Epstein Files Fallout, Nvidia Risks, Burry's Bad Bet, Google's Breakthrough, Tether's Boom

As of November 30, 2025, the Epstein Files Transparency Act has only recently required the Department of Justice to release all unclassified Epstein-related records by around December 19, 2025, so the core “Epstein files” Jason is referring to have not yet been fully released or analyzed. (en.wikipedia.org) Existing official reviews by the DOJ and FBI, completed in mid‑2025, concluded that Epstein died by suicide and reported no evidence of a secret client list or blackmail operation; they did not present or confirm proof that he was an asset of any U.S. or foreign intelligence agency. (theguardian.com) Public discussion in 2024–2025 about possible Mossad, CIA, or other intelligence ties remains speculative and is explicitly denied by senior officials such as former Israeli prime minister Naftali Bennett and by figures like Alan Dershowitz; analyses of these claims emphasize that no concrete documentary or witness evidence has yet substantiated them. (time.com) Because Jason’s prediction is specifically about what will be shown after the Epstein files are fully released and investigated—a process that has not yet occurred—and no authoritative findings currently confirm or definitively rule out intelligence-agency involvement, there is not enough information at this time to judge the prediction as either correct or incorrect.

The prediction’s deadline is “by the end of 2026”, so as of the current date (30 November 2025) there is still over a year left for it to play out.

Public information shows that:

• Huawei already designs and sells Ascend AI accelerators (e.g., 910B/910C) produced by SMIC on a 7 nm process. These chips are used in large clusters, are priced below Nvidia’s top GPUs in China, and are positioned as competitive—though generally behind Nvidia’s latest H100/B200 in absolute performance and efficiency—for certain AI workloads.(tomshardware.com)
• Huawei and SMIC are constrained to older nodes (around 7 nm) because they lack access to EUV lithography; reporting indicates their next Ascend generations are still planned on 7 nm-class tech through at least 2026, with ongoing yield and capacity challenges.(fortune.com)
• Chinese equipment maker SiCarrier has publicly described DUV multi‑patterning techniques that could enable 5 nm‑class production without EUV, and says its tools are used by SMIC and in collaboration with Huawei—but these techniques are openly discussed in industry reporting, not secret or undisclosed.(reuters.com)
• Huawei is reportedly involved in a “secret” network of fabs and proxy firms to expand domestic chip manufacturing under U.S. sanctions, but articles describing this network do not provide evidence of a novel, non‑public lithography technology already deployed in volume production.(bloomberg.com)

So, while parts of the prediction are partially aligned with current reality (Huawei producing domestic AI accelerators at significant scale that are somewhat competitive and cheaper for some markets), the specific claim about a not publicly discussed Chinese lithography technology being deployed in new mainland fabs is not supported by current public evidence, and the time window for the overall prediction has not yet closed.

Because the deadline (end of 2026) lies in the future and there is still ample time for events to change—either toward confirming or refuting the claim—the outcome cannot yet be definitively judged. Hence the result is inconclusive (too early).

As of the current date (November 30, 2025), the prediction concerns a period that has not yet begun: it specifically refers to Huawei’s domestically fabricated high-performance AI chips having a noticeable competitive impact between 2026 and 2027, especially as a lower-cost alternative that becomes a meaningful risk to Nvidia.

While Huawei has started developing and, in some cases, shipping more advanced domestically produced chips (e.g., for smartphones and some AI/5G workloads), the question in this prediction is explicitly about the 2026–2027 impact window and about high‑performance AI accelerators competing with Nvidia at scale. Since we are still in 2025, there is no way to determine yet whether Huawei’s 2026–2027 offerings will:

1. Reach performance levels relevant to the cutting edge of AI training/inference where Nvidia currently dominates; and
2. Achieve broad enough deployment and cost advantage to pose a meaningful new competitive risk to Nvidia during that 2026–2027 window.

Because that time period has not occurred yet, the prediction cannot be evaluated and is therefore too early to call.

It’s too early to evaluate this prediction.

The forecast explicitly concerns what will happen “over the next few years… through at least 2027” to:

1. Google’s core search advertising (overall growth, maintained dominance vs. ChatGPT/OpenAI), and
2. OpenAI’s relative position in the general-purpose AI assistant/model market (whether it loses leadership and share to Google, xAI, and Anthropic by 2027).

As of today (November 30, 2025), only about one week has passed since the podcast release date (November 22, 2025), and we have not yet reached the end of the stated evaluation window (sometime in 2027). Because the claim is about a multi‑year trajectory and end-state, it cannot be definitively judged now, regardless of any short-term signals.

Therefore the outcome is inconclusive (too early to tell).

Available sources confirm that in the audio of the Nov 22, 2025 All‑In episode, Chamath (while talking with Alan Keating about the Polk hand) says they’ll "get Nick to play picture in picture this hand." This line appears in third‑party transcripts of the episode’s audio, but those transcripts only represent what’s said, not what’s shown on screen. (cockatoo.com)

The produced episode’s video is hosted on YouTube (watch URL ending in 4tgV87SM-r0), as referenced by multiple episode listings and summaries, but those pages likewise only describe the topics and segments, not any specific editing choices like picture‑in‑picture overlays. (digitalhabitats.global)

Due to current technical limits I cannot inspect the actual video stream or any frame captures, and a web search does not reveal any reliable viewer reports or notes indicating whether the poker hand was in fact shown via picture‑in‑picture in the released video. (digitalhabitats.global)

Because the episode is already published (so it’s not a timing issue) but there is no accessible, verifiable information about the video’s on‑screen layout, whether the prediction came true cannot be determined from the available evidence. Therefore the outcome is ambiguous rather than clearly right, wrong, or too early to tell.

Available evidence shows the prediction has been announced but not yet verifiably fulfilled as of 30 November 2025.

• In the 22 November 2025 episode “Epstein Files Fallout, Nvidia Risks, Burry's Bad Bet, Google's Breakthrough, Tether's Boom”, Jason explicitly says they are going to have exclusive YouTube content of the besties playing poker with “incredible professional poker players like Jason Koon, Alan Keating, and then I think Phil Hellmuth,” matching the normalized prediction. (cockatoo.com)
• The official All-In site’s description of Episode #252 states that the episode “concludes with a Vegas wrap-up and a tease for exclusive content,” again framing this poker material as future or teased content, not something already released. (allin.com)
• As of 30 November 2025, public indexes and episode/summary aggregators linked from the official site (Tapesearch, Podbay, iHeart, etc.) list recent All-In podcast and YouTube-linked episodes, including poker discussions and mentions of an All-In Poker Tournament, but none describe or link to a standalone All-In Podcast YouTube series or special showing the besties actually playing poker with Jason Koon, Alan Keating, and/or Phil Hellmuth. (tapesearch.com)
• Because direct crawling of YouTube channel pages is restricted, we can’t absolutely rule out that such a video has quietly been uploaded. However, given the prominence of the podcast and of those poker pros, one would reasonably expect some reflection of that release (descriptions, transcripts, blog posts, news blurbs, or episode cross-promotion) outside YouTube; none appears in current search results.

Since (a) the statement is explicitly about future content, (b) only about a week has elapsed since the tease, and (c) there is no independent evidence that the promised exclusive poker videos have been published yet, it is too early to definitively judge the prediction as right or wrong. Therefore the appropriate status is "inconclusive" (too early).