E110: 2023 Bestie Predictions!

The prediction was that in the Students for Fair Admissions (SFFA) cases against Harvard and UNC, the Supreme Court in 2023 would strike down the existing race-based affirmative action admissions policies at those schools as unconstitutional, in a way framed as a win for Asian American applicants.

What actually happened:

• On June 29, 2023, the U.S. Supreme Court decided Students for Fair Admissions, Inc. v. President and Fellows of Harvard College and the companion UNC case.
• The Court’s majority held that the race-conscious admissions programs at Harvard and UNC violated the Equal Protection Clause (and, for Harvard, Title VI) and therefore were unconstitutional. (supreme.justia.com)
• News and university reports uniformly describe this as the Court striking down race-conscious/affirmative action admissions programs at Harvard and UNC. (theguardian.com)
• The cases were brought by SFFA on behalf of Asian American plaintiffs, arguing that the policies harmed Asian American applicants; the Court’s ruling agreed with SFFA that such race-based systems are impermissible, and the outcome has been widely characterized as a victory for SFFA and the Asian American applicants it represented, even though the Court did not separately find intentional anti–Asian American animus. (thecrimson.com)

The central factual claim—that the Court would strike down those race-based admissions policies as unconstitutional in 2023 in the SFFA v. Harvard/UNC cases—is exactly what occurred. The more rhetorical part (“discriminate against Asian Americans”) matches the plaintiffs’ framing that the majority essentially accepted (that using race in this way is unlawful discrimination), even though the Court sidestepped a specific finding of intentional anti-Asian bias.

Given that, the prediction is best classified as right.

The prediction is right.

In Students for Fair Admissions v. President and Fellows of Harvard College and Students for Fair Admissions v. University of North Carolina, the U.S. Supreme Court issued decisions on June 29, 2023 holding that the race-conscious admissions programs at Harvard and UNC violated the Equal Protection Clause of the Fourteenth Amendment (and, for Harvard, Title VI). The Court’s majority opinion concluded that the universities’ use of race in admissions could not be reconciled with the Constitution, effectively ending the form of race-based affirmative action used by elite schools such as Harvard and UNC.

• The Supreme Court’s syllabus and majority opinion state that these admissions programs “cannot be reconciled with the guarantees of the Equal Protection Clause” and therefore “must be invalidated.”

While the Court did not ban every conceivable consideration of race in any context (for example, applicants may still discuss how race affected their lives in essays), it clearly overturned the existing race-based affirmative action regime in college admissions as practiced by Harvard, UNC, and similar institutions in 2023. That matches Chamath’s prediction that it “is going to get repealed.”

Evidence from the 2023–24 cycle shows that Ron DeSantis’ position deteriorated sharply from early‑2023 expectations while Nikki Haley’s standing improved and she ultimately became the main non‑Trump contender.

• Starting point (early 2023): In February 2023 Monmouth found GOP voters essentially split between DeSantis and Trump (33% each) with Haley at just 1%, making DeSantis a co‑frontrunner and Haley a marginal figure.(monmouth.edu) Media and polling throughout early 2023 consistently framed DeSantis as Trump’s strongest alternative.(cnbc.com)
• Poll trajectory in 2023: By late 2023, national and state polls showed DeSantis sliding while Haley rose into double digits. A Marquette national survey of Republican voters in Nov. 2023 had Trump 54%, with DeSantis and Haley tied at 12%—a rise for Haley and a decline for DeSantis since March.(law.marquette.edu) A Wisconsin Marquette poll from June→Oct./Nov. 2023 showed DeSantis dropping from 30% to 18% while Haley climbed from 3% to 11%.(marquette.edu) A Quinnipiac poll in Dec. 2023 likewise found Haley catching DeSantis nationally, both at 11%, with DeSantis down from far higher levels earlier in the year.(axios.com)
• Primary results and candidate status in 2024: In the Iowa caucuses Trump won with 51%; DeSantis was a distant second at ~21% and Haley close behind at ~19%, far from DeSantis’s early “co‑frontrunner” status.(en.wikipedia.org) DeSantis then ended his campaign on Jan. 21, 2024—before New Hampshire—and endorsed Trump, with coverage explicitly noting he had once been seen as Trump’s strongest challenger.(cnbc.com)
• Haley’s emergence as main alternative: After DeSantis dropped out, Haley became the sole major challenger to Trump. She took 43% to Trump’s 54% in the New Hampshire primary, clearly occupying the runner‑up lane.(en.wikipedia.org) She later won the District of Columbia primary by nearly 30 points over Trump, her first primary victory and the only GOP primary Trump lost in 2016 or 2024,(en.wikipedia.org) and went on to win Vermont as well before suspending her campaign.(nypost.com)

Relative to their positions at the start of 2023, DeSantis badly underperformed initial frontrunner expectations and exited early, while Haley rose from low single digits to become Trump’s principal challenger and secure actual primary wins. That matches Chamath’s effective bet: “long Nikki Haley, short Ron DeSantis.”

As of November 30, 2025, the United States has never elected a female president from either major party. Joe Biden (Democrat) served from January 20, 2021 to January 20, 2025, followed by Donald Trump’s second presidency beginning January 20, 2025, with JD Vance as vice president—both men and both Republicans now in office.

Sources discussing Kamala Harris’s 2024 loss explicitly note that, had she won, she would have been the first female U.S. president, implying that the U.S. still has not had a woman president.

• Biden presidency (2021–2025) and Trump’s second presidency beginning in 2025 confirm that all presidents to date are male.
• Coverage of Harris’s 2024 campaign and its aftermath states that she would have been the first female president if elected, underscoring that no woman has yet held the office. (en.wikipedia.org)

Chamath’s prediction is about the ordering of future events (Republican woman elected president before Democratic woman). Since no woman from either party has yet been elected president, the prediction cannot currently be judged as right or wrong. It remains too early to tell, so the outcome is inconclusive.

Evaluating each part of Jason’s chained prediction:

1. GLP‑1/Ozempic and 40–50 lb weight loss in 2023 – Trump’s official 2025 physical lists his medications (cholesterol drugs, aspirin, topical cream) and shows a weight of 224 lbs vs. 244 lbs in 2020, i.e., about 20 lbs lost over several years, not a 40–50 lb drop in 2023. There is no official confirmation that he used a GLP‑1 drug such as Ozempic; media pieces only speculate based on his appearance and unofficial weight numbers.(people.com) This part is at best unproven and the magnitude/timing of the claimed loss is inconsistent with public data.

2. Criminal indictment by Merrick Garland/DOJ before the 2024 election – Attorney General Merrick Garland appointed Jack Smith as special counsel in November 2022 and authorized him to investigate and prosecute Trump.(en.wikipedia.org) Trump was then federally indicted in June 2023 in the classified-documents case and again in 2023 in the election‑obstruction case, both brought by Smith under the Department of Justice, well before the 2024 election.(en.wikipedia.org) Interpreting “indicted by Garland” as “indicted by DOJ under Garland,” this sub‑prediction is substantively correct.

3. Winning the 2024 Republican nomination despite indictment – Trump clinched the delegate majority on March 12, 2024, becoming the presumptive GOP nominee, and was formally nominated at the Republican National Convention on July 15, 2024.(apnews.com) This sub‑prediction came true.

4. Settlement where Trump agrees not to run in exchange for a broad ‘Nixon-style’ pardon – No such bargain occurred. Instead, Trump did run, secured the Republican nomination, and won the 2024 presidential election.(en.wikipedia.org) His federal cases were later dismissed or dropped for legal and policy reasons (special‑counsel appointment ruling, Supreme Court immunity decision, DOJ policy against prosecuting a sitting president), and the Georgia case was dismissed by a new state prosecutor; none of these outcomes involved any agreement by Trump not to run or any comprehensive pardon from another president.(en.wikipedia.org) There is also no evidence of a broad, Nixon‑style pardon covering all his legal exposure.

Because two crucial elements of the chain (the specific GLP‑1/40–50 lb weight‑loss claim and, especially, the non‑running‑for‑pardon settlement) did not happen, the overall prediction is wrong, even though the DOJ indictment and GOP‑nomination pieces were accurate.

Assessment

The prediction was that in 2023 both:

• California would be looking at roughly a $24B budget deficit, and
• San Francisco would also have a gigantic budget shortfall, with significant fiscal pain in the form of spending cuts or other austerity.

California (state)

• In November 2022, the California Legislative Analyst’s Office (LAO) released its 2023–24 fiscal outlook stating that the state would face a $24 billion “budget problem” (deficit) in 2023–24, explicitly defining this as a deficit where resources are insufficient to fund currently authorized services. This is exactly the order of magnitude and fiscal year Sacks referenced. (lao.ca.gov)
• Governor Newsom’s January 2023 proposed 2023–24 budget acknowledged a projected $22.5B deficit, very close to the LAO’s $24B estimate, and proposed to close it through a mix of solutions emphasizing delays, fund shifts, and some spending cuts, rather than new taxes. (src.senate.ca.gov)
• Subsequent reporting on the evolving shortfalls shows that managing these multi‑year deficits has in fact involved broad spending reductions across many programs (e.g., an AP analysis describing multibillion‑dollar state deficits being addressed with “widespread spending cuts across 260 programs,” including trims in housing, broadband, water infrastructure, and education). (apnews.com) While some of the heaviest cuts were embedded in 2024 budget actions, they stem directly from the deficit trajectory identified in 2023–24.

Taken together, California did experience a major budget shortfall of roughly the predicted magnitude (~$24B) and responded with significant belt‑tightening measures and program reductions, matching the spirit of “a lot of pain.”

San Francisco (city)

• In October 2023, Mayor London Breed warned of a “growing structural deficit” and instructed departments to immediately begin addressing an anticipated at least $500M deficit in FY 2025–26, ordering measures such as limiting contract expansions, pausing yet‑to‑start programs, eliminating some vacant jobs, and other cost‑containment steps. (sfstandard.com)
• By December 2023, the city formally projected an $800M deficit over two years and ordered departments to prepare 10% General Fund spending cuts plus an additional 5% contingency, clearly constituting the “gigantic budget shortfall” and significant austerity Sacks predicted. (sfstandard.com)

Conclusion

• The state did in fact confront a multi‑billion‑dollar deficit of roughly $24B in 2023–24, and the solution set included meaningful spending reductions and other painful budget maneuvers.
• San Francisco also entered 2023 in worsening fiscal condition and, by late 2023, was dealing with hundreds of millions in projected deficits and was actively imposing substantial spending cuts.

Because both parts of the prediction — large shortfalls for both California and San Francisco plus significant fiscal pain — materialized in line with the described scale and timing, the prediction is best judged as right.

Key parts of Chamath’s prediction did not come true:

1. Timing and outcome of the first launch. Relativity’s first and only Terran 1 launch from Cape Canaveral did not occur in the third week of January 2023. After scrubbed attempts on March 8 and 11, the rocket finally launched on March 23, 2023, from LC‑16 at Cape Canaveral, reached space but failed to reach orbit due to a second‑stage problem, and Terran 1 was then retired. (en.wikipedia.org)

2. “$10B order book” unlocking and SpaceX‑like valuation. Public reporting around Relativity’s contracts shows a much smaller backlog: about $1.2B in Terran R launch contracts as of mid‑2022, growing to roughly $1.6B by April 2023 and about $2.9B in backlog by 2025 – nowhere near the ~$10B he described. (en.wikipedia.org) Relativity’s last disclosed funding round valued it at around $4.2B, and later commentary still places it in the single‑digit billions, while SpaceX has been valued on the order of $350B in recent private share sales. (en.wikipedia.org) The company therefore did not “unlock” a $10B launch‑services book nor move onto a valuation trajectory comparable to SpaceX.

3. “If it doesn’t, it goes to zero.” The Terran 1 mission failed to reach orbit, yet Relativity did not go to zero. Instead, the company retired Terran 1 and shifted strategy to focus “all in” on the larger Terran R, with billions of dollars in signed contracts, a planned first launch no earlier than 2026, and ongoing operations under a new CEO and a workforce of around 1,200 employees. (en.wikipedia.org)

Because the launch timing was off, the post‑launch outcomes bore little resemblance to either of Chamath’s binary scenarios, the order book never approached $10B, and the company neither rocketed to SpaceX‑like valuation nor collapsed to zero, this prediction is best classified as wrong.

Evidence from 2023 shows that Europe did remain heavily reliant on U.S. LNG as a direct consequence of the Ukraine war and the rapid build‑out of LNG import capacity, and that U.S. gas exporters were major economic beneficiaries.

On dependence:

• EU LNG imports surged after Russia’s 2022 invasion; in 2023 the EU imported about 134 bcm of LNG, which made up roughly 42% of total EU gas imports, more than double LNG’s share in 2021. (aa.com.tr)
• Across all LNG suppliers, the United States was again the largest by far: in 2023 it supplied about 48% of all LNG imported by Europe (EU‑27 plus UK), the third consecutive year it was number one. (theuncontained.com)
• An EU quarterly gas‑market report for Q1 2023 similarly found that LNG already accounted for about 42% of EU gas imports, with the U.S. providing 41.5% of that LNG — by far the biggest single source. (energy.ec.europa.eu)
• Norway remained the largest pipeline supplier, but by 2023 the U.S. had become the second‑largest overall gas supplier to the EU, underlining how central U.S. LNG became in the post‑Ukraine‑war supply mix. (consilium.europa.eu)

On rapid LNG terminal build‑out:

• ACER’s monitoring and other analyses report that since 2022 the EU added over 50 bcm of new LNG import capacity, with total regasification capacity on track to be more than one‑third higher in 2024 than in 2021. Germany in particular went from zero LNG capacity in 2021 to around 13 bcm by end‑2023, thanks largely to fast‑tracked floating terminals. (aa.com.tr)

On the U.S. gas industry as a major winner:

• By 2023 the United States had become the world’s largest LNG exporter; U.S. LNG export volumes hit record levels and total U.S. natural‑gas exports (pipeline plus LNG) rose 12% year‑on‑year, with Europe taking nearly half of U.S. LNG exports. (en.wikipedia.org)
• Watchdogs and analysts explicitly describe U.S. LNG firms as earning “massive profits” from selling gas to Europe at crisis‑elevated prices after the invasion, framing them as major beneficiaries of Europe’s pivot away from Russian pipeline gas. (citizen.org)

The literal phrase “completely dependent on American natural gas” is hyperbolic, since Europe also relied heavily on Norwegian pipeline gas and non‑U.S. LNG (Qatar, Russia, Algeria, etc.). But the normalized version of the prediction—that in 2023 Europe would remain heavily dependent on U.S. LNG because of the Ukraine war and fast‑tracked terminals, and that this would make the U.S. gas industry a major economic winner—is well supported by the 2023 data, so the prediction counts as right overall.

Key measurable part of Sacks’s prediction was that by the end of 2023 U.S. consumers would "tap out" and the U.S. economy would enter a recession once the Fed raised rates to around 5.5%.

What actually happened:

• The Fed did raise the federal funds target range to 5.25–5.50% in July 2023 and held it there through year‑end, matching his rate call. (forbes.com)
• Credit card APRs did reach record highs around or above 20% by late 2023, consistent with his setup about consumer pressure. (bankrate.com)
• Despite that, real U.S. GDP grew about 2.5–3.1% in 2023, with Q4 2023 running at a 3.3% annualized pace; multiple outlets and official data note that the economy "easily skirted" the widely expected recession. (cnbc.com)
• The NBER and NBER-based recession indicators show no U.S. recession months at any point in 2023; the expansion that began after the April 2020 trough continued through at least the end of 2023. (nber.org)
• Unemployment stayed unusually low (around or below 4%) and BLS/press analyses emphasize that strong consumer spending was the main driver of 2023 growth, contradicting the idea that the consumer had "tapped out." (bls.gov)

Because the central outcome he tied to 5.5% rates—consumers breaking and the U.S. entering recession in 2023—did not occur, the prediction is best judged wrong, even though some of the background conditions (high rates, expensive credit) were accurate.

Key macroeconomic indicators and official arbiters show that the U.S. was not in a recession in early 2023, even though the tech sector was under significant stress.

1. No official or widely accepted U.S. recession in 2022–2023

• The National Bureau of Economic Research (NBER), which is the standard body for U.S. business cycle dating, did not declare a U.S. recession in 2022 or 2023. Their list of business cycle peaks and troughs shows the last recession ending in April 2020 and no new recession dated through 2023–2024.
• U.S. real GDP grew in 2022 Q3 and Q4, and continued to grow through 2023 (annual real GDP growth around 2.5% in 2023, with positive quarterly growth rates). Positive, broad-based GDP growth is inconsistent with the claim that the overall economy was already in recession.
• The labor market remained historically tight: unemployment hovered around 3.4–3.7% in early 2023, near multi‑decade lows, with strong job creation—again inconsistent with an economy already in recession.

2. Tech sector weakness was real but sector‑specific

• Major U.S. tech firms (Meta, Amazon, Microsoft, Google, etc.) carried out large layoffs from late 2022 into early 2023 and experienced significant stock price declines from 2022 highs. That supports the sectoral part of the prediction—tech was in a pronounced downturn.
• However, other major sectors (services, travel, hospitality, manufacturing in aggregate) were not in broad contraction; consumer spending and employment remained resilient. The overall economy was better characterized as a slowdown/"near-miss" or "soft landing" rather than an ongoing recession.

3. Why this is judged wrong rather than ambiguous

• The prediction is not merely that tech would be in a recession, but that “the recession is here” for the U.S. economy, just unevenly distributed. By now (late 2025), with full data and NBER’s dating, it’s clear there was no officially recognized or broadly accepted U.S. recession starting in 2022 or early 2023.
• While one can always define “recession” colloquially, economic consensus (GDP, unemployment, industrial production, income, NBER dating) does not support the claim that the U.S. was already in a recession at that time.

Because the central claim — that the U.S. was effectively already in a recession by early 2023 — is contradicted by subsequent data and official business-cycle dating, the prediction is best classified as wrong, even though the narrower observation about a tech downturn was accurate.

Available data show that late‑stage/growth VC did not snap back within a few months of late 2022 and instead stayed impaired for roughly the 2–3 year window Chamath worried about.

• In 2023, late‑stage venture activity and valuations fell sharply: one review notes that late‑stage capital deployed through Q3 2023 was far below 2022 and that median late‑stage valuations were down more than 20% versus end‑2022, with more companies facing down rounds and a still‑dormant IPO market. (eisneramper.com) This fits the idea of a “toxic” late‑stage environment persisting rather than resolving within months.
• Through 2024, the overall VC market remained well below the 2021 peak, with global VC investment in 2024 still ~55% under 2021 and heavily concentrated in a few AI mega‑rounds, indicating that broad late‑stage/growth funding was still constrained. (barrons.com) At the same time, private‑equity/VC exit value in 2024 hit a five‑year low, and commentators highlighted a liquidity drought and stalled exits—conditions that keep late‑stage and growth rounds difficult and valuation resets in place. (spglobal.com) Carta’s 2024 data showed 2022‑vintage VC funds had deployed only 43% of capital after two years, the slowest deployment of recent vintages, underscoring a persistently cautious market rather than a rapid rebound. (carta.com)
• By late 2024–2025, signs of recovery start to appear—e.g., a 2024 uptick in total VC dollars versus 2023 and, in Europe, 2025 data showing venture‑growth and late‑stage valuations rising again and the share of down rounds falling. (afurrier.com) That recovery timing (becoming meaningfully better in 2024–2025 rather than in early 2023) lines up with a roughly 2–3 year impaired period starting in late 2022.

So while conditions are improving by 2025, the evidence shows that the late‑stage/growth market stayed depressed and difficult for about two years (2023–2024) and only gradually began to normalize in the third year—consistent with Chamath’s prediction that the crisis would last years, not a couple of months.

Aggregated late‑stage venture data show that the typical reset for many growth‑stage startups was on the order Chamath described. Carta’s Q1 2023 data found median Series E+ pre‑money valuations down about 82% year‑over‑year, with Series D medians down ~70%, as the 2021 bubble unwound. (carta.com) Follow‑on data from Carta show median Series D valuations falling from roughly $1.07B at the 2021 peak to about $212M by Q1 2024—an ~80% decline. (carta.com) A number of marquee growth‑stage private tech companies that raised at multi‑billion valuations in 2020‑2021 later saw explicit down rounds or marks in the 80–90% range: Klarna from $45.6B to $6.7B (≈‑85%), Byju’s from $22B to around $1B (≈‑95%), PharmEasy raising with about a 90% cut, and quick‑commerce names like Gorillas and Gopuff taking 80–90% valuation haircuts, with Getir falling from $11.8B to $2.5B (≈‑79%). (en.wikipedia.org) While averages across all down rounds were smaller (often 20–40%), analyses explicitly note that elite 2020–2021 “unicorns” frequently suffered much steeper 80–90% markdowns. (fortune.com) Taken together, this supports Chamath’s claim that the true market‑clearing valuations for many growth‑stage private tech companies from the 2020–2021 cohort ended up roughly 80–90% below their peak private marks, even if not every company fell that far and some later recovered.

The prediction that in 2023 Google Search would be the “biggest business loser,” suffering a ~10–15% loss of usage to AI competitors with a material hit to engagement and profitability, did not occur.

Profitability and revenue: Alphabet’s 2023 results show the opposite of a material deterioration. Alphabet reported 2023 revenue of $307 billion, up 9% year-over-year, and the CFO explicitly stated that Search remained the largest contributor to revenue growth.​urn0search6] SEC filings further show that “Google Search & other” revenue for the first nine months of 2023 rose to about $127.0 billion from $119.8 billion in the same period of 2022 (roughly 6% growth), indicating higher—not lower—search monetization.​urn1search7]

Usage and market share: Public traffic and market-share data show no 10–15% collapse in Google search usage. Statista/StatCounter-based summaries have Google at around 92% global search market share in both 2022 and 2023, with Bing edging up only slightly (e.g., 2023 figures around Google 91.6%, Bing 3.4%).​urn0search7urn0search2] A Search Engine Land analysis of StatCounter data found that from February 2023 to January 2024 Google’s worldwide share dipped from 93.37% to 91.47% (a loss of ~1.9 percentage points), while Bing’s share rose from 2.81% to 3.43% (a gain of 0.62 points)—a minor shift, far from a 10–15% usage loss.​urn0search8]

Impact of AI competitors: Subsequent analyses show that even after the launch of ChatGPT and other AI tools, Google search volume continued to grow. Business Insider and Search Engine Land report that Google processes more than 5 trillion searches annually, with search queries more than 20% higher than at the time of ChatGPT’s late‑2022 launch—indicating that AI chatbots have so far had minimal impact on Google’s core search volumes.​urn1news12urn1search1]

Because Google Search’s usage, revenue, and profitability all grew rather than suffering a material 10–15% decline in 2023, the prediction is wrong.

Jason Calacanis predicted that laid‑off tech workers who band together in small founding teams would become “extremely successful” and the “true big winners” coming out of the 2022–2023 tech layoffs.

What has clearly happened so far:

• There was a very large pool of laid‑off tech workers: over 265,000 tech employees were laid off from the start of 2022 through mid‑February 2023 alone, according to layoffs.fyi data summarized by LETA VC. (leta.vc)
• Multiple analyses show a noticeable share of these workers became founders. Clarify Capital’s “From Fired to Founder” survey of 1,007 laid‑off tech workers found that 63% started their own company, with most reporting higher incomes and improved job security. (clarifycapital.com)
• Other studies using LinkedIn data (e.g., BizReport, summarized by TechStartups) estimate that roughly 13 out of every 100 laid‑off workers across major tech firms went on to start businesses, with especially high founder rates from companies like Meta and DoorDash. (techstartups.com)
• Benzinga and other outlets in 2023 explicitly described a “massive wave of new startups” fueled by roughly 150,000 laid‑off tech workers and noted a jump in Y Combinator applications (5x year‑over‑year in January 2023). (benzinga.com)

However, the outcome Jason predicted is much stronger: that this cohort would be “extremely successful” and become the big business winners of the layoffs over the ensuing years. On that, the evidence is not yet in:

• As of late 2025, reporting focuses mainly on the number of new companies and anecdotal income gains for founders, not on large exits, broad unicorn creation, or clear ecosystem‑level dominance by a distinct “layoff founders” cohort.
• The 2024–2025 VC rebound is driven heavily by AI startups in general, with the biggest funding rounds going to firms like OpenAI and xAI; coverage rarely ties these headline companies directly to 2022–2023 layoff cohorts in a systematic way. (reuters.com)
• Historically, it has taken 5–10 years after downturns for the most successful cohorts (e.g., post‑2000 and post‑2008) to clearly emerge as “big winners,” and those outcomes are measured by major exits, sustained profitability, or transformative market impact. We are only ~2–3 years out from the key 2022–2023 layoffs, far too early to judge that kind of long‑run success at a cohort level.
• Even coverage that originally amplified Calacanis’s prediction (e.g., Fortune in January 2023, which quoted him saying such employees‑turned‑entrepreneurs “could win big”) was speculative, and there is not yet follow‑up evidence showing that this group as a whole has indeed become the standout winners of the period. (fortune.com)

Because many laid‑off workers did start companies, the directional part of his thesis (that layoffs would fuel a startup wave) looks supported. But the stronger claim—that these specific founders would, as a cohort, turn into “extremely successful” long‑term winners—cannot be confirmed or disproven yet. The time horizon is too short and the available data do not track this cohort’s exits or long‑term performance.

Therefore, the prediction’s status as of November 2025 is inconclusive (too early to tell).

Evidence cuts both ways:

1. Macro premise ("sticky, elevated" unemployment) was wrong.

   • In 2023 the U.S. unemployment rate fluctuated around 3.4–3.9%, averaging about 3.8% in Q4 versus 3.6% a year earlier. The Bureau of Labor Statistics explicitly characterizes 2023 as a strong labor market with unemployment remaining low by historical standards, not “elevated.” (bls.gov)
   • The Fed’s own series shows unemployment mostly below 4% throughout 2023. (federalreserve.gov)
     → His causal story (“as unemployment becomes sticky at high levels…”) did not materialize.

2. Gig / marketplace labor supply did surge.

   • DoorDash reported that more than 7 million Dashers earned over $15 billion in 2023, and management described Dasher supply as “the healthiest we’ve seen,” highlighting 7M+ Dashers on the Q4 call. (ir.doordash.com)
   • Uber said its “supply position is the best that it’s been,” with over 6.5 million earners in Q3 2023 and the number of drivers up ~30% year‑on‑year by Q4. (roic.ai)
   • Airbnb ended 2023 with 7.7 million active listings, up 18% YoY, and more than 5 million hosts, noting “sustained double‑digit supply growth across all regions.” (news.airbnb.com)
   • Etsy’s active sellers grew 29% in 2023 vs. 2022, a very large increase in marketplace supply. (investors.etsy.com)
     → The core idea that more people would participate on these platforms is borne out, even if we can’t attribute it to higher unemployment.

3. Business performance: strong for most, mixed for Etsy.

   • DoorDash: 2023 orders +24%, Marketplace GOV +25% to $66.7B, revenue +31% to $8.6B, with a huge improvement in adjusted EBITDA and free cash flow—clear evidence it was a major beneficiary of platform growth. (sec.gov)
   • Uber: company revenue rose from about $31.9B in 2022 to $37.3B in 2023, with trips up mid‑20% YoY and improving profitability—again, substantial business benefit. (venuelabs.com)
   • Airbnb: called 2023 “another incredible year,” with record Q4 revenue ($2.2B, +17% YoY) and 99M nights/experiences booked (+12% YoY), plus hosts earning over $57B; listings and host base hit all‑time highs. (news.airbnb.com)
   • Etsy: active sellers surged, but GMS fell ~1.2% in 2023 and has been essentially flat for several years; the company announced layoffs of 11% of staff citing “very challenging” conditions and “flat” GMS. (investors.etsy.com)
     → Three of the four platforms clearly saw strong business performance alongside higher supply; Etsy saw more supply but not meaningfully better demand.

Net assessment:

• The stated macrocondition (elevated, sticky unemployment) was clearly incorrect.
• The mechanism and outcome (more workers/hosts/sellers on gig platforms and stronger performance for major names like DoorDash, Uber, and Airbnb) were largely correct, though Etsy is a notable exception on performance.

Because key parts of the prediction were right (greater participation and strong results for most named platforms) while its explicit macro driver was wrong and Etsy did not clearly “benefit,” the overall call is best classified as ambiguous, rather than cleanly right or wrong.

Starlink did not conduct an IPO in 2023, nor has it gone public as of early/late 2025. Multiple finance and business outlets note that Starlink remains a unit of privately held SpaceX and is not publicly traded, with no confirmed IPO date on the calendar. (nasdaq.com)

Coverage in mid‑2023 explicitly references Chamath Palihapitiya’s prediction on the All‑In podcast that Starlink would IPO in 2023 at a valuation of at least half of SpaceX’s then private valuation, but reports make clear that this remained only a prediction and that Elon Musk would not (and legally could not) commit to such an IPO timeline. (fortune.com)

Subsequent reporting instead describes SpaceX raising liquidity via private share sales and employee tender offers that valued SpaceX itself as high as about $350 billion, while Starlink continues to operate as a non‑public division. (theguardian.com)

Because no Starlink IPO occurred in 2023 at any valuation, the prediction that Starlink would go public in 2023 with a valuation of at least half of SpaceX’s private value (around $75 billion on a $150 billion SpaceX valuation) and provide Musk with that specific form of liquidity and “breathing room” was not fulfilled.

In 2023, Russia and China did deepen economic ties, but public data show no bilateral economic agreement or package of agreements between just these two countries valued around $1 trillion. Official statistics put total two-way trade in 2023 at about $240 billion, far below $1 trillion, and Russian PM Mishustin in December 2023 cited mutual trade of roughly 18 trillion rubles (~$200 billion) plus about 80 joint investment projects worth almost 20 trillion rubles (~$220 billion), i.e., a combined scale under $450 billion and not a single cross‑sector deal of the size predicted. (uk.finance.yahoo.com) Major reporting on Russia‑China relations and energy cooperation in 2023 and after—such as coverage of gas projects like Power of Siberia 2—likewise contains no mention of any $1 trillion Russia‑China economic deal, and the only trade figures around $1 trillion that appear in this context relate to total internal trade among all BRICS members, not a bilateral agreement. (ft.com) Given the absence of evidence for such a mega‑deal and the much lower documented scale of actual agreements and trade, the prediction that Russia and China would conclude a roughly $1 trillion multi‑sector economic agreement in 2023 did not come true.

Apple’s balance sheet and market value around early 2023 support Jason’s core claim that it was financially feasible for Apple to attempt a takeover of Tesla and that acquiring a large legacy automaker like BMW or Volvo would be far easier.

• Apple’s financial capacity (early 2023): As of December 31, 2022, Apple reported about $20.5B in cash and cash equivalents plus ~$145B in current and non‑current marketable securities, i.e., roughly $165B of cash + liquid securities on hand.(sec.gov) Apple’s equity market cap in January 2023 was around $2.0–2.3T.(statmuse.com) This combination gives Apple enormous capacity to finance very large deals using a mix of existing cash, new debt, and/or stock.
• Tesla’s “depressed” valuation: Tesla’s market cap on January 6, 2023 was about $392B, down sharply from its 2021 peak and roughly matching its end‑2022 market cap of about $389B.(statmuse.com) Apple could not literally “almost buy it with cash” using only its then cash+securities (~$165B vs. a ~$390B target, plus any takeover premium), so that specific phrase is an exaggeration. But given Apple’s trillion‑dollar equity base and very strong credit, a mostly‑cash or cash‑heavy bid financed with some combination of balance‑sheet cash and new debt is financially plausible at that size.
• BMW and Volvo are much smaller: BMW’s total market cap at the end of 2022 was roughly €54–55B (≈$60B).(companiesmarketcap.com) Volvo Cars’ market cap at the end of 2022 was about SEK 141B, which corresponds to roughly $13–14B.(stockanalysis.com) Both are well within the scale of Apple’s 2022 cash + securities alone and trivial relative to its ~$2T market cap, making it accurate to say Apple could have easily financed an acquisition of BMW or Volvo in purely financial terms.
• No such deals occurred (but that wasn’t the claim being graded): Apple’s actual M&A activity through 2025 remains focused on smaller tech and AI‑related companies—there is no Apple acquisition of Tesla, BMW, or Volvo on its acquisitions list.(en.wikipedia.org) However, Jason’s normalized prediction, as you framed it, is about feasibility, not that such a takeover would in fact occur.

Overall, the numbers show that Apple was more than capable of financing a bid for BMW or Volvo outright and had the financial firepower to mount a serious, largely cash‑funded run at Tesla at its early‑2023 valuation. The narrow “almost buy it with cash” line is overstated, but the normalized claim that such moves were financially feasible is substantively correct, so this prediction is best classified as right.

• No major health M&A spree, and none of the named targets were acquired. Amazon did complete its previously announced $3.9B acquisition of primary‑care provider One Medical in February 2023, strengthening its primary‑care and virtual‑care footprint. (health.economictimes.indiatimes.com) But from 2023 through late 2025 there is no record of Amazon acquiring Ro, Hims & Hers, Peloton, or Whoop; all continue to operate as independent companies (Hims doing its own acquisitions such as ZAVA, Ro partnering with Lilly, Peloton remaining a standalone public company, and Whoop launching new devices and policies under its own brand). (investors.hims.com) Coverage of healthcare M&A in 2025 also notes that big buyers like Amazon have become cautious and are not doing large new health deals, further contradicting the prediction of an aggressive acquisition push. (businessinsider.com)

• Healthcare is still treated as an experimental or “potential” pillar, not an actual fourth pillar on par with Marketplace, Prime, and AWS. Internal figures reported by The Information (via Business Insider) suggest Amazon’s healthcare unit generated about $2.5B in revenue but lost roughly $1.3B in 2023, and explicitly describe healthcare as a potential fourth pillar, not one that has already reached that status. (theinformation.com) Other analyses say Amazon is still searching for its next big pillar business, with initiatives like Project Kuiper (satellite internet) and Zoox (autonomous vehicles) also framed as candidates, and note that bets such as healthcare have so far struggled to become large, profitable franchises. (livemint.com) Even optimistic coverage only describes healthcare as an area where Amazon could eventually find a fourth pillar, underscoring that this has not yet happened in practice. (geekwire.com)

• Operational reality looks like ongoing experimentation and restructuring, not a mature fourth pillar. Amazon has continued to invest in and reorganize its health efforts (PillPack/Amazon Pharmacy, One Medical, Amazon Clinic integration, new telehealth offerings, and partnerships like Eli Lilly’s LillyDirect and nutrition startup Fay), but these are being refined and restructured—its healthcare business was split into six units and has seen executive turnover and job cuts—rather than celebrated as a core, proven business on the scale of AWS or Prime. (reuters.com) Nearly three years after the January 2023 prediction, health remains a strategic bet with mixed financial performance and no broad consensus that it has become Amazon’s definitive “fourth pillar.”

Given that: (a) Amazon did not buy Ro, Hims, Peloton, Whoop, or similar marquee wellness companies; (b) there was no broad health‑M&A push starting in 2023; and (c) healthcare has not clearly achieved pillar status alongside Marketplace, Prime, and AWS, the prediction that Amazon would quickly make health its fourth pillar via a wave of such acquisitions has not come true.

Key elements of Jason’s prediction have effectively come true, even if the final transaction is still being implemented and the timing slipped slightly beyond 2024.

1. Escalating bipartisan U.S. political pressure in the 2023–2024 cycle
   In April 2024, during the 2024 election cycle, Congress passed the Protecting Americans from Foreign Adversary Controlled Applications Act (PAFACA). The law explicitly targets ByteDance and TikTok, and passed with large bipartisan majorities in both House and Senate before being signed by President Biden on April 24, 2024. (en.wikipedia.org) This matches Jason’s premise of strong, bipartisan political pressure focused on forcing action against TikTok.

2. Forced divestiture of TikTok’s U.S. operations
   PAFACA requires that TikTok’s Chinese parent, ByteDance, divest TikTok’s U.S. operations by January 19, 2025, or face a nationwide ban via removal from app stores and hosting services. (en.wikipedia.org) TikTok and ByteDance sued, but the D.C. Circuit upheld the law in December 2024, and the U.S. Supreme Court upheld it on January 17, 2025, confirming that ByteDance must either divest TikTok or see the app effectively banned in the U.S. (cnbc.com) That is precisely a forced divestiture under duress.

3. Sale rather than status quo; mechanism broadly matches “IPO or sale”
   Instead of an IPO, the chosen path is a sale and restructuring: President Trump’s September 25, 2025 executive order accepts a “qualified divestiture” plan where TikTok’s U.S. app will be operated by a new U.S.-based joint venture that is majority-owned and controlled by U.S. persons. ByteDance and affiliates are to own less than 20% of the new entity. (whitehouse.gov) Reuters reporting in November 2025 similarly states that ByteDance is in the process of divesting about 80% of its U.S. assets, with a consortium led by Oracle and Silver Lake expected to hold the majority stake. (reuters.com) That clearly fits Jason’s “IPO or sale” framing: it is a forced sale.

4. Chinese ownership “largely removed” from the U.S. TikTok business
   Under the approved plan, U.S. investors will control the new U.S. TikTok venture, and ByteDance’s direct equity stake is capped below 20%. (whitehouse.gov) While ByteDance (and thus Chinese ownership) is not completely eliminated, its role is reduced to a non‑controlling minority interest, and there are ongoing concerns over licensing of the algorithm. (reuters.com) That situation aligns well with the prediction’s phrase “largely or entirely removed”—it is largely, though not entirely, removed.

5. Timing nuance
   The compelling political and legal pressure that forces divestiture—PAFACA and its bipartisan passage—did occur during the 2023–2024 election cycle. (en.wikipedia.org) The concrete divestiture structure and implementing executive order came in 2025, but they are a direct consequence of that 2024 law. Given the substance of the prediction (that bipartisan pressure during that cycle would force Chinese owners to divest most of their U.S. TikTok stake via a sale), this is a timing slippage rather than a fundamental miss.

Overall, by late 2025 TikTok’s Chinese parent is under a binding U.S. legal mandate to divest, a divestiture framework has been approved, and control of U.S. TikTok is shifting to a U.S.-controlled entity with ByteDance retaining only a minority stake. That matches the core of Jason’s prediction, so it is best judged as right, with minor caveats on timing and the fact that some Chinese ownership remains as a small minority interest.

Evidence from 2023 shows friction at times but not a clear breakdown of the Biden–Zelensky relationship or a fundamental rift between U.S. and Ukrainian positions.

• In February 2023 Biden made a highly symbolic, risky wartime visit to Kyiv, publicly declaring that the U.S. “stands with” Ukraine and announcing another $500 million in military aid, a strong signal of personal and political solidarity rather than a rupture.【3search18】
• Over the course of 2023, the U.S. provided Ukraine with 34 military aid packages totaling about $24 billion, including advanced air-defense systems, tanks, and other major capabilities; Zelensky publicly thanked the United States and emphasized the importance of U.S. leadership in the pro‑Ukraine coalition.【3search0】
• When Zelensky visited Washington in December 2023, Biden stood beside him at the White House and insisted that the U.S. “cannot leave Ukraine without help” and must keep providing weapons, while Zelensky highlighted battlefield gains—again presenting a united front and shared strategic narrative rather than open divergence.【3search1】
• The most visible 2023 dispute was Zelensky’s public complaint at the July NATO summit that the lack of a clear timetable for Ukraine’s NATO membership was “unprecedented and absurd,” implicitly criticizing cautious members like the U.S. and Germany.【1search0】 However, coverage noted this outburst mainly as negotiating brinkmanship; Zelensky later described the summit outcome as a meaningful success and thanked NATO leaders for support, indicating tension but not a lasting rift.【1search6】
• A December 2023 Washington Post analysis did describe “strain in the relationship” between Kyiv and Washington, primarily due to U.S. congressional delays on funding, but Ukrainian officials still stressed that U.S. partners had never broken their promises and that American support remained crucial—signs of stress within an ongoing partnership, not its collapse.【2search0】
• Subsequent events under Biden reinforce that the personal relationship did not "end" in 2023: in 2024, after Biden exited the re‑election race, Zelensky publicly praised him and his decisions on Ukraine as “tough but strong,” underscoring continued respect and alignment.【3news11】

Taken together, 2023 featured disagreements and pressure tactics but not the clear end of a “bromance” or a decisive, public split in U.S.–Ukrainian positions. The prediction that this close cooperative relationship would break down during calendar year 2023 is therefore wrong.

Evidence shows that Ukraine did in fact launch a large counteroffensive in spring–summer 2023:

• Ukrainian forces began shaping operations and intensified attacks in May 2023, and on June 4, 2023, Ukraine formally launched what was widely described as a major or large-scale counteroffensive against Russian positions in southeastern Ukraine (notably in the Zaporizhzhia and Donetsk regions). Major international outlets and defense analyses consistently refer to this as Ukraine’s counteroffensive or major counteroffensive of 2023.
• The campaign ran through the summer and into the fall of 2023, involving multiple brigades, extensive use of Western-supplied equipment, and significant combat along several sectors of the front—meeting any reasonable description of a “massive” counteroffensive.

On the narrower interpretation tied to the original quote — “there’s going to be a massive Ukrainian counteroffensive in the spring” — this is straightforwardly correct: there was a large Ukrainian counteroffensive initiated in early June 2023, which falls within the commonly used military/political timeframe for a spring–summer campaign season and was widely anticipated and reported as such.

The normalized version you provided adds an outcome clause (stalemate or a return to pre‑invasion lines). That part is only half‑right:

• The counteroffensive did not succeed in pushing Russian forces back to the pre‑February 24, 2022 lines; Russian forces still occupy substantial portions of Ukrainian territory as of late 2025.
• However, the fighting largely resulted in limited territorial changes and a de facto stalemate along much of the front, with only modest Ukrainian gains (e.g., around Robotyne and parts of the southern front) and no decisive breakthrough.

So:

• If we judge strictly by the original prediction (a massive spring counteroffensive will happen), it was right.
• The additional outcome conditions in the normalized text are partially wrong (no restoration to pre‑invasion lines) but partially right (the offensive led to something close to a stalemate).

Balancing these, and given that the core, clearly articulated claim in the quote was about the occurrence and timing of a massive Ukrainian counteroffensive, the prediction is best scored as right overall, with the caveat that the more optimistic territorial outcome did not fully materialize.

Multiple reputable sources report that Ukraine launched a major, widely described as "large-scale" or "massive," counteroffensive against Russian forces in early June 2023, which is late spring by meteorological seasons.

• The BBC reports that Ukraine’s "long-awaited counter-offensive" began in early June 2023, detailing Ukrainian advances and offensive operations along several fronts in the south and east.
• The Institute for the Study of War (ISW) and other military analysis outlets describe the June 2023 operations as Ukraine’s main 2023 counteroffensive, involving multiple brigades and extensive Western-supplied equipment.
• Major news organizations (e.g., Reuters, The New York Times) characterize the June 2023 action as a significant or large-scale counteroffensive, confirming its scale and intent.

Because this substantial counteroffensive did in fact occur in spring 2023 (as commonly used in military and media descriptions, and calendar-wise at the very end of spring/early summer), Sacks’s prediction that “there’s going to be a massive Ukrainian counteroffensive in the spring” is best judged as right.

Evidence from 2023 indicates both parts of the conditional prediction came true in the real-world scenario that actually unfolded.

1. The 2023 counteroffensive led to a de facto stalemate, not a rollback to pre‑Feb. 24, 2022 lines.
   Ukraine’s much‑anticipated 2023 offensive, launched in June, recaptured about 370 km² and a small number of villages but failed to reach key objectives like Tokmak or the Sea of Azov. By early December, multiple outlets and analysts assessed the offensive as stalled or failed, and in November 2023 Ukraine’s commander‑in‑chief Valerii Zaluzhnyi publicly described the war as a “stalemate.” (en.wikipedia.org) This matches the stalemate branch of the predictor’s conditional.

2. By mid‑to‑late 2023, U.S. and Ukrainian strategic interests and preferred war end‑states were visibly diverging.

   • Different war aims and risk tolerances:
     – An April 2023 analysis from the Towson University Journal of International Affairs explicitly argued that the United States and Ukraine already had different ideas of what “winning” meant: Kyiv sought full liberation including Crimea, while Washington was more focused on avoiding escalation and broader confrontation with Russia. (wp.towson.edu)
     – Ukraine’s foreign minister Dmytro Kuleba repeatedly insisted that any peace initiative must restore territorial integrity and reject both territorial cession and a “frozen” conflict, underscoring Kyiv’s maximalist territorial aims. (en.wikipedia.org)

   • Operational and escalation disagreements:
     – U.S. officials were clearly uncomfortable with Ukrainian strikes inside Russia. In May 2023, after drone attacks in the Moscow area, the White House press secretary said the United States did not support attacks inside Russia, even as Ukrainian officials publicly welcomed bringing the war home to Russia and Ukrainian intelligence pursued deep‑strike operations. (washingtonpost.com)

   • Divergent strategic concepts after the failed offensive:
     – A December 11, 2023 New York Times piece (widely discussed and summarized by secondary sources) reported that many U.S. officials wanted Ukraine to adopt a conservative “hold what you have and build up over 2024” strategy aimed at sustainability and eventually creating conditions for negotiations, while Ukrainian leaders still pushed for renewed offensive action and more long‑range strikes to regain territory and keep global attention. (reddit.com) That is a clear example of diverging strategic preferences rather than a fully shared theory of victory.

   • Security architecture and NATO friction:
     – Ahead of the July 2023 Vilnius NATO summit, Kyiv hoped for a concrete pathway or invitation toward NATO membership. The U.S. position—articulated by President Biden—was that Ukraine was “not ready” and that immediate membership was off the table, despite strong support for Ukrainian membership from many allies. This gap over long‑term security guarantees and alliance status became a public point of tension between Kyiv and Washington (and some European capitals). (en.wikipedia.org)

Taken together, by late 2023 there was an openly discussed gap between U.S. and Ukrainian strategic interests and aims: the U.S. increasingly prioritized escalation management, resource limits, and a sustainable defensive posture likely leading toward negotiations, while Ukraine prioritized complete territorial restoration, aggressive operations (including inside Russia), and faster, more extensive Western support. That is a noticeable divergence of strategic interests within the 2023 timeframe the prediction specified.

Because (a) the actual counteroffensive outcome matched the stalemate branch, and (b) a clear and widely reported divergence in strategic interests did in fact emerge during 2023, the prediction is best classified as right.

Across the main inflation measures and major forecasters, 2023 inflation generally fell as fast or faster than the consensus expected at the start of 2023, not more slowly as Chamath predicted.

Key evidence:

1. US inflation fell sharply in 2023.

   • Headline CPI went from 6.5% year‑over‑year in December 2022 to 3.4% in December 2023—almost a halving in 12 months.

     • BLS release for Dec 2022: CPI‑U +6.5% y/y.

     • BLS release for Dec 2023: CPI‑U +3.4% y/y.

     (bls.gov)

   • The Fed’s preferred gauge, headline PCE, fell from about 5.0% y/y in February 2023 to 2.6% y/y by December 2023, very close to the 2% target. (bea.gov)

2. Official forecasts going into 2023 expected more persistent inflation than actually occurred, at least for key advanced‑economy benchmarks.

   • The Federal Reserve’s December 2022 Summary of Economic Projections put median 2023 PCE inflation at 3.1%, with a central tendency of 2.9–3.5%. (federalreserve.gov) Actual year‑end PCE inflation was 2.6% y/y in December 2023—below the Fed’s own 2023 projection and pointing to faster disinflation than that consensus.
   • The IMF’s January 2023 World Economic Outlook Update projected global inflation falling from 8.8% in 2022 to 6.6% in 2023, still well above pre‑pandemic levels, i.e., a gradual, not cliff‑like, decline. (imf.org) By April 2024, IMF Managing Director Kristalina Georgieva stated that headline inflation in advanced economies had fallen to about 2.3% by Q4 2023 from 9.5% 18 months earlier and was easing faster than anticipated—explicitly acknowledging that disinflation beat prior expectations. (reuters.com)

3. Actual inflation outcomes show a large drop from 2022 to 2023 in advanced economies.

   • Compiled IMF/World Bank data (as summarized on Wikipedia) show average annual consumer inflation in the US falling from 8.0% in 2022 to 4.1% in 2023, the euro area from 8.8% to 5.6%, and the UK from 9.1% to 6.8%. (en.wikipedia.org) These are very large single‑year declines in inflation.
   • In the US specifically, a number of commentators dubbed 2023 the “year of disinflation”, with at least some noting that inflation was slowing even faster than they had expected earlier in the year. (seekingalpha.com)

4. Some individual forecasts did slightly underestimate full‑year average inflation, but this does not overturn the broad picture.

   • For example, a January 2023 summary of business/economic forecasts cited an expected 3.8% increase in the US CPI for 2023, whereas the BLS later reported an annual‑average CPI increase of 4.1%—a modest overshoot. (nysut.org) That suggests slightly stickier inflation on that particular metric, but against the backdrop of (a) much lower year‑end readings and (b) official forecasts like the Fed’s PCE projection being too high, the overall theme is that disinflation was at least as rapid as mainstream expectations, and often faster.

5. Did inflation stay above what “people want”?

   • Yes: by the end of 2023, US headline PCE was 2.6% and core PCE 2.9% year‑over‑year—still above the Fed’s 2% target and thus above what central bankers and many consumers would consider “normal.” (bea.gov) But Chamath’s prediction was not just that inflation would remain above 2%; it was specifically that it would not fall as fast or as meaningfully as people/consensus expected. On that relative‑to‑expectations dimension, the evidence above shows inflation generally beat the official consensus in advanced economies.

Putting this together:

• Inflation in 2023 did remain above 2%, so it was still somewhat elevated vs the ideal target.
• However, relative to the major starting‑2023 forecasts (Fed, IMF and others), the speed and magnitude of disinflation in advanced economies—especially the US—were broadly stronger than expected, with prominent institutions later describing the decline as faster than anticipated.

Because his normalized prediction hinges on inflation failing to fall as quickly as consensus expected, and the best evidence shows that in many key measures it actually fell at least as quickly, and often faster, than those expectations, this prediction is best classified as wrong overall.

Evidence from 2023 shows that U.S. inflation fell roughly as much as, or more than, mainstream forecasters expected, despite only moderately strong and cooling wage growth, which contradicts Chamath’s contrarian claim that wage inflation would keep inflation from falling as much as expected.

Key points:

1. What forecasters expected for 2023 inflation (made before or around Chamath’s Jan 2023 prediction):

   • The Federal Reserve’s December 2022 Summary of Economic Projections put 2023 headline PCE inflation at 3.1% and core PCE at 3.5% (Q4/Q4). (federalreserve.gov)
   • The First Quarter 2023 Survey of Professional Forecasters (SPF), published Feb 10, 2023, projected 2023 Q4/Q4 headline PCE inflation at 2.8%, core PCE at 3.0%, and CPI at 3.1%, with those projections already revised down from late‑2022 values. (philadelphiafed.org)

2. What actually happened to U.S. inflation in 2023:

   • U.S. CPI inflation fell sharply: year‑over‑year CPI dropped from about 6.4% in January 2023 to 3.35% in December 2023. (inflation.eu) The average annual CPI rate for 2023 was about 4.1%, down from 8.0% in 2022. (officialdata.org)
   • On the Fed’s preferred PCE measure, inflation declined even more decisively: headline PCE fell from 4.4% in April 2023 to 2.6% by November–December 2023, and core PCE fell from 4.8% to 2.9% over the same period. (indiainfoline.com) By December 2023, headline PCE was 2.6% year‑over‑year, below both the Fed’s earlier 3.1% projection and the SPF’s 2.8% Q4/Q4 forecast, and core PCE (2.9%) was also a bit below the SPF’s 3.0% projection. (philadelphiafed.org)
   • At the global level, the IMF’s January 2024 World Economic Outlook Update explicitly notes that inflation was falling faster than expected in most regions, indicating that consensus forecasts in early 2023 had generally over‑estimated inflation’s persistence. (imf.org)

3. Wage inflation in 2023 did not stay so strong as to block this disinflation:

   • The Employment Cost Index shows 12‑month growth in employer compensation costs declining from a 2022 high of 5.1% to 4.2% in Q4 2023; wage and salary growth for private‑sector workers cooled, even though it remained above CPI. (ilr.cornell.edu)
   • The Atlanta Fed Wage Growth Tracker reports wage growth around 5.2% in September–October 2023, down from much higher peaks (around the mid‑6% range) in 2022, again indicating moderating, not accelerating, wage inflation. (atlantafed.org)

4. Reconciling the prediction with the data:

   • Chamath’s claim was that strong wage inflation in 2023 would prevent inflation from coming down as much as most forecasters expected. In reality, both headline and core inflation fell faster or to lower levels than key forecasters (Fed SEP, SPF) had penciled in, despite wage growth that was elevated but clearly decelerating.
   • The IMF’s later assessment that inflation fell faster than expected in most places reinforces that, ex post, the contrarian bet that inflation would prove too sticky relative to forecasts did not pay off. (imf.org)

Because realized 2023 inflation was at or below the levels forecast by mainstream economists and policymakers—rather than being held up by wage inflation—this prediction is best classified as wrong.

By the end of 2023 there were clear, observable signs of incremental weakening at the margins of the dollar’s unquestioned dominance—of exactly the sort the normalized prediction specified—even though the dollar remained firmly the top reserve currency.

1. Non‑USD settlement in major trade blocs/energy trade

   • China and Brazil reached a deal in March 2023 to conduct trade in their own currencies, establishing a yuan clearing bank in Brazil so transactions could avoid the dollar.(en.mercopress.com)
   • China’s CNOOC and France’s TotalEnergies completed China’s first LNG trade settled in yuan in March 2023, followed by further yuan‑settled LNG trades (e.g., with Engie and others) later in 2023. These were explicitly framed as efforts to promote multi‑currency pricing and reduce reliance on the dollar in energy trade.(brecorder.com)
   • Saudi Arabia’s finance minister stated in January 2023 that the kingdom was open to trading in currencies other than the U.S. dollar for the first time in nearly 50 years—symbolically important given Saudi Arabia’s central role in oil markets and the historical “petrodollar” system.(en.wikipedia.org)

2. Gradual erosion in reserve-share metrics (but not a collapse)

   • IMF COFER data show the dollar’s share of global FX reserves continuing its long, slow decline—from about 70% in 2000 to around 59.2% by Q3 2023.(countercurrents.org)
   • A 2023–24 synthesis of COFER data puts the dollar at roughly 58–59% of allocated reserves by Q4 2023, confirming a multi‑decade downtrend rather than a sudden collapse.(bestbrokers.com)

3. How this lines up with the prediction

   • The prediction (as normalized) only required that by the end of 2023 there be observable signs that the dollar’s de facto reserve‑currency status had begun to weaken, such as increased use of non‑USD settlement in major trade blocs. Those signs clearly existed in 2023 via high‑profile energy and trade deals in local currencies and explicit political moves (Saudi signaling, BRICS rhetoric) to explore non‑dollar options.(en.wikipedia.org)
   • At the same time, major analyses (e.g., the Atlantic Council’s 2024 “Dollar Dominance Monitor”) emphasize that the dollar’s overall dominance in reserves, invoicing, and FX turnover remained secure and that BRICS de‑dollarization made little tangible progress by standard macro metrics.(investing.com)

Taken together, 2023 did not mark a collapse of dollar dominance, but it did deliver exactly what Friedberg described: visible, widely discussed early moves toward diversification away from the dollar—especially in energy and South–South trade—consistent with the “beginning of the end” of completely unquestioned dollar reserve dominance. Hence, on the normalized criterion (“observable signs”), the prediction is best scored as right.

Market behavior in 2023 still treated the U.S. dollar primarily as a safe‑haven asset rather than as a risk asset. During the U.S. regional‑bank turmoil in March 2023, investors dumped risk assets and rotated into the dollar, explicitly described as a “safe haven” bid for USD, which is the opposite of risk‑asset behavior. (ofx.com) As confidence in the banking system returned later that month, the dollar fell as risk appetite improved, again matching traditional safe‑haven dynamics (strong in risk‑off, weaker in risk‑on). (investing.com)

Looking at broader correlations, analysis of the U.S. Dollar Index (DXY) versus the S&P 500 shows a strongly negative 1‑year correlation (around –0.8) over the period spanning late‑2022/2023, meaning the dollar tended to rise when equities fell and vice versa—classic safe‑haven, not risk‑asset, behavior. (thesteadytrader.com) Contemporary and subsequent market commentary continues to classify the USD as a core global safe‑haven currency alongside gold and the yen, with investors still flocking to dollars in periods of heightened risk. (ecmarkets.mu) Because 2023 data and narratives show the dollar retaining its inverse relationship to risk assets and its safe‑haven role, the prediction that it would begin trading more like a risk asset that year did not materialize.

Evidence from 2023–2025 supports Jason’s directional claim that U.S. “exceptionalism” has held up or strengthened relative to rivals, while Russia–China–Saudi efforts have not produced a credible alternative currency order.

1. U.S. relative strength has remained high or improved.
   IMF updates in 2024–2025 describe the U.S. as the clear growth out‑performer among major advanced economies, with stronger productivity, a leading tech sector, and upgraded growth forecasts, while the euro area and other advanced peers face weaker prospects. (imf.org) This keeps the U.S. far ahead in size and influence among advanced economies.

2. Dollar dominance remains intact; no rival reserve system has emerged.
   IMF COFER data for 2024–2025 show the U.S. dollar still accounting for roughly 56–58% of allocated global FX reserves, with the euro around 20% and the renminbi only about 2%. The IMF notes that, after adjusting for exchange‑rate moves, the dollar’s share is “little changed,” and it remains the primary reserve and invoicing currency. (leap-insights.org) Central banks are slowly diversifying and using slightly more yuan and gold, but this is a gradual multipolar drift, not a regime change.

Experiments around de‑dollarisation—Saudi‑China yuan oil settlements, petroyuan narratives, and regional payment systems—have been modest. Analyses stress that Gulf currencies remain pegged to the dollar, about half of world trade and ~80% of oil sales are still dollar‑denominated, and yuan‑priced oil contracts since 2018 have not significantly dented dollar dominance. (thenationalnews.com) BRICS has promoted using national currencies and launched BRICS Pay as messaging infrastructure, but Kremlin and BRICS statements in 2024–2025 explicitly say there is no imminent common BRICS currency; de‑dollarisation talk has not yielded a new reserve currency. (reuters.com) Overall, Russia–China–Saudi and the broader BRICS camp have not “moved the currency” in the sense of displacing the dollar‑centric order.

3. Rival bloc behavior shows self‑inflicted constraints and internal friction.
   Russia’s invasion of Ukraine produced heavy sanctions, a war‑driven fiscal and industrial tilt, and rising dependence on China. By 2025, roughly 40% of Russia’s federal spending is on defense and security, crowding out social and economic development, while sanctions and capital controls damage long‑term growth. (en.wikipedia.org) Meanwhile, secondary U.S. sanctions led about 98% of Chinese banks to stop processing direct yuan payments from Russia by mid‑2024, sharply curbing the very financial links meant to undercut the dollar. (en.wikipedia.org) Chinese exporters have also used Russia’s dependence to sharply raise prices on war‑related goods—an 87% price increase to Russia versus ~9% elsewhere—highlighting opportunistic, asymmetric relations rather than a tightly aligned bloc. (ft.com)

Saudi Arabia and Russia’s oil partnership has produced deep OPEC+ cuts that propped prices somewhat but also pushed Saudi GDP into contraction in 2023 and cut Russian oil‑and‑gas revenues, illustrating costly self‑inflicted adjustments. (re-russia.net) OPEC+ and BRICS meetings have been marked by quota disputes, over‑production by some members, and resistance from countries like Brazil and India to an explicitly anti‑U.S., China‑dominated line. (ogj.com) These dynamics fit the prediction’s spirit that these states would undercut each other’s leverage and struggle to form a coherent, effective counter‑order.

Taken together, as of late 2025 the U.S. still enjoys clear economic and financial primacy, and the Russia–China–Saudi/BRICS efforts have not created a viable alternative reserve currency or fundamentally shifted the global currency order, while suffering from sanctions, internal frictions, and costly policy choices. Although the very long‑term future is uncertain, the prediction about the direction of the “coming years starting from 2023” has been borne out so far, so it is best classified as right given current evidence.

Evidence supports that this prediction largely came true.

1. Increasing sophistication of Western weapons to Ukraine (early 2023)

   • In January 2023, the U.S. and several European allies agreed to send modern Western main battle tanks (M1 Abrams, Leopard 2, Challenger 2) to Ukraine, a major step‑up from earlier equipment.
   • Through early 2023, the U.S. and its allies also expanded deliveries of longer‑range precision munitions (e.g., GLSDB was announced, and later in 2023 the U.S. provided ATACMS), advanced air defenses (Patriot systems), and more sophisticated armored vehicles and artillery.
   • This clearly fits “more and more sophisticated weapons, more and more support.”

2. Support for a major Ukrainian counteroffensive in spring/summer 2023

   • The widely reported 2023 Ukrainian counteroffensive began in early June 2023, following months of Western training and equipment deliveries specifically intended to prepare new Ukrainian brigades for offensive operations.
   • Western officials and media throughout spring 2023 described Western assistance as “shaping” or “setting conditions” for this expected counteroffensive, including the coordinated tank and armored vehicle packages announced in January–February 2023.
   • While the main push started in early June (late spring/early summer), the intent and preparation through spring 2023 were clearly to enable a large counteroffensive.

Given:

• Trend: clear escalation in sophistication and scale of Western weapon deliveries into 2023.
• Culmination: those deliveries were explicitly framed by U.S. and allied governments as preparation for, and support of, a major Ukrainian counteroffensive that launched in mid‑2023.

The core substance of the prediction — continued escalation of sophisticated Western support leading up to a major Ukrainian counteroffensive in spring 2023 — is borne out by events. The exact start date of the offensive edges into early summer, but in common political/military discourse it is still referred to as the spring–summer 2023 counteroffensive that Western aid was designed to enable.

Conclusion: The prediction is best judged as right.

Available performance data for 2023 show that early‑stage private technology investing (seed–Series A) did not outperform other major asset classes.

The closest broad benchmark for early‑stage venture is the Cambridge Associates U.S. Venture Capital Index. For calendar year 2023, that index had a negative return of ‑3.4%, as venture continued to work through valuation resets, weaker exits, and a difficult fundraising environment.(cambridgeassociates.com) This is inconsistent with being the best‑performing asset class.

By contrast, major public and liquid asset classes were strongly positive in 2023:

• The S&P 500 delivered a total return of about 26.3%.(slickcharts.com)
• The Nasdaq‑100, heavily weighted to large‑cap tech, was up about 53.8% on a price basis.(nasdaq.com)
• Bitcoin returned roughly 155% in 2023, and financial press coverage explicitly notes that Bitcoin outperformed every other major asset class that year.(statmuse.com)

Within private markets, 2023 was broadly described as a weak year for venture capital: global startup and early‑stage funding fell more than 40% year‑over‑year, and industry commentary emphasizes a challenging exit environment and continued markdowns rather than outsized realized gains.(news.crunchbase.com) Even if early‑stage tech deals held up somewhat better than later‑stage venture on paper, there is no evidence they generated returns exceeding those very strong public‑market and crypto benchmarks, and the main venture index being negative essentially rules this out.

Because at least one clearly recognized major asset class (Bitcoin) and multiple others (Nasdaq‑100, S&P 500) had far higher 2023 returns than venture capital, Jason’s prediction that early‑stage private tech investing would be the best‑performing asset class of 2023 was wrong.

Evidence on short‑term U.S. government yields in 2023 matches Chamath’s prediction.

• 3‑month T‑bills: Average monthly yields in 2023 were already around his range and moved above 5% starting in April 2023; December 2023’s average 3‑month T‑bill rate was about 5.39%. (otm.finance.harvard.edu)
• 6‑month T‑bills: The 6‑month Treasury bill rate on December 31, 2023 was about 5.07%, right in the 4.5–5%+ band he described. (ycharts.com)
• 2‑year Treasuries: The 2‑year Treasury yield was about 4.58% at year‑end 2023, again within his “roughly 4.5–5%” range. (ycharts.com) Other 2023 data show 2‑year yields spending much of the year around 4–5%, peaking above 5% in early October and sitting near 4.7% by late November. (jpmorgan.com)

These are standard U.S. government securities, considered very low credit risk when held to maturity; price volatility on short‑dated bills is minimal, and even 2‑year notes carry much less risk than equities or long‑duration bonds. Over 2023, investors who rotated into short‑term Treasuries could indeed earn roughly 4.5–5% annualized with low risk while “waiting for things to become more certain,” as Chamath suggested. Therefore, the prediction is right.

Neel Kashkari did project that the fed funds rate might need to rise to about 5.4%; he reiterated this outlook in a February 7, 2023 CNBC interview, placing himself above the then‑consensus terminal rate of ~5.1%.(cnbc.com) In practice, the Federal Reserve did not reach 5.4% in Q1 2023: after hikes on February 1 and March 22, the target range ended Q1 at 4.75–5.00%, and was later raised to 5.00–5.25% on May 3 and then to 5.25–5.50% in July.(ycharts.com)

However, Sacks’s substantive prediction was about what short‑term T‑bill yields would look like in that rate environment: that they would be roughly 5–5.5% risk‑free. That part proved accurate. Data for 3‑month T‑bills show monthly average yields of 4.63–4.77% in January–March 2023, rising to 5.01% in April and then staying in the 5.0–5.5% band (about 5.01–5.47%) from April through December 2023.(otm.finance.harvard.edu) Similarly, the 1‑year Treasury yield averaged 4.94% in 2023, with an intrayear high around 5.44% and a year‑end level near 5.35%, again matching the 5–5.5% range he described.(app.macrotrends.net)

So while the exact timing/level of Kashkari’s 5.4% call in Q1 did not materialize, the world that actually unfolded—fed funds in the mid‑5s later in 2023—produced short‑term T‑bill yields very close to 5–5.5% for much of the year. Sacks’s economic claim that a Kashkari‑style rate path would make short‑term T‑bills yield about 5–5.5% risk‑free was borne out by realized 2023 T‑bill yields, so the prediction is best judged as right on its core content.

Evidence from 2023–2025 shows that the U.S. did continue very large, multi‑year infrastructure and industrial-policy spending programs, largely justified in terms of economic stimulus, energy transition, and national/security‑related supply‑chain resilience—and other major economies followed similar paths.

In the U.S., the Bipartisan Infrastructure Law (IIJA), CHIPS and Science Act, and Inflation Reduction Act (IRA) together underpin hundreds of billions in ongoing federal infrastructure outlays and crowd in large private investment. By November 2024 the White House reported that these laws had already catalyzed about $756 billion in public infrastructure spending and $1 trillion in private investment, including $568 billion announced for more than 66,000 infrastructure projects (roads, bridges, water systems, etc.), with disbursements stretching through the mid‑2020s. (bidenwhitehouse.archives.gov) Independent analyses note roughly $550 billion in new federal infrastructure projects from the IIJA plus at least $370 billion (and potentially far more) for clean‑energy and grid‑related infrastructure via the IRA, all structured as multi‑year programs rolling out beyond 2023. (wri.org) Treasury and DOE reviews in 2023 describe rapid growth in clean‑energy and energy‑infrastructure investment driven by these laws (including loan programs specifically to retool or replace energy infrastructure), confirming that large, stimulus‑like infrastructure spending continued throughout 2023 and after. (energy.gov) At the same time, the CHIPS Act has awarded multibillion‑dollar grants in 2023–2024 to companies like Intel, TSMC, and GlobalFoundries to build or expand fabs in the U.S., justified explicitly on economic and national security / supply‑chain resilience grounds—another form of large, security‑motivated infrastructure/industrial investment. (en.wikipedia.org)

Other major economies have mirrored this pattern. The EU Chips Act, adopted in 2023, is designed to mobilize around €43 billion in public and private investment in Europe’s semiconductor ecosystem, explicitly framed as strengthening security of supply, resilience, and technological sovereignty—essentially a security‑driven infrastructure/industrial policy program. (consilium.europa.eu) Japan has committed massive subsidies for semiconductor plants (e.g., over 1 trillion yen in support for TSMC fabs and a broader ~10‑trillion‑yen chip and AI plan) to rebuild its chip industry and secure supply chains amid geopolitical tensions, again explicitly tied to economic security. (reuters.com) These show that not only the U.S., but also other major economies, have indeed pursued sustained, large‑scale infrastructure/industrial spending in 2023 and beyond, closely aligned with stimulus, resilience, and national‑security objectives.

Taken together, this matches Friedberg’s prediction that “we continue to have significant infrastructure spending from both the stimulus and security point of view” in 2023 and thereafter, so the prediction is best judged as right.

Data for 2023 show that energy equities significantly lagged most other major asset classes and were among the worst performers:

• The S&P 500 Energy sector returned ‑4.8% in 2023, ranking 10th out of 11 sectors, with only Utilities doing worse (‑10.4%). That clearly makes energy one of the worst‑performing equity sectors.【0search0】
• The broad U.S. energy‑equity ETF XLE had a negative total return of about ‑0.63% in 2023, confirming that U.S. energy equities as a group were roughly flat to slightly down for the year.【0search1】
• By contrast, the overall S&P 500 total return was about +26.3% in 2023, so energy massively underperformed the main U.S. equity benchmark.【4search1】
• The Callan Periodic Table summary for 2023 notes that all of the major asset classes it tracks (U.S. and international stocks, bonds, real estate, etc.) had positive returns, with the worst (cash) still around +5% real—meaning a negative‑return sector like energy equities underperformed essentially all of those broad asset classes.【1search1】
• A mid‑2023 Goldman Sachs–based review likewise cited crude oil and energy stocks as among the worst‑performing assets of the year to that point, reinforcing that energy was a notable laggard versus the broader opportunity set.【3search0】

Taken together, this evidence matches Jason’s prediction that energy equities would underperform most other major asset classes and end up as one of the worst‑performing categories in 2023.

Evidence from 2023 market performance contradicts most parts of this prediction:

• Technology stocks: The S&P 500 Information Technology sector returned +56.4% in 2023, making it the best‑performing S&P sector, not a sector having a “tough year.” (visualcapitalist.com)
• Energy stocks: The S&P 500 Energy sector returned ‑4.8% in 2023, one of only three sectors with negative returns, so energy did relatively underperform in a strong equity year and can fairly be called weak. (visualcapitalist.com)
• High‑yield (junk) debt: The ICE BofA U.S. High Yield Index delivered about +13.5% total return in 2023 (13.46% in one widely cited series), and Barron’s notes high‑yield bonds returned roughly 13.2% in 2023, outpacing core U.S. aggregate bonds and many other fixed‑income segments. This is strong performance, not “the biggest asset class that is going to get pressured.” (morganstanley.com)
• Relative asset‑class stress: Cross‑asset return tables for 2023 show notable underperformers elsewhere—oil, broad commodities, and Chinese equities all had double‑digit negative returns, while broad U.S. bonds were positive—so junk debt was far from the most pressured major asset class. (visualcapitalist.com)

Because tech actually boomed, high‑yield credit performed strongly, and only the energy call was directionally right, the overall prediction is wrong.

Evidence shows that downtown San Francisco’s office market has indeed become highly distressed since 2023, but lenders have not ended up owning “virtually all” or even a large share of the office towers.

• Severe distress, but not universal: San Francisco’s office vacancy rate climbed to around 36–37% in 2024, the highest of any major U.S. market, and office attendance remains far below pre‑COVID levels.(axios.com) CMBS/CRE loan data show that the San Francisco metro has the highest overall CRE distress rate among large U.S. markets (about 22.1% as of September 2024), with office loans the most troubled property type nationwide.(businesswire.com) This validates the “toxic debt / lots of distress” part of the thesis, but that is only half of the prediction.

• How distress has actually resolved: Instead of mass foreclosure with banks taking title, most troubled downtown office assets have been resolved through distressed sales or loan trades to other private owners:

  • The 22‑story tower at 350 California Street sold at roughly a 75% discount to its previous estimated value, to SKS Real Estate Partners and a partner—not to a bank or lender REO.(sfgate.com)
  • In 2024, 23 downtown office buildings changed hands for about $916 million; coverage by CBRE and local press notes that much of this activity involved distressed assets being bought by institutional and wealthy investors, at steep discounts.(sfchronicle.com) Again, these are transfers between private owners, not bank takeovers.
  • The vacant tower at 199 Fremont (300 Howard) was sold in 2025 for about $111 million to DivcoWest and Blackstone, far below pre‑pandemic valuations but still a private‑equity deal, not lender ownership.(sfchronicle.com)
  • At 353 Sacramento Street, New York Life and Lincoln Property bought the distressed loan at a major discount; reporting emphasizes that the building “was never foreclosed on,” illustrating how lenders prefer selling the note to becoming the property owner.(sfstandard.com)
  • Brokers are marketing large office loans such as those on 45 Fremont Street and Market Center (555–575 Market) on behalf of banks, highlighting that the strategy is to dispose of the debt or facilitate a recapitalization, not to foreclose and hold the real estate on bank balance sheets.(ipgsf.com)

• Banks are reducing exposure, not accumulating towers: Industry coverage describes big lenders (e.g., ING, JPMorgan, Goldman Sachs, Capital One) selling office‑backed loans at deep discounts to investors in order to cut their commercial office exposure.(reddit.com) That is the opposite of a scenario where “virtually all of downtown” ends up owned by banks.

• Scale of distress vs. “most towers owned by banks”: Even in the securitized‑loan universe, only about one‑fifth of San Francisco CRE loan balances are classified as distressed; that is substantial but still far from “most” of the downtown office stock.(businesswire.com) Meanwhile, marquee towers like Salesforce Tower, 555 California, 525 Market, and many others remain in the hands of REITs, pension funds, insurance companies, and other private owners—not banks or special servicers as long‑term owners.

• Timeframe: The prediction (made in early 2023) used language like “soon” and “over the next few years.” Nearly three years later, the pattern is clear: widespread distress and big value write‑downs, but lenders overwhelmingly try to avoid ending up as owners, instead pushing note sales, restructurings, or third‑party acquisitions. With transaction volumes rising and some early signs of leasing recovery, the window in which banks might still end up owning “virtually all” or even a large fraction of downtown towers is shrinking rather than expanding.(globest.com)

Because (a) serious distress did materialize but (b) banks/lenders do not own a large share of downtown San Francisco office towers, and many distressed assets have gone to other investors instead, the core ownership outcome described in the prediction has not come true as of late 2025.

Evidence since 2023 shows that U.S. commercial real estate—especially office—entered exactly the kind of “reckoning” Chamath described: falling values, high vacancies, and financing stress.

• Office vacancies hit record highs. Moody’s data show U.S. office vacancy reaching around 19.6% by Q4 2023 and over 20% across the top 50 metros in 2024, an all‑time high, driven by persistent remote and hybrid work. (aa.com.tr) National Association of Realtors’ 2025 insights still describe office vacancy as being at record levels, and other coverage notes that only a few prime submarkets are recovering while most cities remain under severe strain. (nar.realtor) This matches the prediction of “high vacancies” beginning in this period.

• Office values have fallen sharply. CoStar’s repeat‑sale indices, summarized by the St. Louis Fed, show that by early 2024 U.S. office prices were down about 34% from their peak. (stlouisfed.org) CoStar’s later releases report value‑weighted office pricing down roughly 43–44% from its late‑2021 high by Q3 2024 and Q3 2025, confirming a deep, prolonged drawdown in office valuations. (costargroup.com) These are large, structural declines rather than a mild soft patch.

• Financing and banking stress tied to CRE, particularly office. Reports in 2024–2025 describe a sharp drop in U.S. commercial real estate investment volumes (down more than 50% in 2023, the lowest since 2012), with vacant offices explicitly cited as driving falling values and growing concern for banks holding CRE loans. (aa.com.tr) S&P cut its outlook on several regional banks primarily because of heavy CRE exposure, flagging office loans as especially risky due to falling values and higher vacancies. (investopedia.com) A Reuters analysis in 2025 finds office loans as the main weak spot in regional bank portfolios, with office‑loan delinquencies at a record level (around 11.8%) and banks actively trimming office exposure—clear signs of financing stress. (reuters.com)

• Industry outlooks label office as the most strained CRE segment. Major forecasters like CBRE and Moody’s characterize the 2023–2025 period as a muted or difficult environment for commercial real estate, with office singled out as bearing "the most strain," facing rising vacancy, falling values, and restricted lending appetite. (cbre.com)

Taken together, by and after 2023 the U.S. commercial real estate market—especially the office sector—has in fact undergone a pronounced downturn with exactly the features Chamath anticipated (declining prices, record vacancies, and financing stress). That makes his prediction right in substance and timing ("the beginning of the reckoning" starting in 2023).

Available data show a clear deterioration in U.S. consumer credit quality over 2023, especially in unsecured products, consistent with Friedberg’s call, even though it did not become a full-blown systemic crisis.

• Credit cards and autos (core of his thesis): The New York Fed reported that in 2023 credit card delinquencies “surged more than 50%,” with serious (90+ day) delinquencies rising from just over 4% of balances at the end of 2022 to 6.4% by Q4 2023—about a 59% jump. Total household debt reached $17.5T and the share of debt 90+ days past due rose from ~1.0% to ~1.4%. Fed researchers emphasized that transitions into delinquency for credit cards and auto loans were above pre‑pandemic levels and signaled “increased financial stress, especially among younger and lower‑income households.”(cnbc.com)(cnbc.com)(idcfp.com)(newyorkfed.org) Separately, auto loans showed similar strain: by late 2023, 7.7% of auto loan balances were 30 days late, the highest since 2010, indicating many borrowers were failing to keep up with payments.(forbes.com) These are sizable, broad‑based deteriorations in consumer credit performance and defaults in exactly the segments he highlighted.

• BNPL: The BNPL sector is opaque, but surveys summarized by the Federal Reserve and others indicate that around one‑fifth of BNPL users had already made at least one late payment in 2023, with that share rising further in 2024, highlighting meaningful repayment problems among users of this form of consumer credit as it expanded.(fool.com) Regulators like the CFPB also found BNPL borrowers were more likely to have delinquencies on traditional credit products, reinforcing the picture of a stressed subset of consumers relying on layered credit.(nortonrosefulbright.com)

• Mortgages and overall system: Where Friedberg slightly over‑shot was in how broad the wall would be. Mortgage performance in 2023 remained exceptionally strong: the MBA reported overall mortgage delinquency at or near the lowest levels in the survey’s history (e.g., 3.37% in Q2 2023, the lowest since 1979), and CoreLogic described mortgage delinquency rates as hovering near record lows, with serious delinquencies around 1.1%—far below post‑2008 crisis norms.(mba.org)(newslink.mba.org)(worldpropertyjournal.com) The New York Fed likewise noted that while delinquencies ticked up, overall delinquency rates in 2023 were still below 2019 levels.(reuters.com) So the “wall” was not system‑wide or mortgage‑driven.

On balance, though, Friedberg’s core prediction—that sometime in 2023 there would be a significant deterioration in U.S. consumer credit performance, with many consumers unable to meet obligations and a noticeable spike in defaults in categories like credit cards (and related unsecured credit)—did occur. The magnitude of the deterioration in credit cards and auto loans, and the Fed’s own characterization of rising delinquencies and financial stress, align well with his call, even if mortgages stayed unusually healthy. Therefore, the prediction is best classified as right.

Evidence from multiple data sources shows that U.S. consumer credit delinquencies/default‑like measures rose noticeably over 2023.

• The New York Fed’s Household Debt and Credit Q2 2023 report shows the share of credit card balances flowing into serious delinquency (90+ days) rose from 3.35% in Q2 2022 to 5.08% in Q2 2023, and auto loans from 1.81% to 2.41%. Overall serious-delinquency transition for all debts rose from 0.84% to 1.16%.(newyorkfed.org)
• The Q3 2023 report shows the same pattern: serious delinquency transitions increased across most debt types, with credit card serious delinquencies rising from 3.69% in Q3 2022 to 5.78% in Q3 2023, and auto from 2.02% to 2.53%.(newyorkfed.org) These are large percentage increases from already-depressed pandemic-era levels.
• Experian’s review of 2023 consumer credit reports that delinquency rates increased in 2023 across 30–59, 60–89, and 90+ day buckets, even while remaining below pre‑pandemic levels.(aol.com)
• By late 2023, financial press coverage noted that delinquency rates on credit card loans had reached their highest level since 2011, reflecting rising borrower distress as balances and interest costs climbed.(ft.com)
• An AP report on Q4 2023 similarly describes record credit‑card debt and climbing household delinquency rates, especially among younger and lower‑income borrowers.(apnews.com)

While definitions of “default” vs. “delinquency” vary, in consumer‑credit reporting these rising serious delinquencies and charge‑offs are exactly the phenomena the prediction referred to. Given the clear, broad-based, and historically large increases over 2023, the claim that U.S. consumer credit defaults would "rise meaningfully" during 2023 was borne out.

Multiple 2023 data points and analyses describe exactly the kind of household “belt‑tightening” and reduced discretionary spending Jason predicted.

• Macro/retail context. S&P Global’s February 2023 outlook said that “belt‑tightening consumers” would weigh on U.S. retail sales, with real retail sales expected to be roughly flat in 2023 as high inflation and weakened demand strained household budgets. (spglobal.com) An October 2023 ACCA analysis likewise warned that “an episode of belt‑tightening by American consumers is emerging as a major threat,” highlighting run‑down excess savings, rising credit‑card balances, and the likelihood that many households would have to cut spending. (abmagazine.accaglobal.com)

• Documented household austerity behaviors. Kearney’s Consumer Institute Q1 2023 briefing found 69% of consumers reducing spend in one category to fund another and explicitly framed 2023 as a “belt‑tightening environment” driven by inflation and trade‑offs. (prnewswire.com) Company and survey data echoed this: Conagra’s CEO reported in July 2023 that Americans were “hunkering down” and buying fewer items as inflation squeezed budgets, (fortune.com) and Numerator’s 2023 SNAP analysis showed lower‑income households buying notably fewer groceries and trading down across categories to cut absolute costs. (globenewswire.com)

Although total consumer spending still grew in 2023, the emergence of a widely discussed trend of consumer belt‑tightening and austerity—especially in discretionary categories—is well documented. That matches Jason’s prediction about a notable austerity trend among consumers during 2023.

The prediction was that in 2023 and into the 2024 cycle, Donald Trump’s influence in the GOP would continue to wane, shown by weaker endorsements and more open defiance from Republican figures. The opposite happened on the key metrics the predictor named.

1. Endorsement effectiveness stayed extremely high. Compilations of Trump’s endorsements show that his Republican primary success rate in 2024 remained in the mid‑90% range (around 96–98%), in line with or even slightly above previous cycles, indicating that a Trump endorsement was still a powerful asset inside GOP primaries rather than a waning force. (grokipedia.com)

2. GOP elites largely rallied to him rather than defying him. After Trump’s first 2023 indictment, Republican leaders and rivals mostly closed ranks around him, with coverage noting that the party remained “firmly in the hold” of Trump and MAGA Republicans rather than peeling away.(washingtonpost.com) Following his 2024 New York felony conviction, top Republicans—including some previous skeptics such as Mitch McConnell and Susan Collins—again rallied to his defense, and House Speaker Mike Johnson publicly attacked the prosecution and trial on Trump’s behalf, underscoring continued deference rather than open rebellion.(politico.com)

3. He consolidated organizational and electoral control of the party. In early 2024, Trump effectively forced out RNC chair Ronna McDaniel and moved to install loyalists like Michael Whatley and Lara Trump in party leadership, described as part of his effort to solidify control over the GOP apparatus.(ft.com) Electorally, he then dominated the 2024 Republican primaries—winning Iowa in a historic landslide for a non‑incumbent, decisively winning New Hampshire, quickly clinching the delegate majority, and becoming the GOP nominee again—demonstrating overwhelming rank‑and‑file support rather than a continuing decline in influence.(en.wikipedia.org)

Taken together, Trump’s endorsement power, the behavior of GOP officials, his control over party machinery, and his sweeping victory in the 2024 primaries all point to a reassertion and consolidation of influence within the Republican Party, not a continued waning. Therefore, the prediction is wrong.

On direction and timing, the prediction is essentially borne out by 2023–2025 developments.

1. Multiple new cell and gene therapies have been approved and brought to market since 2023

• In 2023 the FDA approved a record seven cell and gene therapies, including five gene therapies for rare genetic diseases, according to the Alliance for Regenerative Medicine’s State of the Industry briefing.(biospace.com)
• Elevidys, the first gene therapy for Duchenne muscular dystrophy, received accelerated approval in June 2023 for boys aged 4–5, with a major label expansion in June 2024 to patients 4 and older.(fda.gov)
• Casgevy (exagamglogene autotemcel), a CRISPR-based gene-edited cell therapy, was approved in December 2023 for severe sickle cell disease and in January 2024 for transfusion‑dependent beta thalassemia.(cgtlive.com)
• Lenmeldy (atidarsagene autotemcel), a gene‑edited cell therapy for early‑onset metachromatic leukodystrophy, was approved by the FDA in March 2024.(cgtlive.com)
• Beqvez (fidanacogene elaparvovec), a one‑time AAV gene therapy for hemophilia B, gained FDA approval in April 2024.(investors.com)
• Lifileucel (Amtagvi), the first FDA‑approved tumor‑derived T‑cell (TIL) therapy, was approved for metastatic melanoma in February 2024.(en.wikipedia.org)
• In November 2025 the FDA approved Novartis’s Itvisma, an intrathecal SMA gene therapy that broadens their existing SMA gene therapy franchise.(reuters.com)

These examples, plus the documented record year of approvals in 2023 and expectations of further double‑digit CGT approvals annually, show that many new cell and gene therapies have indeed reached the market starting in 2023.(pharmalive.com)

2. Significant build‑out of manufacturing and delivery infrastructure

• The global cell and gene therapy manufacturing market was about 18.1 billion dollars in 2023 and is projected to grow to roughly 97.3 billion dollars by 2033, reflecting rapid expansion of capacity and specialized infrastructure.(biopharminternational.com)
• Industry analyses report that around 80 percent of cell therapy manufacturers have expanded capacity by adding new facilities, with more than 40 percent of those facilities dedicated to T‑cell manufacturing and most capacity located in North America, indicating large‑scale build‑out.(globenewswire.com)
• Major pharma companies are investing heavily in new CGT plants, such as Johnson & Johnson’s planned investment of over 2 billion dollars in a new biologics and cell therapy manufacturing site in Wilson, North Carolina.(pr.comtex.com)
• The CGT sector attracted 15.2 billion dollars of investment in 2024, up 30 percent from 2023, with thousands of developers and about 2,000 active clinical trials; industry leaders highlight automation, robotics and new distribution models as key to overcoming manufacturing and delivery bottlenecks.(trial.medpath.com)
• On the delivery side, companies like Vertex are explicitly creating networks of authorized treatment centers to administer Casgevy, with an initial set of centers already activated and more being added, reflecting a reconfiguration of care delivery for gene‑edited therapies.(forbes.com)

Some reporting notes that uptake has been slower than scientific progress because existing health systems were not built for these complex one‑time treatments, underscoring that infrastructure is still catching up.(axios.com) But the combination of record numbers of new cell and gene therapy approvals and clear evidence of large, ongoing investments in manufacturing plants, logistics, quality infrastructure, and specialized treatment centers fits the predictor’s claim that these modalities would become more mainstream with more products on the market and a changing infrastructure and delivery system starting from 2023.

Friedberg’s normalized prediction was that by 2023 generative‑AI media would be far enough along that we’d see prominent examples of: (1) fully AI‑written symphonic music, (2) AI‑written, published long‑form fiction, (3) short films using AI‑generated scripts, and (4) early AI‑driven interactive games with dynamically generated content. The year and surrounding period do in fact feature clear examples in each category.

• AI‑written symphonic / orchestral works. Long before 2023, orchestras were already performing AI‑generated symphonic music, e.g. the 2019 Ars Electronica project in Linz where Mahler’s unfinished 10th Symphony was followed by several minutes of “Mahleresque” music written by MuseNet‑based software and played by an orchestra, widely covered by AFP and others. (thelocal.at) In 2023, a college orchestra in Pennsylvania performed “Symphony (After Beethoven),” a live premiere of an AI‑completed version of Beethoven’s 10th Symphony created by David Cope’s Experiments in Musical Intelligence system. (moultrieobserver.com) Also in October 2023, the Munich Symphony Orchestra performed an AI‑composed piece as part of the “Tapestry of Spaces” project, where ChatGPT/GPT‑4 generated melodies from images and musicians performed the resulting score at Serviceplan’s Innovation Day. (roastbrief.us) These show that by 2023, fully or largely AI‑written orchestral works were being performed and written up in mainstream outlets – exactly the sort of “AI‑written symphony” Friedberg had in mind, even though some precede 2023.

• AI‑written, published novel/novella. In 2023, Stephen Marche (under the pseudonym Aidan Marchine) released Death of an Author, a crime novella published by Pushkin Industries and widely promoted as being 95% generated by AI systems (ChatGPT, Sudowrite, Cohere). The New York Times described it as “arguably the first halfway readable A.I. novel,” and coverage in outlets like Wired, New Scientist and others made it a high‑profile example of AI‑written long‑form fiction. (en.wikipedia.org) At the same time, Reuters and Business Insider reported in February 2023 that more than 200 Kindle titles on Amazon already listed ChatGPT as an author or co‑author, illustrating a genuine “boom in AI‑written e‑books.” (reuters.com) So the prediction that we would see AI‑written, commercially published narrative books in 2023 clearly came true.

• Short films based on AI‑generated scripts. By early 2023, filmmakers were already releasing short films explicitly using ChatGPT‑generated scripts. The Catalyst, a 5‑minute 2023 short by Blake Ridder, is described on IMDb as “loosely based on a screenplay written by ChatGPT” and is listed there with its January 17, 2023 release date. (pro.imdb.com) Ridder and related blogs explicitly market it as a short film created from a ChatGPT screenplay. (lifeboat.com) This sits on top of a broader wave of AI‑script experiments (e.g., GPT‑3‑written short Date Night, covered by BuiltIn in 2021), showing that by and through 2023, “short films with AI‑created scripts” were a visible reality rather than a speculative future. (builtin.com)

• Early AI‑driven interactive video games with dynamic content. Generative‑AI games arrived even earlier than Friedberg suggested. AI Dungeon, first released in 2019, is a text adventure where stories and events are generated on the fly by large language models in response to the player’s free‑form input; the game reached over a million players and is widely cited as an “endless world” of AI‑generated narrative. (en.wikipedia.org) Academic and design work during 2023 further explored generative AI as a dynamic storytelling companion in games like Dungeons & Dragons, explicitly envisioning generative models that create unfolding narratives and character dialogues based on player actions. (arxiv.org) These are precisely the “early AI‑based interactive video games… where content is dynamically generated for the user” that he was referring to, even if the earliest high‑profile examples predated 2023.

Taken together:

• Orchestras performing AI‑generated symphonic music;
• A widely covered AI‑generated novella (Death of an Author) plus a rapid rise of ChatGPT‑written books on Amazon;
• Short films publicly marketed as based on ChatGPT or other AI‑written scripts; and
• Existing, popular AI‑driven games like AI Dungeon and growing research prototypes on generative game content

all existed and were prominent by and during 2023. Friedberg’s colorful phrasing about the “first” such works is historically off (many of these experiments began years earlier), but the normalized claim—that 2023 would be a year when generative‑AI media reached the point of visible, cross‑medium examples of AI‑written symphonies, novels, scripts, and dynamic games—is directionally accurate. Hence, the prediction is best scored as right overall.

Evidence by late 2025 shows that large generative‑AI providers have indeed faced exactly the kind of large‑scale, potentially existential copyright litigation Jason predicted.

Most clearly, Anthropic (developer of Claude) faced a certified US class action over using pirated books to train its models. Judge Alsup certified a class potentially covering up to 7 million works, with statutory damages that legal analysts estimated could reach hundreds of billions to over $1 trillion—described as an existential threat that could financially ruin Anthropic or even the broader AI industry if fully imposed.(arstechnica.com) Anthropic itself told the court it was under “inordinate pressure” to settle in order to avoid a potentially business‑ending trial; it then agreed to a proposed settlement of about $1.5 billion plus destruction of disputed training data—a huge payout, but far smaller than the theoretical maximum.(wired.com) This is exactly the scenario of being “sued into oblivion” in the sense of facing litigation whose potential damages could have wiped out the company, even though it likely survives via settlement.

OpenAI itself has also been hit with multiple major copyright and related suits over its training data and outputs: class actions by authors, a landmark lawsuit from The New York Times, and claims by other rightsholders, all challenging the legality of using copyrighted material for training GPT models.(en.wikipedia.org) Additional cases include a UK copyright/database‑rights lawsuit by Mumsnet over scraped forum content, and a German ruling that ChatGPT’s training on song lyrics violated national copyright law, resulting in damages.(thetimes.co.uk) Other major generative‑AI players like Midjourney and Stability AI are likewise embroiled in large copyright suits from artists and major film studios.(en.wikipedia.org)

While none of these companies has yet been literally destroyed, the prediction—as normalized to “major, potentially existential litigation over copyrighted training data on a scale that could severely threaten the business”—has clearly materialized. Multiple leading providers have been, and still are, defending against lawsuits whose claimed damages and legal theories plausibly threatened their core business models. On that basis, Jason’s prediction is best scored as right.