Yen Carry Trade, Recession odds grow, Buffett cash pile, Google ruled monopoly, Kamala picks Walz

Chamath’s claim was that, over roughly the month after early August 2024, equity‑trading algorithms would mechanically have to sell about $160B of global equities, which would then induce “many hundreds of billions” more in selling by others, so that “the preponderance of the market action will be to continue to sell.”

What actually happened in that period looks inconsistent with that scenario:

1. Global equity prices rose meaningfully over the forecast window. The MSCI ACWI index gained about 2.6% in August 2024 and another 2.7% in September 2024, indicating broad-based equity appreciation rather than sustained net selling pressure. (statmuse.com) The S&P 500 was similarly up ~2% in August and ~2.5% in September 2024, again suggesting buyers, not persistent forced sellers, dominated. (statmuse.com) It is hard to reconcile a month-long, multi‑hundred‑billion mechanical equity dump with globally rising equity indices and no corresponding prolonged drawdown.

2. Fund‑flow data show net inflows into equities over much of that month, not large net outflows. In the week around the initial yen‑carry shock (reported August 9), global equity funds did see outflows of about $6.3B, with nearly $99B rushing into money‑market funds amid the volatility spike. (reuters.com) But a Reuters/LSEG piece on September 6 noted that the week to September 4 was the first weekly global‑equity outflow in four weeks, implying the prior three weeks (i.e., most of August after the shock) had net inflows into equity funds. (y94.com) Separately, the Financial Times reported that ETF flows hit a record in August 2024 with about $129.7B of net inflows, and equity ETFs saw strong buying, which is the opposite of a market dominated by mechanical equity selling. (ft.com) Together, these flow data contradict the idea that the “preponderance” of activity in that month was forced equity liquidation on the scale Chamath described.

3. Market behavior quickly flipped from a forced‑selling shock to a strong rebound. A yen‑driven meltdown and CTA deleveraging did occur: Goldman Sachs’ Scott Rubner estimated that systematic strategies had already dumped roughly $109B of global equity futures in the month up to mid‑August, and expected some continued selling into autumn. (investing.com) However, within that same mid‑August period, global stocks logged their strongest weekly performance since November 2023, with the S&P 500 up about 3.9% for the week, the MSCI World having its best week since early November, and volatility (VIX) dropping back down. (ft.com) That pattern—rapid rebound, easing volatility, and strong weekly gains—is inconsistent with a month‑long dominance of algorithm‑driven selling.

We do see some systematic selling and rebalancing (CTAs’ $109B unwind, later pension and 401(k) re‑allocations out of equities), but public data do not support Chamath’s much larger, specific claim that algorithms had to sell another $160B in equities over the following month and thereby pull through “many hundreds of billions” more in net selling, nor his conclusion that selling would dominate market action over that entire period. Instead, prices and flows suggest that after an initial shock, net buying and recovery characterized most of that month.

Given the available evidence, the prediction that equity markets would experience continued, large, mechanically driven net selling on the order he specified over the subsequent month did not materialize, so this forecast is best classified as wrong.

At the time of the prediction in early August 2024, USD/JPY was around 146–147 (for example, 1 USD = 146.67 JPY on 9 August 2024).(exchange-rates.org) Over the following weeks the yen actually strengthened for a time, with the rate falling toward 141–143 and reaching a 2024 low of about 140.66 per dollar in mid‑September.(exchange-rates.org)

However, over the subsequent months and into 2025 the medium‑term direction matched the prediction: USD/JPY moved well above the 147–150 area, trading above 151 by late October 2024 and around 153–155 in November 2024, and ending 2024 near 156–157.(exchange-rates.org) In 2025 the pair reached highs around 158–159 (for example, 158.35–158.87 in January) and has mostly traded above 150 in late 2025, with recent levels around 155–157 yen per dollar and a six‑month gain of roughly 9–10% for the dollar versus the yen.(exchange-rates.org)

Japan did begin raising policy rates in 2024–2025, bringing the Bank of Japan's short‑term rate to about 0.5% by January 2025, but this still left Japanese rates far below U.S. levels and the yen remained structurally weak.(cnbc.com) Overall, from August 2024 to late 2025 the U.S. dollar does indeed buy meaningfully more yen than at the time of the podcast, so the prediction about further yen depreciation against the dollar has been borne out in broad terms despite interim fluctuations.

Key parts of Chamath’s prediction did not materialize, even though one sub-component (rate cuts) did:

1. Recession call and data becoming “more obvious” in Q3–Q4 2024

   • Official U.S. real GDP growth was positive in every quarter of 2024. BEA data show real GDP increasing 1.6% in Q1 2024 (advance), 2.8% in Q2 (advance, later revised around 3.0%), about 3.0–3.1% in Q3, and 2.4% in Q4; full‑year 2024 growth was 2.8%. (bea.gov)
   • The National Bureau of Economic Research (NBER), the standard arbiter of U.S. recessions, has not dated any new recession after the April 2020 trough; its business‑cycle chronology and derivative indicators show expansion (recession indicator = 0) through at least Q3 2025. (nber.org)
   • Because GDP and other broad indicators remained solidly positive and no NBER recession was declared, it did not become “more obvious” in Q3–Q4 2024, via official data or revisions, that the U.S. had already been in recession as of mid‑2024. This central claim of the prediction is therefore contradicted by realized data and the main official arbiter.

2. Powell “will have to cut” rates in Q3–Q4 2024, question being the size

   • The Federal Reserve did cut rates in that window: at its September 18, 2024 meeting, the FOMC lowered the federal funds rate by 50 basis points (from 5.25–5.50% to 4.75–5.00%), the first cut since the early COVID period. (cnbc.com)
   • Chamath framed the open question as whether Powell would “overreact” and cut 75–100 bps vs 25 bps. In reality, the Fed chose an intermediate path—a 50 bps move with projections implying roughly 100 bps of total easing by the end of 2024—so the actual outcome did not match either extreme he posed. (cnbc.com)

Overall, the core thesis—that the U.S. was already in recession and that this would be clearly borne out by official data and revisions in Q3–Q4 2024—was not borne out. Although he was directionally right that the Fed would start cutting rates in that period, the main, testable part of the prediction (the existence and recognition of a mid‑2024 recession) was wrong. Given that sufficient time and data have now accumulated, this prediction is best classified as wrong.

BEA data show that real U.S. GDP growth was positive in the quarters around the podcast date: it increased at an annualized rate of about 3.0% in Q2 2024, 3.1% in Q3 2024, and 2.4% in Q4 2024, followed by a contraction of –0.6% in Q1 2025 and then a rebound of 3.8% growth in Q2 2025. (apps.bea.gov) Thus, there was only one quarter of negative real GDP (Q1 2025) before growth turned positive again; at no point by roughly August 2025 did the U.S. experience two consecutive quarters of negative real GDP growth, the “technical recession” criterion specified in the normalized prediction. Additionally, business‑cycle dating series and NBER communications indicate no new recession episodes after the brief COVID-19 recession in 2020, reinforcing that no recession—technical or official—was recorded in that 2024–2025 window. (nber.org) Therefore, Chamath’s expectation that the U.S. would “probably be in a technical recession” about a year after the August 2024 discussion did not come to pass.

Chamath’s claim was that by around August 2025, there was a “pretty decent chance the market will be up” relative to early August 2024 levels.

Using the podcast date (August 9, 2024) as the reference point:

• On August 9, 2024, the S&P 500 closed at 5,344.16, and the Nasdaq Composite closed at 16,745.30.【2search3】2search7】2search2】
• By late August 2025, both indices were clearly higher. On August 29, 2025, the S&P 500 closed at 6,460.26 (about 21% above its Aug 9, 2024 level), and the Nasdaq Composite closed at 21,455.55 (roughly 28% higher than Aug 9, 2024).【1search7】1search0】1search5】
• News reports also note that by September 9, 2025, the S&P 500 (6,512.61) and Nasdaq (21,879.49) were at or near record highs, confirming that major U.S. equity indices were significantly above their early‑August‑2024 levels.【2news13】

Whether or not a “technical recession” occurred is not necessary to evaluate the directional market call. The key observable claim—that major U.S. equity indices would likely be higher by around August 2025 than they were in early August 2024—did in fact materialize. Therefore, this prediction is best classified as right.

Friedberg’s normalized prediction was that there was a greater‑than‑not (>50%) probability the U.S. would enter a recession within about one year of the August 9, 2024 episode (by ~August 2025).

By that date, the commonly used recession criteria were not met:

1. Two consecutive quarters of negative real GDP did not occur. Real U.S. GDP grew at annualized rates of about 3.1% in Q3 2024 and 2.4% in Q4 2024, then fell 0.5% in Q1 2025 and rebounded 3.3% in Q2 2025. That is, only one quarter (Q1 2025) was negative, and it was followed by strong positive growth, so the “two negative quarters” rule-of-thumb for a recession was not satisfied.

   • Q3 2024: +3.1% real GDP growth.
   • Q4 2024: +2.4%.
   • Q1 2025: –0.5% (contraction), largely attributed to a surge of imports ahead of new tariffs.
   • Q2 2025: +3.3% rebound.

   (apnews.com)

2. No officially dated NBER recession in that window. The National Bureau of Economic Research’s Business Cycle Dating pages still report the most recent peak in U.S. economic activity as February 2020 and the trough as April 2020, with no later peaks or troughs announced as of late 2025. (nber.org) Since NBER is the standard arbiter of U.S. recessions and has not declared a new recession after the short 2020 downturn, there is no recognized recession starting between August 2024 and August 2025.

3. Broader data and commentary describe a slowdown, not a recession in that period. Full‑year 2024 U.S. real GDP growth is estimated around 2.8%, with projections around 2% for 2025, indicating continued (if slower) expansion rather than an outright contraction. (en.wikipedia.org) While some analysts and even an NBER research paper discussed elevated probabilities (e.g., a model‑based 71% chance the U.S. was in recession as of May 2025), these were probabilistic assessments and not confirmations that a recession had actually begun. (linkedin.com)

Given that (a) the two‑negative‑quarter GDP criterion was not met within the forecast window, and (b) the NBER has not subsequently dated any recession in that period, the U.S. did not enter a generally recognized recession between August 9, 2024 and August 9, 2025. Consequently, a forecast that the probability of such a recession was greater than 50% turned out to be wrong in outcome (the event did not occur in the stated timeframe), even if many contemporaneous forecasters viewed recession risk as elevated.

The prediction was conditional: if the U.S. entered a recession sometime between roughly August 2024 and August 2025, then the federal government would roll out significant countercyclical programs that kept equities rallying or at least “strong” despite that recession.

On the first part – whether a recession actually occurred in that window – the evidence is mixed and not cleanly resolved:

• The NBER-based recession indicator used by the St. Louis Fed remains at 0 (expansion) through at least October 2025, meaning no official post‑2020 recession has been dated yet. (fred.stlouisfed.org)
• Major forecasters (IMF, RBC, UBS, various banks) describe a sharp slowdown / soggy or near‑zero growth and elevated recession risk in 2025, but generally do not say a clear, broad U.S. recession is underway; some emphasize the distinction between a “growth recession” and a formally defined recession. (livemint.com)
• Academic work using unemployment/vacancy classifiers estimates around a 71% probability the U.S. was already in recession as of May 2025, but that is probabilistic model evidence, not a widely accepted determination. (arxiv.org)

On the second part – government support and equity performance – the picture is also not cleanly attributable to what Friedberg described:

• The U.S. did suffer a sharp stock market crash in early April 2025 tied to Trump’s broad “Liberation Day” tariff shock, but by late August 2025 the S&P 500 had rebounded and was up about 9–10% year‑to‑date, implying equities were again relatively strong. (en.wikipedia.org)
• Federal actions in this period were dominated by aggressive tariff policy, talk of future tax cuts, and assurances from Treasury and the Fed that the system was sound and that tools were available if needed; there is little evidence of large, new, clearly countercyclical market‑support programs of the kind seen in 2008 or 2020 rolled out specifically in response to an acknowledged recession. (ft.com)

Because (1) it is not clearly established that the U.S. was in a formal, broad‑based recession during the specified period, and (2) any subsequent equity strength cannot be unambiguously tied to distinct, recession‑response government programs, the prediction cannot be judged definitively true or false on the available record. Hence the outcome is best classified as ambiguous rather than right, wrong, or merely “too early.”

By roughly one year after the discussion (around August 2025), the U.S. had not experienced two consecutive quarters of negative real GDP growth, nor had the NBER declared a new recession.

Evidence:

• BEA data show real GDP was positive throughout 2024, with Q4 2024 growing at about 2.3–2.4% annualized. (apps.bea.gov)
• Real GDP then declined in Q1 2025 (around −0.5 to −0.6% annualized after revisions), but rebounded in Q2 2025, growing about 3.3–3.8% annualized, so there was only one negative quarter followed by strong growth, not two negative quarters in a row. (bea.gov)
• The NBER’s Business Cycle Dating Committee listings and later summaries show no new recession peak/trough dates after the April 2020 trough, meaning no officially dated recession in 2024–2025 as of late 2025. (nber.org)

Since the realized outcome by about August 2025 was that the U.S. had not entered a recession under either the common two‑negative‑quarters rule of thumb or the official NBER dating, the prediction that a recession was “more likely than not” by that time ended up being wrong in hindsight.

As of November 30, 2025, the key trigger condition for Jason’s prediction has not occurred, and the time window he specified has not started.

1. Apple has not (yet) lost the Google default‑search TAC deal as a remedy of the DOJ antitrust case.

   • Judge Amit Mehta’s remedies ruling in the U.S. search‑monopoly case against Google (United States v. Google LLC, 2020) in September 2025 barred exclusive arrangements and required Google to share some search data, but explicitly allowed Google to continue paying partners like Apple to be the default search engine. (barrons.com)
   • Reporting around the case and Apple’s own filings repeatedly describe Apple as still benefiting from a multibillion‑dollar default‑search agreement with Google (about $20B/year), and Apple has been actively trying to protect that revenue stream in the remedies phase. (theverge.com)
     Since the deal continues, the condition “after losing the default‑search TAC deal with Google as a result of this antitrust case” has not been met.

2. Apple has not bought an existing search engine or launched a full Google‑style first‑party web search engine.

   • There is no record through late 2025 of Apple acquiring DuckDuckGo, Brave Search, or any similar web search engine. Public reporting and reference sources continue to list DuckDuckGo and Brave Search as independent companies/products, not Apple subsidiaries. (en.wikipedia.org)
   • In fact, Apple’s services chief Eddy Cue has told the court that Apple has no plans to build a standalone Google‑rival search engine, calling such an effort economically and technically impractical; Apple instead emphasizes AI‑driven search features layered on top of existing providers. (theverge.com)
     This undercuts the likelihood of the prediction but doesn’t make it logically false, because it concerns Apple’s future behavior after a potential loss of the Google deal.

3. Time horizon: “within a few years after losing the deal.”
   Jason’s prediction is explicitly conditional and long‑dated: only after Apple loses the TAC deal as a result of this case does a “within a few years” clock start. Given that (a) Apple has not yet lost the deal, and (b) the contract itself is reported to run at least into 2026 with possible extensions, the specified time window for the predicted acquisition has not even begun. (reuters.com)

Because the triggering event (loss of the TAC deal) has not occurred and the stated “few years” window has not started, there is not yet a factual basis to declare the prediction either correct or incorrect. It remains too early to tell, so the outcome is inconclusive.

Public information on the remedies decision in United States v. Google LLC (the DOJ’s search antitrust case) shows that Judge Amit Mehta rejected structural breakup remedies (such as forced divestiture of Chrome or Android) and instead imposed behavioral / conduct remedies with a multi‑year duration.

• A September 10, 2025 Congressional Research Service legal sidebar explains that in the September 2, 2025 remedies decision, the court “rejected the plaintiffs’ proposals for structural relief” (including divestiture of Chrome and contingent divestiture of Android) and instead ordered a set of behavioral remedies (data sharing with rivals, search and ad syndication, limits on exclusive contracts, disclosure of ad‑auction changes, and a technical committee), with the final judgment to run six years. (congress.gov)
• The DOJ’s own September 2, 2025 press release describes the outcome as the court prohibiting Google from entering or maintaining certain exclusive contracts, requiring it to share index and user‑interaction data and to offer search and text‑ad syndication services—all classic conduct remedies—without any order to divest business units. (justice.gov)
• Commentary from outlets like The Guardian characterizes the decision as avoiding breakup (no divestiture of Chrome or Android) and instead imposing relatively limited penalties, criticized by some as a “slap on the wrist,” again underscoring that no Ma Bell–style structural split was ordered. (theguardian.com)

As of November 30, 2025, the final judgment at the trial‑court level is a set of behavioral constraints lasting several years, with no structural breakup ordered, which matches Chamath’s prediction that the “big O” outcome (a Ma Bell–style structural breakup) would be off the table and that the remedy would look more like a consent‑decree‑style conduct regime. Appeals may continue, but the remedies decision we have validates the prediction in substance.

The remedies in the DOJ’s search-monopoly case against Google were decided on September 2, 2025. Judge Amit Mehta ordered behavioral remedies (banning certain exclusive contracts, requiring Google to share some index and interaction data, etc.) while rejecting the DOJ’s proposed structural break‑up of Chrome/Android. The DOJ, under the Trump administration, publicly framed this as a “significant win,” indicating support from the Republican-led executive branch and allied state attorneys general, but that only covers one side of the aisle. (justice.gov)

Congressional and broader political reaction was explicitly mixed, not broadly bipartisan in favor. A Congressional Research Service summary notes that reactions to the remedies decision were mixed and that some Members of Congress criticized the ruling as providing insufficient relief and began talking about new legislation instead—signaling clear dissent rather than unified support. (congress.gov) Prominent Democrats in particular opposed the outcome: Senator Amy Klobuchar (D‑MN) said the limited remedies show why additional rules for Big Tech are needed and used the decision to argue for passing stronger legislation, rather than endorsing the court’s remedy. (theverge.com) Senator Elizabeth Warren (D‑MA) went further, calling the remedies a “slap on the wrist” that fail to hold Google accountable and explicitly arguing the judge should have broken up the company. (warren.senate.gov) News coverage likewise reported that lawmakers were among those describing the penalties as a “slap on the wrist” and expressing deep disappointment at the weak sanctions. (theguardian.com)

Because key Democratic lawmakers publicly criticized the remedies as inadequate rather than “really supporting whatever happens here,” while support came mainly from the Republican DOJ and some industry groups, the outcome was not one of broad, bipartisan public backing for the court’s action. That contradicts Chamath’s prediction that both the Democrats and the Republicans will really support whatever happens here, so the prediction is wrong.

Evidence shows that Tim Walz did remain Kamala Harris’s running mate through the 2024 Democratic National Convention and beyond, contrary to the prediction that he would have to drop out before or during the convention.

Key facts:

• The 2024 Democratic National Convention in Chicago was held August 19–22, 2024, with Kamala Harris as the presidential nominee and Tim Walz as the vice‑presidential nominee listed on the official convention record. (en.wikipedia.org)
• On August 6, 2024, the Democratic Party formally certified Kamala Harris and Tim Walz as the official Democratic nominees for president and vice president, respectively, after the delegate roll call. (democrats.org)
• During the convention itself, Walz formally accepted the Democratic Party’s vice‑presidential nomination in a televised speech, explicitly thanking delegates for nominating him as vice president. (dw.com)
• Post‑election coverage and later analyses consistently refer to Walz as Harris’s 2024 vice‑presidential running mate and the Democratic vice‑presidential nominee, with no record of him withdrawing or being replaced on the ticket. (en.wikipedia.org)

Because Walz remained on the ticket through the Democratic National Convention and the November 2024 election, the prediction that he "might have to drop out" before or at the convention was not borne out.

Tim Walz was selected as Kamala Harris’s running mate on August 6, 2024 and remained on the Democratic ticket through the November 2024 election, so the condition of the prediction was met.

However, the predicted outcome — recurring, publicly reported protests by Gold Star families at Harris–Walz campaign events, effectively dogging “all of their events” and creating an ongoing media headache through Election Day — did not occur.

What did happen is a concentrated controversy around Donald Trump’s August 26, 2024 visit to Arlington National Cemetery with Gold Star families of the 13 service members killed at Abbey Gate. Harris criticized Trump’s team for filming there as a political stunt, and several of those Gold Star families responded with videos, a joint statement, and later a Trump campaign ad sharply attacking Harris and the administration’s handling of the Afghanistan withdrawal. This generated a burst of negative coverage for Harris across outlets from the AP and Boston Globe to numerous conservative sites, but the conflict was framed around her comments on Trump’s Arlington visit and the 2021 withdrawal, not around protests at her rallies. (en.wikipedia.org)

By contrast, contemporaneous reporting on Harris–Walz campaign events shows repeated disruptions by pro‑Palestinian / Gaza protesters and “uncommitted” activists, including at major rallies in Detroit and Phoenix and at a New York City support event, as well as organized Gaza‑focused protests around the Democratic National Convention and in Minneapolis. These stories consistently describe Gaza-related activists — not Gold Star families — as the source of recurring protest pressure on the Harris–Walz campaign. (wdet.org)

Extensive news searches turn up Gold Star families’ videos, letters, and media statements criticizing Harris, but no evidence of them repeatedly protesting at Harris–Walz campaign stops in person, and no narrative that Gold Star family protests became a constant presence “at all of their events.” Given that such a pattern would have been newsworthy and is absent from major coverage, the specific, event-focused and recurring-protest scenario described in the prediction did not materialize. Therefore, the prediction is best classified as wrong.