E91: SoftBank's $21B+ Vision Fund loss, signals of a bubble, macro picture, Trump raided by FBI

Breaking the prediction into components:

1. "Shallow technical recession" by Q2 2022 – This matches the data. BEA shows real U.S. GDP fell at an annualized –1.6% in Q1 2022 and –0.6% in Q2 2022, i.e., two consecutive negative quarters (a common “technical recession” rule of thumb). (bea.gov)

2. Return to positive real GDP in Q3 or Q4 2022 – This also happened. Real GDP turned positive in Q3 2022 (+2.6% annual rate, advance estimate) and remained positive in Q4 2022 (+2.7% in the second estimate). (bea.gov)

3. "Double-dip" recession in 2023 as rate hikes kick in (another downturn 6–9 months later) – This did not occur. BEA data show real GDP grew in every quarter of 2023: Q1 (~1.1% advance), Q2 (2.1% second/third estimate), Q3 (4.9% third estimate), and Q4 (about 3.2–3.4%), with full‑year 2023 real GDP up 2.5% from 2022. (bea.gov) The National Bureau of Economic Research’s official business‑cycle chronology (last updated March 14, 2023) lists the most recent recession as February–April 2020 and does not show any new recession starting in 2023. (nber.org)

Because the central, forward‑looking part of the prediction was a renewed recession in 2023 that never happened, the overall prediction is best classified as wrong, even though the short‑term call about a 2022 technical recession followed by a brief return to growth was accurate.

Evidence shows that while Fed rate hikes in 2022 did sharply hit parts of U.S. housing/construction, the broader U.S. economy did not exhibit clear, economy‑wide weakness by early‑to‑mid 2023.

• On the construction side, single‑family housing starts fell 32% from March–December 2022, and single‑family units under construction began their fastest sustained decline since 2006–09 starting in mid‑spring 2022, consistent with a strong and early impact from Fed tightening on that segment of construction. (fredblog.stlouisfed.org)
• However, macro indicators in early–mid 2023 remained very strong:
  • The unemployment rate was just 3.4% in April 2023—matching the lowest level since 1969—and stayed in the mid‑3% range through mid‑year, indicating a very tight labor market rather than broad weakness. (bls.gov)
  • Real GDP grew at a 2.2% annualized rate in Q1 2023, and for all of 2023 real GDP expanded 2.5%, faster than in 2022, which is inconsistent with a recession or clear economy‑wide downturn. (bea.gov)
  • Retrospective assessments note that the widely predicted 2023 recession never arrived; instead, the U.S. economy “defied expectations” with solid growth and a soft‑landing narrative. (investopedia.com)
• Even within construction, employment was not collapsing in 2022–23 in a way that would signal a broad downturn: BLS reports construction employment reached an all‑time high of about 8.0 million jobs in 2023, surpassing the 2006 housing‑boom peak, and AGC data show 45 states added construction jobs between June 2022 and June 2023. (bls.gov) Other analysts explicitly described this cycle as unusual because homebuilding and construction employment proved more resilient than past tightening episodes, helping explain why GDP and job growth “kept on chugging” despite sharply higher rates. (axios.com)

Because the predicted broad economic weakness by early‑to‑mid 2023 did not materialize—the economy remained in expansion with very low unemployment and positive GDP growth—the specific lagged‑propagation claim from construction to a clearly weak overall economy is best judged as wrong, even though the first step (construction being hit early) was partly correct.

Sacks’ prediction has largely borne out.

1. Republicans rallied around Trump after the Mar‑a‑Lago raid. Within days of the 8 August 2022 FBI search, Republican leaders attacked the FBI and DOJ and publicly defended Trump, framing him as a victim of a politicized investigation. (en.wikipedia.org) Polling showed GOP voters consolidating behind him: a Politico/Morning Consult poll found 58% of Republican primary voters preferred Trump as the 2024 nominee—his highest level of post‑presidency support—and 71% said he should run; an Economist/YouGov poll saw his favorability among Republicans jump from 45% to 57% in a week. (newsweek.com) This is consistent with Sacks’ claim that the raid spurred Republicans to rally around Trump.

2. He became “very hard to beat” and won the 2024 GOP nomination. Through late 2023 and early 2024, Trump held dominant leads in Republican primary polling in key states such as New Hampshire, South Carolina, and Georgia, often exceeding 55–60% support versus much lower numbers for rivals like Nikki Haley and Ron DeSantis. (en.wikipedia.org) On March 12, 2024 he officially clinched the Republican nomination by surpassing the delegate threshold. (forbes.com) This matches Sacks’ forecast that, after the raid, Trump would be very difficult to defeat for the 2024 Republican nomination.

3. Of the polarized outcomes Sacks named, the ‘White House’ branch occurred. In November 2024, Trump won the U.S. presidential election against Kamala Harris, securing 312 electoral votes to 226 and winning the popular vote 49.8% to 48.3%. (en.wikipedia.org) He took office again in January 2025. This is exactly the “White House” end of Sacks’ “big house or White House” framing.

4. The ‘big house’ (incarceration) outcome has not occurred, and legal actions did not derail his candidacy. Trump was convicted in New York state court in May 2024 on 34 felony counts related to hush‑money payments, but in January 2025 he received an “unconditional discharge,” meaning no jail time, fines, or probation. (reuters.com) Other criminal cases—federal election‑obstruction and classified‑documents prosecutions—were dismissed or dropped after his 2024 victory, in part because DOJ policy bars prosecuting a sitting president and because a judge found the special counsel’s appointment invalid. (en.wikipedia.org) None of these proceedings prevented him from securing the nomination or winning the presidency, so the caveat Sacks added (“unless the FBI comes up with ironclad evidence…that derails his candidacy”) did not materialize in the sense of blocking his political comeback.

Taken together, the central elements of Sacks’ prediction—that the Mar‑a‑Lago raid would rally Republican support around Trump, make him very hard to beat for the 2024 GOP nomination, and lead to a polarized outcome where he either returns to the White House or faces serious legal ruin—have come true in the form of a strong rally, an easy nomination win, and an eventual return to the presidency rather than incarceration. Therefore, the prediction is best classified as right.

Portion about 10–15% believing evidence was planted
Multiple polls after the Mar-a-Lago search show that at least this share of Americans believed or leaned toward believing the “planted evidence” claim:

• A YouGov poll summarized by The Washington Post found that shortly after the search, 38% of Republicans said the FBI planted evidence at Mar‑a‑Lago, 23% disagreed, and 39% were unsure. With Republicans making up roughly a third of adults, that alone implies ≈11–13% of all U.S. adults affirmatively believed evidence was planted. A subsequent YouGov poll the next month found a majority of Republicans said it was at least “probably true” the FBI planted classified documents, which would put the national share at or above 15%. (washingtonpost.com)
• A national survey summarized by Statista (Sept 3–6, 2022) reported 9% of all Americans saying it was definitely true the FBI planted sensitive documents at Mar‑a‑Lago. That figure excludes additional respondents who said it was “probably true,” so the total share who believed the claim is necessarily higher than 9%, plausibly in the low‑ to mid‑teens. (statista.com)

Taken together, these data support Chamath’s prediction that at least 10–15% of the U.S. population would believe the documents were planted.

Portion about an outcome that leaves people unsatisfied
The legal and political trajectory of the case fits his contention that whatever the outcome, it would not be broadly seen as satisfactory:

• Early on, public opinion was sharply split: a Northeastern University survey conducted days after the search found 51% of Americans approved of the raid and 27% disapproved, but 84% of Democrats approved while 64% of Republicans opposed it—showing immediate, intense partisan division. (news.northeastern.edu)
• A Quinnipiac poll in late August 2022 found about half of Americans thought Trump’s conduct was criminal and that he should be prosecuted, with 85% of Democrats and 52% of independents saying he should be charged—versus strong opposition among Republicans. (thepeninsulaqatar.com)
• By June 2023, an AP–NORC poll found 53% of adults believed Trump had done something illegal in his handling of the documents, but only 23% of Republicans agreed; views on his guilt and the propriety of prosecution remained deeply polarized. (apnorc.org)
• Legally, the case ended inconclusively for accountability advocates: Trump was federally indicted over the documents in 2023, but Judge Aileen Cannon dismissed the case on Appointments Clause grounds in July 2024. (en.wikipedia.org) Democrats and many legal experts condemned the ruling as “breathtakingly misguided” and “stunning and wrong,” while Republicans celebrated it as vindication and a major legal victory. (missouriindependent.com)

So although some Republicans were clearly pleased by the dismissal, the broader pattern matches Chamath’s point: no resolution commanded widespread, cross‑partisan satisfaction. One side’s “win” was treated by the other as proof of system failure.

Portion about both extremes getting angrier and more polarized
Evidence also supports his forecast of intensified anger and polarization on both ends of the spectrum:

• Surveys show a widening partisan gulf in trust toward the FBI and DOJ. A Marquette Law School national poll found Republicans far less confident in the FBI than Democrats; large shares of Republicans report “very little” or “no” confidence in the bureau, while Democrats mostly express high confidence. (law.marquette.edu)
• After Trump’s subsequent indictment on the classified-documents charges, an ABC News/Ipsos and CBS/YouGov polling summary found around two‑thirds to three‑quarters of Republicans believed the charges were politically motivated and said they must support Trump to oppose his enemies—while majorities of Americans overall (and overwhelming majorities of Democrats) thought his retention of the documents was a serious national‑security problem and that he should be charged. (axios.com)
• Broader opinion data around this period show majorities saying the country is on the “wrong track,” ranking “threats to democracy” as a top national problem and expressing high willingness to protest—indicators of elevated, cross‑ideological anger. (statista.com)

Assessment

• The quantitative threshold in the prediction—at least 10–15% believing the evidence was planted—is met by multiple independent polls. (washingtonpost.com)
• The qualitative claims that the investigation’s outcomes would not bring a broadly legitimizing resolution, and that both political extremes would grow angrier and more polarized, are borne out by the sequence of a polarizing raid, a controversial indictment, and a highly disputed dismissal that each side alternately denounced or celebrated.

While some wording (e.g., “nobody will be satisfied”) is somewhat hyperbolic, the core substance of Chamath’s forecast—significant belief in planted evidence, no outcome widely accepted as fair, and heightened polarization on both ends—matches what happened closely enough to count as essentially correct.

Chamath framed the “middle path” as one where nothing happens legally to Trump from the Mar‑a‑Lago raid, but public trust in DOJ continues to erode.

On the legal outcome, that forecast did not hold:

• The August 8, 2022 FBI search of Mar‑a‑Lago recovered over 13,000 government documents, including 337 marked classified. (en.wikipedia.org)
• On June 8, 2023, Trump was federally indicted in United States v. Trump on 37, later 40, felony counts (willful retention of national defense information, obstruction‑related offenses, false statements, etc.), a case explicitly arising from his retention of classified documents at Mar‑a‑Lago and evidence gathered through that investigation and search. (en.wikipedia.org)
• Trump surrendered, was arrested, booked, and arraigned in June 2023, with charges carrying potential multi‑year prison sentences—an unprecedented federal criminal case against a former president. (en.wikipedia.org)
• The case was later dismissed in July 2024 on the ground that Special Counsel Jack Smith’s appointment was unconstitutional; DOJ dropped its appeal after Trump’s 2024 re‑election, and in February 2025 the FBI returned boxes seized in the 2022 search. (en.wikipedia.org)

Even though the charges were ultimately dismissed and no conviction or sentence resulted, being indicted on dozens of federal felonies, arrested, arraigned, and subjected to a multi‑year prosecution is a major legal consequence flowing directly from the Mar‑a‑Lago documents investigation and raid. That is materially different from “nothing happens” or “no significant legal consequences.”

On the trust in DOJ piece, the direction of his call is broadly consistent with the data:

• Surveys since 2022 show low and worsening confidence in the justice system generally; Gallup found overall confidence in the U.S. judicial system at a record‑low 35% in 2024, with Trump‑related legal battles cited among the contributing factors. (apnews.com)
• Evaluations of DOJ’s performance have been largely negative: a 2024 summary of Gallup and Pew data reports roughly seven in ten Americans rating DOJ’s performance as only fair or poor, and Pew finds the public split roughly evenly between favorable and unfavorable views of DOJ. (forbes.com)
• By 2025, only about 20% of Americans express strong confidence that DOJ and the FBI will act fairly and nonpartisan under Trump, and a Marquette national poll finds 55% saying DOJ has filed unjustified cases against Trump’s political opponents—clear evidence of widespread skepticism about DOJ’s fairness. (apnews.com)

So, while Chamath was directionally right that trust in DOJ would be further eroded, the central, falsifiable part of his prediction—that the Mar‑a‑Lago investigation would lead to effectively no significant legal consequences for Trump—has been contradicted by the subsequent indictment and prosecution. Overall, this makes the prediction wrong.