← Paulo Macro hub  ·  Research hub  ·  Research library

Fly on the Wall with Shrub & Cloudbear

"Revisiting the bonds, AI, and risk" — a month after the MUM episode: the long bond is the trade everyone mocks, so Le Shrub keeps a small, asymmetric TLT option bet that Bessent, "two out of two" on crude and the yen, can manage the long end too; the Anthropic IPO "has to succeed," so expect a month of bullish AI stunts — while Paulo doubts it makes the window, and sees the West "speaking Portuguese": Brazil-style fat real yields pulling fully-funded pensions out of equities and into bonds.
2026-SEP-10 · PauloMacro (Substack, PAID) · audio conversation (36:14) with Substack auto-transcript — Substack player positions, no YouTube deep-links · ↗ Listen / read original · transcript · actionable insights
Two speakers. A conversation between Le Shrub (author of the "Claudification Overdrive" piece; owner of the TLT bond-optionality trade) and Paulo Macro ("Cloudbear"), on Paulo's Substack. The Substack transcript carries no speaker labels and breaks on pauses rather than speaker turns, so attributions below are inferred from content (as on the Aug-10 episode). The (m:ss) markers are positions in the Substack audio player.
One-line take: the tape is "a bull market in memes" — $5,000 stimulus checks ("a cool trillion… more than we threw at COVID") into 30s rounding 5.3 and 10s knocking on 5, so "no stopping this train means it goes off the rails." Le Shrub: the midterm risk window delivered "a massive chop," and the trade everyone makes fun of is bonds — which is exactly why he is "still in the TLT trade": a defined-risk option position ("I'm risking 30 bips total… 0.3% of my portfolio") that could make "10 times my money," held because Cramer, a Barron's cover, Lacy Hunt and Dalio have all turned bond-bearish, and because Bessent is "two out of two" managing crude and the yen (MUM) — "the only thing that's left is the bonds." The odds he gives: ~50-50 yields hold here, 20-30% he gets bonds higher; "the trade is asymmetric, focus on that." But the unsolvables (fiscal irresponsibility, Iran/Ukraine geopolitics) make any fix short-term. On AI, his "Claudification Overdrive" call: Wall Street will bombard you with bullish AI narratives because the Anthropic IPO has to succeed — OpenAI's Navier-Stokes "millennium problem" claim (~$22mn of tokens for a $1mn prize: "the ROI is negative") is exhibit one, "more millennium problems… a miracle drug" to come in 30 days — and it will get done ("don't underestimate the greed"), the only question being whether it is "the absolute meme top." Paulo dissents on the window: DeepSeek just shipped a model "right behind Astra" at "1.4% of the cost," Congress is leaning on OpenAI, data-center opposition runs "four or five to one," and if the deal slips past October into a Democratic midterm sweep "that deal comes back a lot cheaper" — while a pulled mega-IPO would itself mark "a near-term low" by removing the supply overhang. He also sees 2026 re-running 2023's bond script (Druck and PTJ on the fiscal, "we are going to sell the bonds hard") and the EMification of the West: with yields "you don't have to squint to see," over-funded US pension plans can defease into 5.5% long bonds and TIPS, a slow rotation out of equities that supports bonds and lets the dollar "slow bleed" rather than collapse — "one more way we are actually speaking Portuguese." Both: own "some real assets somewhere at the back of your pocket."

1. Stocks & names mentioned

TickerNameResearchViewWhat's saidSource
TLTiShares 20+ Year Treasury Bond ETFQT · SA · STK · FAPositiveLe Shrub — a small, defined-risk tactical trade in long-bond optionality, put on ~1-2 weeks earlier and flat since: "I'm risking 30 bips total… if it's going to work and I'm going to make 10 times my money, well, I'm going to do it." Held as a contrarian bet ("I can't sell a trade that Kramer hates"; Barron's cover, Lacy Hunt and Dalio all bond-bearish) on Bessent repeating his crude/yen management on the long end ("he's two out of two"): ~50-50 yields hold, 20-30% they fall — "the trade is asymmetric." He owns no bonds otherwise ("I don't even have T-bills").read ↗
AnthropicAnthropic (private)NeutralSplit call on the IPO, not on the company. Le Shrub ("Claudification Overdrive"): the S-1 "has to succeed," so expect weeks of bullish AI narrative, and "it's very obvious to me it's going to get done" — valuation went from "a trillion" to "two trillion" in two months and "no one blinked"; the question is whether it is "the absolute meme top." Paulo: "not obvious to me that this thing is going to get done" — DeepSeek, semis off highs, safety/guardrail noise and a possible Democratic midterm sweep could push it out and bring it back "a lot cheaper"; a missed window would itself mark "a near-term low in the market."read ↗
DeepSeekDeepSeek (private, China)NeutralPaulo — evidence, not a stance: a new DeepSeek release in the last 12 hours that nobody noticed because of the bond sell-off, "literally right behind Astra. And 1.4% of the cost. And it's ahead of Fable" — "these open models are almost there and they cost a fraction." A way for the AI-IPO trade to "get snaked" before the bankers get the deal away.read ↗
OpenAIOpenAI (private)NegativeLe Shrub: exhibit one of the pre-IPO narrative push — hours after his piece, OpenAI claimed to have solved Navier-Stokes, a millennium problem, then it emerged it had used human researchers' methodology, "burned, like, 22 million in tokens for a $1 million math prize… the ROI is negative" ("we spent $2 trillion to steal someone's math homework"). Paulo adds congressional letters accusing it of misleading, and data-center opposition now "four or five to one."read ↗

2. Talking points

0:31A bull market in memes — $5,000 checks into a 5% long end

3:15The midterm risk window delivered chop — and bonds are the trade everyone mocks

5:09Bessent should manipulate quietly — why Shrub is still in TLT

7:052026 is 2023's bond script — "22 Jump Street"

9:37Claudification Overdrive — the Anthropic IPO has to succeed

12:12A WeWork S-1 moment? Not when Claude summarizes the S-1

14:38How the deal gets snaked — DeepSeek, Congress, NIMBY, the midterms

16:49Never underestimate the fee drive — "the only company that's going to get nationalized is you"

22:05The trade: bond optionality, 30bps of risk, sized to be wrong

26:04EMification — Brazil's 7% real yields and the defeased pension fund

31:08How does Bessent stick-save the long end before the midterms?

35:23The 9/11 tell

3. In plain English

TLT — iShares 20+ Year Treasury Bond ETF Positive

TLT is a fund that holds long-dated US government bonds, so it rises when long-term interest rates fall and drops when they rise. With the 30-year yield near 5.3% and nearly every well-known investor saying bonds are a disaster, Le Shrub has placed a small bet the other way.

The bet is built with options, so the most he can lose is what he paid — about 0.3% of his portfolio — while a big bond rally could return around ten times that. His reasoning is that the Treasury Secretary has already “managed” oil prices and the yen through a crisis, and may do the same to long-term rates with tools like buying back old bonds, issuing fewer long ones and letting banks hold more. He puts the odds of real success at only 20-30%, but the payoff is lopsided, and when everyone from TV hosts to magazine covers hates a trade, he is reluctant to be on the same side. He is clear it is a trade, not a belief in bonds: deficits and war-driven inflation mean he would not own them long term.

Anthropic — Anthropic (private) Neutral

Anthropic, the maker of the Claude AI models, is preparing a stock-market listing at a reported value of around $2 trillion. The two hosts disagree about whether it happens on schedule, not about the company itself.

Le Shrub thinks banks and fund managers earn so much from the deal that they will flood investors with upbeat AI news until it is done — and that it will be done, possibly marking the peak of the AI mania. Paulo is less sure: a cheap new Chinese model, political pressure on AI companies, local opposition to data centers and a possible change of control in Congress in November could delay it and force a lower price. His twist is that a delay could actually help the wider market for a while, because a very large new share sale soaks up money that would otherwise go into existing stocks.


Key points extracted from the paid PauloMacro Substack audio conversation (Substack's transcript saved in transcript.txt) for personal study. Not investment advice. © PauloMacro / Le Shrub for source material.