E252

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

2025-11-22 spoken.md · speaker-labeled ▶ watch ← E251 all episodes E253 →

1
ideas born
4
ideas moved
7
captures · 3 voices
0
dissenting
+69.2
conviction added
-78.9
decay · 108 silent

Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 108 ideas nobody mentioned gave up this week; it applies only when an episode is processed.

Tier crossings

conviction thresholds crossed by this episode — 65 / 45 / 15 · ideas born here show where they landed. why 65 isn't always green

▲ +47.8 🤖 Nvidia risk: special-purpose silicon + Huawei born at watch 47.8

Kill dates that landed since E251

1 hit · 2 partial · 0 miss — windows that closed after 2025-11-14 and up to 2025-11-22, auto-scored against price data and never hand-set. verdict · R · α

ideaverdictRαclosed
📈 Deregulation re-rates the regulated old economy, not tech PARTIAL +8.9% -6.8 2025-11-16
🏛️ Pharma ad ban plus FCC licence pressure guts broadcast TV HIT +90.7% +75.1 2025-11-16
📈 San Francisco's governance flip turns the city around PARTIAL +12.6% -3.0 2025-11-16

Who moved the board

each voice's force on conviction this episodesupports and opposes, weighted exactly as the replay applied them · share = % of this episode's movement

Friedberg
Friedberg
2 captures · 59% of movement · 1 idea born
+41.0 → net +41.0
Chamath
Chamath
4 captures · 35% of movement
+24.1 → net +24.1
Jason
Jason
1 capture · 6% of movement
+4.1 → net +4.1

What got argued (4 ideas)

ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode

NEW NVDA 🤖 Nvidia risk: special-purpose silicon + Huawei dormant 10 CONTESTED ▲ +47.8 0.0 → 47.8

Inference fragments into special-purpose chips (TPU, Groq, Inferentia, Meta silicon) and Huawei — using undisclosed Chinese lithography advances — starts rivaling Nvidia in select markets: announcements in 2026, impact by 2027. Nvidia's general-purpose dominance erodes at the decode layer.

plays NVDA ·primary evals 2027-11-22
Chamath
Chamath support ×1 sentiment ▶ 36:01
There's going to be a highly fragmented layer of decoding chips that exist in the marketplace. GROK is one, TPU is one, Microsoft has a spin, Amazon has Inferentia... we're going to get to disaggregated decode pretty quickly.
Friedberg
Friedberg support ×2 explicit_prediction 24mo horizon ▶ 40:46
There's going to be differentiated chips for different markets and different applications. And that's where there's a risk to Nvidia. ... my early prediction for 2026 is Huawei, where I think that there's lithography technology that exists in China that is not publicly discussed,
GOOGL 🤖 Google's AI comeback pays off in 2025 closed 77 CONTESTED ▲ +8.1 70.1 → 78.3
Chamath
Chamath support ×2 explicit_prediction ▶ 36:01
TPU is an incredible product... This latest spin is very profound... they went from like 8% share to like 16% share of the entire chat market as of like this last month... Google has done an absolutely incredible job in defending search.
Jason
Jason support ×2 explicit_prediction ▶ 38:04
I think the search franchise is going to grow and that Google is not going to lose to ChatGPT... the short in all of this, if you were going to put on the pair trade, is short OpenAI... And I would be long Google, Grok and Anthropic.
SPY 🌍 Risk-off window: November to February closed 41 ▲ +8.1 49.3 → 57.4
Chamath
Chamath support ×2 sentiment ▶ 23:35
My gosh, this short Bitcoin thing has been a bonanza... Is it below 90? It is, right? It's like 87?... Watch out below.
NVDA 🤖 AI compute capex supercycle has 12-24 months of runway green 87 ▲ +5.1 79.8 → 84.9
Friedberg
Friedberg support ×2 sentiment ▶ 25:47
is incorrect. There is value creation because they are generating revenue from those chips this year, six years later. ... Burry is incorrect in thinking that they're hiding anything because it's all there,
Chamath
Chamath support ×1 sentiment ▶ 29:07
Google and Facebook and Microsoft and OpenAI and X are not going to be in the business of generating negative revenue output tokens just for the sake of it... they are already keenly aware of the value of these output tokens. They know the revenue it's generating.

Episode digest

written during extraction and stored in data/extractions/ep252.json — the auditable source of truth, including everything market-adjacent that did not earn a capture

Live from the Venetian (Vegas F1 week), Sacks absent. Nvidia's blowout quarter (+62% y/y) frames round two of the Burry fight: Friedberg's Accounting Corner dismantles the depreciation-fraud claim on textbook GAAP grounds and Chamath adds the output-token-economics rebuttal — both attached to the capex supercycle. Gemini 3 ships and retakes the benchmarks: Chamath cites the 8%→16% chat-share jump and TPU's profundity; Jason commits to an explicit pair trade — long Google/Grok/Anthropic, short OpenAI (private-side short not tradeable, GOOGL long is). Friedberg plants his 2026 flag: special-purpose silicon plus a Huawei lithography black swan is the real risk to Nvidia (new idea nvidia-silicon-fragmentation, impact by 2027). Chamath gloats on the Bitcoin breakdown ('below 90... watch out below') — reinforcing his risk-off-into-February call. Tether dinner with Paolo Ardoino: 500M users, +30M/quarter, 95%+ margins, 'incredible business' — private, no instrument; Friedberg's counterpoint (margins like that invite competition; rate cuts are headwinds) noted but not captured. Epstein-files release (427-1) and the Alan Keating poker segment are digest-only. Labels: clean — guest Alan Keating properly labeled from his 15:38 cameo onward.