather than red because the prediction was explicitly flagged as "exaggeration, directionally meant seriously" when made. 🟢 Dan (~31:38): "You make a bad tweet [in the UK] and the police seem to be showing up your door." Real phenomenon. The UK's Communications Act 2003 §127 and Public Order Act 1986 have been used to prosecute social-media speech, including multiple high-profile cases in 2024–2025 (Allison Pearson, the Southport-riots prosecutions, the Bee on the Wall case). Directionally accurate. 🟡 Sean (~10:43): "Caviar is on trend right now." Half-confirmed. There's been a real TikTok/Instagram-driven "caviar bump" in 2024–2025 (multiple trade-press pieces flagged it; Imperial Caviar and The Cure Company ran the social campaigns). Whether it qualifies as "on trend right now" depends on the recording date — the trend may have already peaked. 🟢 Dan (~30:19): "Never bet against humanity / never doubt human ingenuity." This is a running theme of the show (Dan: "we've said that's been a running theme of ours"), and the inheritor of my Because Molecules closer "never bet against ingenuity." Promoting from UNSURE to Confirmed in the show's Canon RunningBits file. Final score: 5🟢 / 2🟡 / 0🔴. One-line summary: A diagnostic episode with no real errors — the macro figures are accurate, the directional claims hold up, and the only yellows are an in-show prediction revision and a "is this still hot?" trend question. Chapters 0:00 — Cold open: three weeks to get numb 0:17 — Sign-on / Houston heat / NWSL 1:30 — Mud season in Crested Butte 2:30 — Have we all gotten numb? (the body version) 3:00 — Fuel prices vs. take-home pay (FRED) 5:00 — Hardship vs. going backwards 6:30 — Building the bellwether basket: streaming subs 8:20 — DoorDash: pickup vs. delivery as the belt-tightening signal 10:00 — Birkin bags, caviar, and signals from the rich 14:15 — The cure for high prices is high price 15:00 — The September thesis (oil + heat + crops) 16:30 — "The Adams keep moving" — smuggler's paradise 18:30 — Food inflation propagation, rice's reach 20:00 — The hosts themselves as bellwethers: unit-price scanning 21:30 — Coupon-scan volume as alt-data 23:30 — Maslow's hierarchy: rent, safety, gym subscriptions 25:30 — Dating apps and gamification fragility 27:20 — The body could puke it up 30:15 — Never bet against humanity / sex or drugs 30:45 — AI safety, free expression, and oblique prompts 31:35 — UK police and bad tweets 33:30 — WSJ chart: US debt > GDP, 1942–today 37:15 — Japan as the comparison case (232–260% of GDP) 39:00 — Drive it down vs. expand out of it 39:30 — The 10x-engineer-for-everyone thesis 42:00 — Mind virus and the messaging problem 42:30 — Hormuz callback / we got used to it 43:30 — Birth rate, immigration, growth levers 44:20 — 300k-from-side-hustles as the new normal 46:10 — The Marco Rubio "bottle it up" line 47:25 — Wrap: go build something

  7. Apr 27

    Rules for Thee, Alpha for Me

    Two Hundred Bucks and a Stock Tip Hello dear show notes readers! This week Dan and I sit with a question that has been bothering me for a while and got sharper this week thanks to a perfect natural experiment: a US Army NCO got charged for making four hundred thousand dollars on Polymarket betting on the Maduro raid he was part of. Around the same time, Senator Markwayne Mullin bought stock in Chevron — the only US-listed company with Venezuelan oil operations — days before the same operation. The senator's position is up about twenty-two points over the S&P. The soldier is going to prison. The senator is going to a committee hearing. We get into why that gap exists, what the data on congressional trading actually shows (it's more nuanced than the headlines), why the STOCK Act has a two-hundred-dollar fine and zero prosecutions in roughly a decade, and whether the answer is more rules or more sunlight. Spoiler: I came in believing one thing and Dan talked me partway out of it. We also fall down a few side roads we couldn't resist — the French guy who spoofed a Polymarket weather contract by holding a hairdryer to the airport temperature sensor (still my favorite grift of 2026), the term-limits and lifetime-appointments question, and a riff about reading David Deutsch and Rory Sutherland in the same week. Dan plants a seed at the end about something we've both been wrestling with: what happens to investing when you bolt a probabilistic machine (LLMs) onto a deterministic one (the legacy quant playbook). That one's coming in a future episode. Thanks for sticking with us. We are, by Dan's accurate diagnosis, "the weird ones because this is what we choose to do with our weekend." Go build something. — Sean Books Discussed The Beginning of Infinity by David Deutsch — Dan's "infinity hotel" reference is from this book. Dense but rewarding. Alchemy: The Surprising Power of Ideas That Don't Make Sense by Rory Sutherland — I'd reread it tomorrow. Sutherland: please write another one. Money Stuff (Bloomberg newsletter) by Matt Levine — Dan's source for the Tesla-SpaceX-merger reasoning. Subscribe. Tools & Platforms Mentioned Polymarket — prediction market; site of both the Maduro-raid bet and the hairdryer hack Kalshi — regulated prediction market Capitol Trades — primary public source for congressional trade disclosures Quiver Quantitative — best aggregated data and portfolio-return modeling on Congress Companies Discussed Chevron (CVX) — the Mullin trade Tesla (TSLA), SpaceX, xAI, Cursor, US Steel — all in the Musk-merger riff Renaissance Technologies, Citadel, Jane Street, AQR — the legacy-quant universe in the AI-investing plant Google Cloud — financial-services partnerships (CME, Citadel) Links & References NANC — Subversive Democratic Trading ETF — tracks Democratic Congress trades; trailing 12-month return ~30%. KRUZ — Subversive Republican Trading ETF — tracks Republican Congress trades; trailing 12-month return ~18%. Pelosi Tracker (pelositracker.app) — the dedicated single-name tracker. STOCK Act (2012) — full text on Congress.gov Hawley press release — PELOSI Act advances out of Senate committee, July 2025 Ziobrowski et al. (2004) — Abnormal Returns from the Common Stock Investments of the U.S. Senate — the original ~85 bps/month finding. Belmont et al. (2022) — Capitol Losses — the post-STOCK-Act revised verdict. WSJ — Soldier charged with $400K Polymarket bet on Maduro capture (search "Polymarket Maduro soldier") Hilbert's Hotel / "infinity hotel" — the concept Dan referenced from The Beginning of Infinity Unqualified Fact-Check 🔴 Sean (~17:18): "Back before 2004 when they passed the Stock Act, I think it was 2004…" The STOCK Act was passed in April 2012, not 2004. The 2004 date is the Ziobrowski et al. study, which analyzed 1993–1998 Senate trading data and found ~85 basis points/month abnormal returns. Different document, ten years apart, both important — Sean conflated them. (The directional point — that congressional alpha looked larger before the legislation forced disclosure — is correct.) 🟢 Sean (~09:43): "I'm sure that there's a system where if you're an NCO, you can only go to..." Correct. US Army NCOs run E-4 (Corporal) through E-9 (Sergeant Major of the Army). Above that are warrant officers (W-1 to W-5) and commissioned officers (O-1 to O-10). NCOs do max out at E-9 — Sean was right (and Sean explicitly asked us to fact-check this on-mic). 🟢 Sean (~12:08–12:23): "NANC, plus about 30%... S&P plus 25%... Republicans... trailing the market by a good 8 percentage points." Approximately correct. Trailing-12-month: NANC ~28.78–31.54%; KRUZ ~18.14%; SPY ~25%. Sean was within a point on each. 🟡 Sean (~17:55): "Top decile of Congress trades at about plus 30%, which is impressive that that's over every era." Directionally correct but slightly oversold. The Ziobrowski (1993–1998) and Belmont et al. (2022) papers find that the top decile of Congress members significantly outperforms the market, but +30% as a constant across all eras is a rough approximation. The actual top-decile alpha varies by year and era; +30% is roughly right for the most-active high-frequency traders in recent years. 🟢 Sean (~20:07): "$200 fine... zero prosecutions on these 200-plus incidents since [2012]." Correct on both counts. STOCK Act late-filing fine is $200 (with a member-applied waiver path); per multiple Sludge / Business Insider audits, dozens of late-filing violations have been documented and zero criminal prosecutions have occurred. 🟡 Sean (~19:42): "since 2014, [late reports have] been going up every year from single digits. Now we're seeing about 50 trades a year that aren't reported on time." The trend direction (rising late-filings) is correct. The specific year-over-year shape is approximate — different audits use different definitions of "late." 50/year is in the right neighborhood. 🟢 Sean (~12:14): "Mark Wayne Mullen right there with his Mark Wayne Mullen trade wasn't in great size — only $50,000 when I saw reported." Correct. Mullin's Chevron disclosure was $15,001–$50,000 (Capitol Trades / Senate disclosure), purchased Dec 29, 2025; the position is up ~24% with SPY up ~3% over the same window. Final score: 1🔴 / 2🟡 / 4🟢. One-line summary: Strong on the data, sloppy on the date — the STOCK Act is a 2012 law, not a 2004 one, but the directional argument holds. Chapters 0:00 — Cold open / cemetery bike ride 2:24 — Sean's lawyer routine: setting up a non-US VPS 3:48 — The rules don't apply to them: Senator vs. Soldier 6:18 — The Charles de Gaulle hairdryer hack 8:13 — Polymarket fraud cases and Special Forces NCO 11:22 — NANC vs. KRUZ vs. SPY: the Congress-tracker ETFs 13:35 — No incumbents: term limits across all three branches 16:30 — The STOCK Act, $200 fines, and zero prosecutions 21:00 — What would real reform look like? 24:10 — Sunsets, DOGE, and reform-by-iteration 26:55 — David Deutsch, Rory Sutherland, and the infinity hotel 28:30 — The Tesla/SpaceX merger prediction 36:30 — Plant: probabilistic vs. deterministic in AI-driven investing 44:00 — Wrap and sign-off

  8. Apr 20

    Prometheus Politicking

    Stealing Fire for the People Hello dear show notes readers! Dan's been thinking about Prometheus — not as a myth, but as a political archetype. The person who steals fire and hands it to the people is a very different animal than the one who steals it and keeps it for themselves. That framing runs underneath the whole hour. We start with Dan's top-down vs. bottoms-up read of American politics, wander through Christopher Alexander's A Pattern Language and what organic growth looks like (London streets vs. Manhattan grid, Dresden rebuilt block by block), and land on the labor market — where Dan drops the line that carries the rest of the episode: if people are in pain, they'll vote. The question is what they vote for. So we spent most of the hour sketching a reform menu — tax code, trade, safety net, government structure, civic fabric — and what candidate we'd actually want running on it. Burke shows up, as does the idea that representatives were supposed to be deliberators, not delegates. Somewhere around the 51-minute mark I asked Claude on-mic to stitch all of this into a candidate profile, which you'll find right below. Thanks for sticking with us. A little heavier than usual, and we think the weight is earned. Cheers, Sean "If people are in pain, they'll vote." The Promethean Candidate (v0.1) A first-pass profile of the candidate we'd actually want to vote for. Not a platform. A posture. Core disposition Deliberator, not delegate — per Burke, willing to disagree with constituents when the evidence demands it, and willing to explain why. Organic, not imposed — comes from the community they represent, not parachuted in. Signals less, builds more — comfortable being boring on cable news. Trade Regulatory parity on imports — foreign producers shouldn't enjoy lower compliance costs than domestic ones. Strategic, not blanket, tariffs — targeted at genuine national-security or parity issues. Honest about the losers of trade liberalization, and willing to fund real transition support rather than empty retraining rhetoric. Tax Code Radical simplification — fewer brackets, fewer deductions, shorter forms. Hostile to rent-seeking by tax-prep intermediaries (see: Intuit). Willing to raise revenue where the math demands it, not just cut. Safety Net A real floor, with fewer strings — simpler programs, less means-testing theater. No benefit cliffs that punish the single mom for earning a raise. Unemployment paid as a lump sum when it buys real mobility, not drip-fed weekly as a posture of distrust. Housing supply taken seriously as a safety-net issue, not just a market one. Civic & Communal Fabric A mandatory year of service out of high school — military, community, or public works — explicitly designed to mix people across geography and class. Tax incentives for the physical places where people actually gather: pubs, pickleball courts, civic clubs, third places. Civic education reform, taken personally. What the profile is not Not a platform of purity tests. Not a single-issue candidate. Not someone who thinks the answer is more viral moments. — v0.1. We'll steel-man it, stress-test it, and keep building. Proposed Constitutional Edits The structural changes the Promethean Candidate can't deliver alone — these sit above the candidate level and would need amendment, major statute, or a genuine constitutional moment. One six-year presidential term, no reelection — space to plan without campaigning through the job. Upper and lower age limits on office — not just a floor; the cognitive-decline ceiling is overdue. Sunset clauses on every bill touching the power of the purse — nothing funded in perpetuity by default. Gerrymandering gone — county lines or nothing. Districts should follow geography, not incumbents. Campaign advertising gated to a short window before the vote — paid advertising limited in time; earned media and direct voter engagement are unaffected. Books Discussed A Pattern Language by Christopher Alexander Companies & Organizations Mentioned Intuit / TurboTax — the tax-code-complexity lobby in residence Bureau of Labor Statistics — the jobs data in question Links & References BLS — Employment Situation — the non-farm payroll data behind Sean's charts Edmund Burke, Speech to the Electors of Bristol (1774) — the "deliberator, not delegate" source Ray Oldenburg — The Great Good Place (third places) Prof G Markets — Catherine Edwards interview — the labor-market conversation Dan referenced Unqualified Fact-Check 🔍 We said some things. Here's how we did. 🟢 = Nailed it | 🟡 = Close enough | 🔴 = Whiffed it 🟢 Burke really did say that. Dan attributed to Edmund Burke the idea that a representative is a deliberator, not a delegate. Real — from Burke's 1774 Speech to the Electors of Bristol, where he argued his constituents owed him their trust to exercise independent judgment. Textbook-accurate. 🟢 BLS revisions are the real story. Sean's chart and Dan's read on it — the economy has been adding fewer jobs than originally reported and benchmark revisions have been consistently downward — is directionally accurate and matches multiple quarters of labor data. 🟢 Third places — Ray Oldenburg. Dan's attribution is correct. The Great Good Place (1989). Right concept, right source. 🔴 Intuit's market cap. Sean said Intuit was "10 to 20 billion." Off by roughly an order of magnitude — actual market cap is around $97–109B. The larger point about Intuit lobbying to keep tax prep complicated is accurate, but the number was way off. 🟡 Pattern Language as political metaphor. Using Christopher Alexander's A Pattern Language as a metaphor for organic political movement isn't something Alexander himself argues, but it's a fair extension of his framework. Partial credit. Final Score: 3 green, 1 yellow, 1 red. One order-of-magnitude miss, one stretched-but-earned metaphor, and Burke holding the scorecard together. We'll know Intuit's market cap next time. Chapters 0:00 - Cold Open (Burke on Representatives) 0:40 - Show Open & Welcome 1:34 - Headlines, Reality, and the Photographer's Frame 3:53 - Bottoms-Up vs. Top-Down: China and the US 5:47 - A Pattern Language and Organic Cities 10:19 - Dresden, Bauhaus, and Rebuilding 13:35 - "History Rhymes" 14:27 - Labor Market Warning Signs 16:40 - The Generational Squeeze 18:47 - Building an Ideal Candidate 19:03 - Reform Menu: Tax Code (and the Intuit Problem) 19:43 - "If People Are in Pain, They'll Vote" 23:01 - Reform Menu: Trade and Import Parity 24:03 - Reform Menu: The Safety Net 28:33 - Sunset Clauses on Every Bill 29:44 - Prometheus, Burke, and Deliberation 31:30 - Reeducating on What Government Is For 34:49 - The Performative Culture War 35:26 - Reform Menu: Government Structure, Term Limits, Age Limits 38:53 - Gerrymandering and Campaigns 43:07 - Family Size and Communal Fabric 46:34 - A Mandatory Service Year 50:02 - The Lost Third Places 51:29 - Host Note: Ideal Candidate Profile 53:51 - Closing Thoughts & Sign-Off

See All (51)

Ratings & Reviews

4.8
out of 5
4 Ratings

About

Hello and welcome to Unqualified Advice, an entertaining show for entertainment purposes. Join us as we talk about running our small businesses, what we've been learning, and how we're applying lessons from academia and real life as entrepreneurs and investors.

Information

  • Creator
    Sean Filipow and Daniel Hatke
  • Years Active
    2024 - 2026
  • Episodes
    51
  • Rating
    Explicit
  • Copyright
    © 2024
  • Show Website
    Unqualified Advice