Trump AI Speech & Action Plan, DC Summit Recap, Hot GDP Print, Trade Deals, Altman Warns No Privacy

The prediction explicitly concerns "years to come" after the August 2025 speech—that is, it is about the dominant framing of AI policy over a multi‑year horizon. As of today (November 30, 2025), only about four months have passed since August 1, 2025. That is far too short a period to assess whether this framing will remain dominant for several years. Even if the “AI race”/“space race” framing is currently influential, we cannot yet determine whether it will continue to be the primary lens for U.S. and global AI policy debates over the forecasted time span. Therefore, the accuracy of this prediction is too early to judge.

As of November 30, 2025, it is too early to determine whether the entire remaining term through January 2029 will be primarily defined by AI, crypto, immigration, and tariffs.

Evidence so far does show these four areas are central early priorities:

• AI: Trump signed Executive Order 14179, “Removing Barriers to American Leadership in Artificial Intelligence,” on January 23, 2025, revoking prior AI rules and ordering an AI Action Plan to keep U.S. leadership in AI.(en.wikipedia.org) Subsequent speeches and a detailed AI Action Plan, along with draft and proposed orders to preempt state AI regulation, show AI is a marquee policy area.(reuters.com)
• Crypto: The administration created a Presidential Working Group on Digital Asset Markets and floated a U.S. cryptocurrency reserve (Bitcoin, Ethereum, Solana, XRP, Cardano, etc.) as part of a push to make the U.S. the “Crypto Capital of the World.”(en.wikipedia.org) This places crypto clearly among the signature economic-tech initiatives.
• Immigration: Trump’s first‑day Executive Order 14159, “Protecting the American People Against Invasion,” aggressively tightened immigration enforcement, including expanded expedited removal and penalties for undocumented immigrants.(en.wikipedia.org) Congress also passed the One Big Beautiful Bill Act, massively increasing funding for border security, detention capacity, and deportations through 2029, underscoring immigration as a core agenda item.(en.wikipedia.org)
• Tariffs / Trade: The administration’s second‑term trade framework, the Mar‑a‑Lago Accord, centers on broad tariffs, currency policy, and reworking trade relationships to shrink U.S. trade deficits, with tariffs explicitly a key tool.(en.wikipedia.org) New and proposed tariffs are already affecting sectors like AI hardware and data‑center construction.(foreignpolicy.com)

However, the prediction is about how the whole term “is going to be earmarked” and what will be most important “for the next three and a half years.” Only about ten months of that period have elapsed by November 30, 2025. Other unforeseen issues (e.g., foreign crises, economic shocks, domestic unrest) could still emerge and overshadow these priorities. Because the full 3.5‑year horizon has not yet passed, the prediction cannot be definitively classified as right or wrong, even though early evidence is consistent with it.

Therefore the status is inconclusive (too early to fully evaluate).

By the time of (and even before) the August 1, 2025 episode, there were already multiple AI systems explicitly tuned to particular religious beliefs and available to the public:

• Hindu-specific: GitaGPT is a GPT‑3–based chatbot trained on the Bhagavad Gita to answer life questions using its verses, described as a ChatGPT‑like platform for “Gita ka gyaan” (wisdom from the Gita).(timesofindia.indiatimes.com)
• Islam-specific: A range of bots such as QuranGPT, HadithGPT, and Islam & AI focus on answering questions from the Quran and/or Hadith; a 2023 feature explicitly refers to such systems as “religious AI chatbots,” noting they are trained on Islamic scriptures and used by tens of thousands of users across many countries.(thenationalnews.com)
• Christian-specific: Multiple BibleGPT products exist, including a web service positioning itself as an “AI Bible companion” that provides scripture‑rooted answers, and an iOS/Android app BibleGPT: Study Bible with AI that markets “AI-powered answers to your Biblical and Christian questions,” offered with free and paid tiers on mainstream app stores.(biblegpt-la.com)
• Catholic-specific: Magisterium AI, created by Longbeard, is explicitly branded as “Catholic AI,” built on a curated database of about 27,000 Catholic documents and marketed as a tool to answer questions on Catholic doctrine; it has tens of thousands of monthly users and is integrated into the commercial Hallow prayer app.(washingtonpost.com)

These systems are all commercially or publicly available AI tools, intentionally trained or constrained to specific religious corpora and faith traditions, matching the idea of “religious AI” tuned to users’ beliefs. Because such products clearly existed (and have grown more numerous) by late 2025, the prediction that we would “see AI that’s tuned to people’s religious beliefs” has, in substance, come true.

As of November 30, 2025, we are before the prediction’s deadline of August 2027. The claim was that the major copyright/AI‑training lawsuits pending as of August 2025 (e.g., against OpenAI and similar companies) would be settled or otherwise finally resolved within 1–2 years, i.e., by August 2027.

Even if some individual cases have progressed or settled already, the prediction is explicitly about what will have happened by that future date, not by today. Since that date has not yet arrived, we cannot know whether all or even most of the “major lawsuits” will ultimately fit his description.

Because the time window for the prediction has not expired, its accuracy cannot yet be evaluated. Therefore the correct status is inconclusive (too early).

The prediction hinged on the idea that the New York Times–Amazon AI licensing deal in mid‑2025—reported at $20–25M per year—was the peak of this type of training‑data/content‑licensing deal, and that comparable deals signed in future years would see per‑year dollar values “only go down from here.” (wsj.com)

What we can see as of 30 Nov 2025:

1. Size of the NYT–Amazon deal
   Multiple reports (summarizing the original WSJ scoop) say Amazon will pay The New York Times $20–25M annually under a multi‑year AI content‑licensing agreement for training models and powering Alexa and related products. (wsj.com) This is the specific deal Chamath was reacting to.

2. Other large AI–publisher deals already in market
   Before this NYT–Amazon deal, there were already larger AI training‑licensing arrangements on a per‑year basis:

   • News Corp–OpenAI: widely reported at >$250M over 5 years (≈$50M/year). (allmo.ai)
   • Reddit–Google & Reddit–OpenAI: coverage of Reddit’s filings and press reports imply ~$60–70M per year from OpenAI and ~$60M/year from Google for data licensing. (allmo.ai)
   • Dotdash Meredith–OpenAI: estimated fixed component around $16M/year. (allmo.ai)
     So even at the time of the podcast, $20–25M/year was not the highest per‑year figure in the ecosystem, but the user’s normalized prediction focuses on the future trend from that point.

3. Deals signed after Aug 1, 2025
   Since the podcast date, there have been new AI content‑licensing deals, but most have undisclosed financial terms:

   • Perplexity–Gannett / USA TODAY Network (Perplexity Publisher Program) – terms not disclosed. (businesswire.com)
   • Gannett–Microsoft AI licensing agreement via Microsoft’s Publisher Content Marketplace – announced in Gannett’s Q3 2025 results, but again no dollar figures given. (businesswire.com)
   • Perplexity–Getty Images multi‑year visual‑content licensing deal – no public pricing. (futureweek.com)
     Survey pieces that catalog the training‑data market as of October 2025 list the major known deals (News Corp, Reddit, Dotdash Meredith, NYT–Amazon, etc.) and explicitly note that only a few dollar figures are public and most contracts are confidential. (allmo.ai)

4. No post‑August 2025 deal with a clearly higher disclosed per‑year value
   In the public record through November 2025, no newly announced publisher/AI training‑data deal after Aug 1, 2025 has a disclosed annual payment that clearly exceeds the NYT–Amazon $20–25M/year range. However, because most newer deals do not publish financial terms at all, this absence of evidence is not strong evidence that the underlying economics are actually lower.

5. Time horizon and data limitations
   Chamath’s statement was about “from here” going forward and “future years”. Only about four months of dealflow have elapsed since the podcast, and the small number of post‑August deals with confidential pricing makes it impossible to robustly determine a trend in per‑year license values. Industry overviews emphasize that the market is still evolving, with a wide spread from single‑digit millions per year to nine‑figure multi‑year packages, and that public numbers should be treated as directional only. (allmo.ai)

Why the result is "inconclusive" rather than right or wrong:

• To decisively call the prediction wrong, we would need at least one post‑Aug‑2025 training‑data licensing deal, clearly comparable in scope, with a disclosed annual value above ~$25M/year. No such publicly documented example exists yet.
• To call it right, we would need evidence that subsequent comparable deals are consistently priced below that level (or that $20–25M/year is clearly a ceiling). Because most newer contracts do not report financials, and the observation window is short, we also lack that evidence.

Given the short time since the prediction and the opacity of deal terms, the direction of per‑year pricing for future AI training‑licensing deals cannot yet be determined from available data, so the outcome is best classified as inconclusive (too early to tell).

The prediction specifies a horizon of about five years from August 2025, i.e., until roughly August 2030. As of late November 2025, less than one year has elapsed since the prediction date. Today is still in calendar year 2025, with multiple independent sites confirming that 2025 is the current year and we are in Q4 2025. (en.wikipedia.org)

Because the prediction is explicitly about what will happen over the next five years, it is too early to determine whether patents and copyrights will have become practically unenforceable or economically insignificant by 2030. Even though there are active debates and legal cases about AI, copyright, and training data, these are part of an evolving legal landscape whose outcome by 2030 cannot yet be observed or measured in 2025. Therefore, the correct status of this prediction as of now is inconclusive (too early to tell).

The prediction hinges on whether future AI–publisher content-licensing deals, on average, exceed the ~$20M/year level of the cited New York Times deal.

1. Baseline deal (~$20–25M/year)
In late July 2025, reporting revealed that Amazon’s multiyear AI licensing deal with The New York Times is worth $20–25 million per year, allowing Amazon to train its AI models and surface NYT content in products like Alexa. This is the specific deal and order of magnitude the podcast referred to as the starting point.【2news12】【2search3】

2. Pre‑existing context already spans well above and below $20M/year
Before this NYT–Amazon disclosure, major media–AI deals already covered a wide value range, including:

• OpenAI–News Corp: reported $250M+ over five years (≈$50M/year) for access to WSJ, NY Post, and other News Corp outlets.【0search10】【0search9】
• Google–Reddit: about $60M per year for content used in AI and search.【0search3】
• Several other OpenAI publisher deals (Axel Springer, Financial Times, Dotdash Meredith, etc.) where only partial or rough figures (often “tens of millions” total or $5–10M/year ranges) are public.【0search3】【0search10】
  This shows that even before the NYT–Amazon benchmark, deal sizes were not uniformly trending around $20M/year; they ranged both above and below it.

3. Post–Aug 1, 2025 deals rarely disclose comparable dollar amounts
After the podcast (Aug 1, 2025) and up to Nov 30, 2025, there are new AI–publisher arrangements, but their effective annual economics are not disclosed:

• Gannett (USA TODAY Network) – Perplexity: strategic AI content-licensing deal plus ad‑revenue share; no dollar figures given.【5search1】【5search3】【5search8】
• Perplexity’s Comet Plus model: a $42.5M initial revenue pool to be shared across participating publishers, with plans to grow over time—again, a pool and rev‑share, not clear per‑publisher annual license numbers.【5news12】
• People Inc. (formerly Dotdash Meredith) – Microsoft and Gannett – Microsoft: participation in Microsoft’s new Publisher Content Marketplace, a pay‑per‑use / marketplace model rather than a disclosed fixed yearly check.【3news13】【3news14】【3search1】【3search3】【3search8】
  None of these provide transparent, per‑year dollar figures comparable to the NYT–Amazon arrangement.

4. Why the prediction can’t be scored yet
Jason’s claim is about a future trend in average dollar values for similar AI–major‑publisher licensing deals relative to the ~$20M/year NYT benchmark. To test that, we’d need:

• A meaningful set of post‑July‑2025 deals of similar type (AI company ↔ major publisher), and
• Clear, per‑deal financial disclosures so we can compare whether the average annualized value is rising or falling.

As of Nov 30, 2025:

• Only a few months have elapsed since the baseline NYT–Amazon numbers became public.
• Most subsequent AI–publisher deals either don’t disclose dollar values or use rev‑share/marketplace structures whose realized annual payouts aren’t yet known or reported.
• There is no robust data set from which to compute or qualitatively assess an “average” trend above or below $20M/year.

Because of the short time horizon and the lack of sufficient, comparable post‑baseline deal data, we cannot determine whether the average value of such deals is actually trending up or down relative to the ~$20M/year benchmark. The prediction therefore remains too early to evaluate, not clearly right or wrong.

As of November 30, 2025, there is no evidence that any jurisdiction has formally granted a professional license (such as bar admission or a medical license) directly to an AI system or model.

Medicine:

• Some AI systems have passed or achieved high scores on medical licensing exams (e.g., the Xiaoyi robot in China passing the national medical licensing exam; OpenEvidence’s model scoring up to 100% on the USMLE), but these systems are not recognized as licensed physicians. They function as decision-support tools for human doctors, not as independently licensed professionals. (en.wikipedia.org)
• U.S. regulators are explicitly reinforcing that only humans can be licensed clinicians. California’s AB 489, signed into law in October 2025, prohibits AI systems from presenting themselves as licensed health professionals and extends existing title-protection rules to AI developers and deployers. The California Medical Board has emphasized that only a “natural person” may be a licensed physician in the state. (medscape.com)

Law:

• In the U.S., unauthorized practice of law (UPL) statutes in all states restrict legal practice to individuals who are licensed attorneys, and existing case law and guidance frame the practice of law as limited to natural persons who have been admitted to the bar. (americanbar.org)
• While LLMs have passed bar-exam-style benchmarks, and legal AI tools like Harvey or others are widely used, they operate under the supervision of human lawyers rather than holding any bar license themselves. Regulatory reform efforts in states like Utah, Arizona, and Washington expand business models and tech use but still premise actual legal practice on human licensees. (reuters.com)

Emerging proposals, not law:

• The U.S. Healthy Technology Act of 2025 (H.R. 238) would allow AI/ML systems to qualify as practitioners eligible to prescribe drugs under specified conditions, but as of late November 2025 it remains only introduced and referred to committee, with no passage in either chamber. (congress.gov)

Given this:

• The specific event Friedberg predicts — AI models being formally allowed to obtain professional certifications like bar admission or medical licensing, to a degree that a Polymarket contract could cleanly resolve on that fact — has not occurred by November 30, 2025.
• However, his prediction is explicitly open-ended (“there will be a date when this happens”) with no time horizon, so the fact that it has not yet occurred does not make it wrong; it is simply untested so far.

Because the outcome is about an unspecified future date and there is still ample time for such regulatory changes to (or not to) occur, the correct classification today is “inconclusive (too early)”, not right or wrong.

Scope of the prediction
Chamath said that from Q3 2025 onward the U.S. macro environment would look like Q2 2025, specifically: (1) a large surplus (fiscal/trade), (2) strong GDP expansion similar to Q2, and (3) inflation that continues to moderate.

1. “Large surplus” did not persist

• Trade: Official BEA data show the U.S. continued to run large trade deficits, not surpluses. The goods-and-services deficit was $60.2B in June, $78.3B in July, and $59.6B in August 2025, with the year‑to‑date deficit actually larger than in 2024. None of these months show a surplus. (bea.gov)
• Current account: For Q2 2025 itself, the U.S. still had a current‑account deficit of $251.3B (3.3% of GDP); it narrowed versus Q1 but remained firmly negative, not a surplus. (bea.gov)
• Fiscal: For fiscal year 2025 (ended Sept. 30), the federal government ran an annual deficit around $1.8T, only slightly smaller than 2024. July alone saw a $291B deficit, and October posted a $284B deficit, even though some months like April and June had temporary surpluses. The overall environment is one of persistent large deficits, not a sustained “large surplus.” (finance.yahoo.com)
  Given this, the central claim that Q3 and beyond would feature a large surplus similar to (his characterization of) Q2 is clearly false.

2. “Good GDP expansion” is likely but not fully observable yet

• Q2 2025 real GDP grew at about 3.3–3.8% annualized, a strong rebound from the Q1 contraction. (tradingeconomics.com)
• The Commerce Department’s advance Q3 GDP estimate was delayed by the government shutdown and is scheduled for Dec. 23, 2025, so as of Nov. 30 we do not yet have an official Q3 GDP print. (reuters.com)
• High‑frequency tracking (e.g., Atlanta Fed estimates around 4% annualized for Q3) suggests growth probably has remained solid, but that’s still model‑based, not realized data. (reuters.com)
  So this part of the prediction may be directionally right, but it isn’t fully verifiable yet.

3. Inflation has not kept “moderating” through Q3

• CPI: Year‑over‑year CPI inflation in 2025 starts near 3% and dips into the 2.3–2.4% range in March–April, but then rises back to about 3.0–3.1% by September 2025, indicating a re‑acceleration rather than a continued decline in Q3. (inflationtool.com)
• PCE (Fed’s preferred gauge): Headline PCE inflation rises from 2.5% in May to 2.7% in August, and core PCE YoY drifts up from about 2.6% in April to roughly 2.9% by August, with monthly readings of 0.2–0.3%. That is “sticky” or slightly higher inflation, not ongoing moderation. (bea.gov)
  Thus, the claim that inflation would continue to moderate into Q3 and beyond does not match the realized data; it flattened and then ticked up.

Overall assessment
Even allowing that Q3 GDP itself looks likely to be strong, two core pillars of Chamath’s forecast—a sustained large surplus and continued disinflation into Q3+—are clearly contradicted by the data we now have. Since these elements were central to his description of the future macro environment “similar to Q2,” the prediction as stated is wrong rather than merely unproven or ambiguous.

• Was there a qualifying Fed cut?
  – After Q2 2025, the Fed held the funds rate at 4.25%–4.50% in May, June, and July 2025.(investopedia.com)
  – The first cut after that was on September 17, 2025, when the FOMC reduced the target range by 25 bps to 4.00%–4.25%, its first rate cut since December 2024.(uk.finance.yahoo.com)
  This matches the “first Fed interest‑rate cut following this discussion (post–Q2 2025)” condition.

• Immediate market reaction (“the second they cut”):
  – Around the announcement and projections, the Dow jumped as much as ~+1% intraday, the S&P 500 flipped from red to green, and the Russell 2000 (small caps/risk assets) briefly surged over 2% before paring gains.(finance.yahoo.com)
  – By the close that day, the move had partly faded: Dow finished about +0.5–0.6%, S&P 500 –0.1%, Nasdaq –0.3%, while the Russell 2000 held a small gain.(finance.yahoo.com)
  So intraday there was a clear, fast, upside spike—especially in the Dow and small‑cap/risk assets—even though it did not all hold into the close.

• Very short‑term follow‑through:
  – On September 18, 2025 (the next session), four major U.S. equity benchmarks (S&P 500, Dow, Nasdaq 100, Russell 2000) all closed at all‑time highs together, with the Russell 2000 up ~2.5%, described as a surge in risk‑on sentiment “after the Federal Reserve delivered its first interest‑rate cut this year and signaled more ahead.”(bloomberg.com)
  – Pre‑market that morning, futures for the S&P 500 and Nasdaq were already up roughly 0.8%–1.1%, and coverage explicitly linked record‑level opens to the prior day’s Fed cut.(actionnewsjax.com)

• Assessment vs. the prediction:
  Jason’s claim was that “the market will rip the second they cut that,” i.e., that the first cut would trigger a sharp, immediate risk‑asset rally.
  – There was an instantaneous pop right after the cut (Dow ~+1%, Russell 2000 >+2% intraday, S&P flipping positive) and then a broad, record‑setting rally within one trading day directly attributed in coverage to that first cut and the signal of more easing.(apnews.com)
  – While the initial “rip” partially faded by the close of the announcement day and was not uniformly strong across every index at settlement, the overall pattern—sudden upside spike in key indices and risk assets followed by record highs immediately afterward—is consistent with his directional, qualitative prediction.

Given that the first post–Q2 2025 rate cut did coincide with a rapid, noticeable upside move in major equity indices and risk assets in the immediate aftermath and into the next session, the prediction is best judged as right (even if the intraday move was volatile and not all of it stuck into the first day’s close).