Hurricane fallout, AlphaFold, Google breakup, Trump surge, VC giveback, TikTok survey

Evidence as of late 2025 shows that the 2020 IMO sulfur cap has significantly increased radiative forcing over the oceans and likely accelerated recent warming, but it is too early and too uncertain to say that the rate of ocean warming in the 2020s (and beyond) is roughly double its pre‑2020 value.

Key points:

• The 2020 IMO regulation cut the sulfur content of marine fuel from 3.5% to 0.5%, reducing shipping-related SO₂ emissions by ~60–80%, and thus reducing reflective sulfate aerosols over the oceans. This removes a cooling mask and adds positive radiative forcing.⁽¹⁾⁽²⁾
• A high-profile study by Yuan et al. (2024) estimates a radiative forcing of about +0.2 ± 0.11 W/m² averaged over the global ocean from the shipping sulfur cut, and explicitly states that this could lead to a doubling (or more) of the warming rate in the 2020s compared with the rate since 1980, with strong spatial variability. NASA’s own summary repeats that this forcing “could lead to a doubling of the warming rate in the 2020s.”⁽³⁾⁽⁴⁾ This is strongly consistent with Friedberg’s mechanism and order of magnitude, but still framed as a projection.
• Other recent modeling studies find much smaller effects. Yoshioka et al. (2024) estimate a global aerosol effective radiative forcing from IMO 2020 of only ~0.13 W/m², yielding ~0.04–0.05 °C of additional global warming over several decades—equivalent to just a few extra years of greenhouse‑gas warming, not a sustained doubling of the long‑term rate.⁽⁵⁾ Jordan et al. (2024) similarly conclude that IMO 2020 accelerates global warming by roughly 2–3 years, not by a factor of two in trend.⁽⁶⁾ A multi‑model intercomparison (Partanen et al. 2024) also reports relatively weak forcing, comparable to only a couple years of recent CO₂ increases.⁽⁷⁾ These results conflict with the much larger effect inferred by Yuan et al.
• Observationally, global ocean heat content has set new records in 2020–2024 and shows an ongoing multi‑decadal acceleration, with 2024 again the hottest ocean year on record.⁽⁸⁾ However, disentangling how much of the change in the rate of ocean warming since 2020 is due specifically to shipping aerosols versus greenhouse gases, El Niño, and other factors is still under active debate. Some prominent scientists (e.g., Michael Mann) argue that ocean heat content is increasing steadily but not accelerating as dramatically as claimed by studies like Hansen et al., underscoring current scientific disagreement on the magnitude of any recent rate jump.⁽⁹⁾
• Statistically, robustly establishing that a decadal warming rate has doubled requires analyzing the full decade with enough signal-to-noise to overcome internal climate variability. As of November 2025 we are only halfway through the 2020s, and the prediction also extends “and going forward,” which refers to time periods that have not yet occurred.

Because:

1. the main paper that directly supports a potential “doubling” frames it as could lead to, not as an already-demonstrated fact,⁽³⁾⁽⁴⁾
2. other state-of-the-art studies find a significantly smaller acceleration,⁽⁵⁾⁽⁶⁾⁽⁷⁾ and
3. only half the decade has elapsed, making trend detection and attribution inherently premature,

the claim that the rate of ocean warming in the 2020s and thereafter will be roughly twice the pre‑IMO‑2020 rate cannot yet be judged as clearly right or wrong. It remains scientifically inconclusive (too early).

Sources

1. IMO sulfur reductions and post‑2020 average fuel sulfur content changes.⁽²⁾
2. Yoshioka et al. 2024 on ~80% reduction in shipping sulfur emissions.⁽²⁾
3. Yuan et al. 2024, Communications Earth & Environment (NASA NTRS abstract): +0.2 ± 0.11 W/m² over global ocean, potentially doubling 2020s warming rate vs. since 1980.⁽³⁾
4. NASA 2024 Year-in-Review summary restating Yuan et al.’s “could lead to a doubling of the warming rate in the 2020s.”⁽⁴⁾
5. Yoshioka et al. 2024, Atmospheric Chemistry and Physics: ERF ≈ 0.13 W/m², ~0.04–0.05 °C extra warming over coming decades.⁽⁵⁾
6. Jordan et al. 2024, Earth’s Future: IMO 2020 regulations accelerate global warming by ~2–3 years (~0.05 °C between 2020 and 2029).⁽⁶⁾
7. Partanen et al. 2024, Atmospheric Chemistry and Physics: multi‑model ERF from IMO 2020 comparable to a few years of recent CO₂ increase.⁽⁷⁾
8. Recent ocean heat content syntheses showing record OHC in 2020–2024 and ongoing acceleration.⁽⁸⁾
9. Coverage of Hansen’s higher-acceleration claims and Mann’s critique that ocean heat content is not accelerating as sharply as suggested.⁽⁹⁾

Citations: (1)–(9) correspond respectively to: 2, 2, 3, 3, 2, 2, 2, 1, 3 in the tool output: ⁽¹⁾[turn2search5] ⁽²⁾[turn2search0] ⁽³⁾[turn3search0]/[turn2search3] ⁽⁴⁾[turn3search1] ⁽⁵⁾[turn2search0] ⁽⁶⁾[turn2search8] ⁽⁷⁾[turn2search7] ⁽⁸⁾[turn1search12] ⁽⁹⁾[turn3search4]

This prediction is about multi‑year to multi‑decadal changes in the frequency of rare storms (events with historical return periods of 1-in-100 or 1-in-500 years) driven by persistently warmer oceans. The podcast aired on 11 October 2024, and we only have about one year of subsequent data (through 30 November 2025), which is far too short to robustly evaluate a statement about how often such rare events will occur on average.

A few relevant points from the science:

• U.S. data do show a long‑term increase in heavy precipitation and extreme one‑day rainfall events over the past century, with especially strong increases since the late 20th century. This is physically linked to warmer oceans and a more moisture‑laden atmosphere, and is expected to continue. (epa.gov)
• The IPCC and U.S. National Climate Assessment project that rare heavy‑precipitation events become more frequent as global warming progresses, but the quantified increases are generally factors of ~1.3–2 (or a few times more frequent) for many heavy‑rain extremes over the 21st century, not an immediate shift to nationwide 1‑in‑100‑year storms happening every 2–3 years. (en.wikipedia.org)
• The Asheville flooding from Hurricane Helene in September 2024 has been characterized as an event with an annual recurrence interval greater than 1,000 years in parts of the southern Appalachians—i.e., extremely rare in the historical statistics. (content-drupal.climate.gov) One such extreme event soon after the prediction is directionally consistent with concerns about more severe storms, but a single event doesn’t establish a new long‑term average frequency.
• A 2023 analysis cited by the Washington Post found that about 20% of the U.S. has already seen roughly a fourfold increase in the likelihood of 100‑year rainfall (from ~1% to ~4% yearly probability), which would correspond to something like a 1‑in‑25‑year event on the old scale—substantial, but still not at the level of “every 2–3 years” nationally, and it is regional rather than U.S.-wide. (washingtonpost.com)
• By definition, estimating or re‑estimating 100‑year or 500‑year return periods requires many decades of observations and/or detailed modeling, and NOAA guidance emphasizes that these probabilities are statistical constructs evaluated over long periods, not something that can be judged from one or two years of weather. (nwrfc.noaa.gov)

Taken together: current science and observations are consistent with the direction of Friedberg’s concern (warmer oceans → more intense and more frequent extreme precipitation and some record‑breaking events). However, his specific quantitative claim—that across the United States, 1‑in‑100‑year storms will occur roughly every 2–3 years and 1‑in‑500‑year events every few years—cannot be confirmed or falsified from only one additional year of data, and existing studies do not yet document such an extreme nationwide frequency shift. Therefore, it is too early to say whether this prediction is right or wrong.

Chamath’s prediction was that, as a result of the DOJ’s antitrust actions against Google, there would be a structural remedy (e.g., divestiture or breakup of business units) rather than Google remaining intact with only behavioral remedies.

What actually happened in the key DOJ cases as of 30 November 2025:

1. Search monopoly case (United States v. Google LLC, 2020 case)

   • After finding Google had illegally maintained a monopoly in search in August 2024, Judge Amit Mehta held a remedies trial in 2025. (en.wikipedia.org)
   • In his September 2025 remedies ruling, Mehta explicitly rejected DOJ’s request to force Google to divest Chrome or break up Android, calling such forced divestiture an overreach. (theguardian.com)
   • Instead, the court imposed behavioral / conduct remedies: banning Google from exclusive search-distribution deals for Search, Chrome, Gemini, and Google Assistant; prohibiting certain tying and pre‑load conditions; and requiring Google to share portions of its search index and related data with competitors. Google remains intact as a single company owning Chrome, Android, and its core businesses. (androidheadlines.com)
   • Commentators widely characterized the outcome as a relatively lenient, conduct-focused remedy rather than a breakup. (theguardian.com)

2. Ad‑tech monopoly case (United States v. Google LLC, 2023 case)

   • Judge Leonie Brinkema ruled in April 2025 that Google illegally monopolized parts of the ad‑tech stack (publisher ad server and ad exchange). (en.wikipedia.org)
   • The DOJ is seeking structural remedies (divestiture of AdX and possibly parts of Google Ad Manager), but as of late 2025 no remedies order has been issued. Hearings on remedies concluded in fall 2025, and Brinkema has signaled reluctance to order a breakup, showing preference for behavioral and interoperability remedies instead. (grokipedia.com)

Across the DOJ’s major Google cases, no structural breakup or forced divestiture has actually been imposed to date. The one case where remedies have been decided (search) resulted in Google remaining intact with conduct-based and data‑sharing obligations, not the kind of structural remedy Chamath predicted. Therefore, given the current record, his prediction is wrong.

As of November 30, 2025, it is too early to determine whether Chamath’s five‑year prediction (deadline October 11, 2029) will come true.

Key facts:

• The U.S. Department of Justice (DOJ) and state attorneys general have active antitrust cases against Google, notably over search and advertising practices. In the search case (filed 2020), a federal judge ruled in August 2024 that Google illegally maintained a monopoly in search, and the case moved into a remedies phase where potential solutions — including behavioral and structural remedies — are being considered, but no final remedy order has yet been issued.
• Reporting on the remedies phase consistently describes that possible outcomes could include changes to default search agreements or aspects of Google’s ad tech stack, but as of late 2025 there has been no court‑ordered breakup or mandated divestiture of Chrome, Android, or major ad businesses, nor any other confirmed “forced structural remedy.” (This is based on current coverage of the DOJ v. Google proceedings and follow‑up reporting on the remedies phase.)

Because the prediction explicitly allows the full window through October 11, 2029, and we are only in 2025 with remedies still under consideration and litigation ongoing, the correct classification today is:

• Result: inconclusive – The prediction could still become true or false; the outcome of the DOJ’s remedies process and any appeals is not yet known, and the deadline has not arrived.

The prediction was conditional on a Republican—specifically Donald Trump—winning in 2024 and taking office in January 2025. That condition was met: Trump defeated Kamala Harris in the November 5, 2024 election and was inaugurated as the 47th president on January 20, 2025.(en.wikipedia.org)

The substantive claim was that, under that Republican administration in 2025, it would become easier to get M&A done than under the prior Democratic administration.

Evidence shows a meaningful easing of merger enforcement and greater receptivity to large deals in 2025 relative to the Biden years:

• While the FTC and DOJ initially kept the Biden‑era 2023 Merger Guidelines in place in February 2025—contrary to expectations of an immediate rules rollback(mayerbrown.com)—subsequent practice shifted.
• A July 2025 analysis notes that new Trump‑appointed antitrust leaders explicitly rejected the Biden “anti‑merger” agenda and began rolling back tactics such as routine “prior approval” requirements and other policies that had discouraged many transactions, characterizing these as a departure from the previous administration’s merger policies.(forbes.com)
• A Wall Street Journal report in late November 2025 describes corporate dealmaking as “bigger and bolder under Trump,” with U.S. deal value up over 40% year‑over‑year to $1.9 trillion, twice as many deals over $10 billion, and only three mergers challenged that year versus an average of about six annually under Biden—a roughly 50% reduction in challenges. The article attributes this shift to more lenient antitrust enforcement and a willingness to clear or settle large transactions, including in media and tech.(wsj.com)
• Specific cases reinforce that high‑profile deals became easier to clear: DOJ quickly ended its probe into T‑Mobile’s $4.4 billion acquisition of UScellular, a move widely read as a sign of decreased scrutiny under Trump, and T‑Mobile is pursuing additional acquisitions.(businessinsider.com) Likewise, Trump reversed earlier opposition to Nippon Steel’s takeover of U.S. Steel by negotiating political conditions (e.g., a golden share and governance guarantees) and allowing the acquisition—one that had faced significant resistance under Biden—to close in June 2025.(en.wikipedia.org)
• Trump also fired the remaining Democratic FTC commissioners in March 2025, leaving the agency in full Republican control—an institutional change expected to soften the stance on major mergers, particularly in tech and other concentrated industries that had been focal points for Biden‑era enforcement.(washingtonpost.com)

At the same time, enforcement did not disappear: antitrust suits against Big Tech (e.g., the Meta monopolization case) and a continued focus on “pocketbook” sectors like housing, healthcare, and groceries persisted, and several law‑firm and practitioner memos describe Trump’s second‑term approach as still “robust” but more case‑by‑case, with the stricter guidelines retained for now.(wilmerhale.com) This nuance, however, is fully compatible with the forecast: the claim was not that antitrust enforcement would vanish, only that M&A would be easier to get done than under the Biden administration.

Given:

• The realized condition (Trump taking office in January 2025), and
• Multiple independent indicators of materially more permissive merger enforcement and a surge in large‑scale M&A activity relative to Biden’s term, including fewer challenges, rollback of prior‑approval requirements, faster clearances, and high‑profile deals that likely would have faced more resistance previously,

the prediction that “there will be an opening up of M&A” and that it would be easier to get M&A done in 2025 under a Republican administration than under the preceding Democratic one is best judged as right.

Available evidence from the 2024 U.S. election cycle supports Chamath’s framing that state-backed social‑media influence campaigns caused information chaos but did not clearly swing the election outcome.

Multiple foreign governments, notably China and Russia, ran coordinated online influence operations using fake personas, networks like China’s “Spamouflage,” and covertly funded influencers to push divisive narratives and undermine confidence in U.S. institutions across platforms such as X and TikTok. These efforts aimed to stir domestic divisions, promote conspiracy theories, and erode trust in elections rather than overtly support a single candidate in China’s case, while Russia focused on boosting Trump and damaging Democrats via paid American influencers and disinformation channels. (en.wikipedia.org)

However, U.S. cyber‑ and election‑security officials, as well as independent researchers, reported no evidence that foreign operations or hacking materially affected vote tallies or the overall outcome of the 2024 presidential election. CISA director Jen Easterly stated there was “no evidence of activity that has the potential to materially impact the outcome of the presidential election,” and academic work similarly notes no evidence that foreign hackers have successfully altered vote counts, while think‑tank analysis finds AI‑enabled foreign influence campaigns had limited engagement and did not transform their effectiveness. (justice-integrity.org)

At the same time, the U.S. response to TikTok—culminating in the Protecting Americans from Foreign Adversary Controlled Applications Act and a de jure nationwide TikTok ban based on concerns over propaganda, misinformation, and foreign influence operations—illustrates how these platforms became focal points for public mistrust, polarization, and political conflict. (en.wikipedia.org) Taken together, the record shows: (1) state‑linked use of influencers and algorithmic amplification on platforms like TikTok contributed to significant social and informational chaos, and (2) there is no credible evidence that these efforts were sufficient to swing the U.S. election result—matching Chamath’s prediction that such campaigns “probably wouldn’t swing an election, but … would cause chaos.”