E224

Trump vs Harvard, Nvidia export controls, how DEI killed Hollywood with Tim Dillon

2025-04-19 spoken.md · speaker-labeled ▶ watch ← E223 all episodes E225 →

2
ideas born
7
ideas moved
15
captures · 5 voices
2
dissenting
+159.5
conviction added
-129.4
decay · 122 silent

Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 122 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

▲ +65.3 🏛️ Nvidia's Asia revenue is diverted China demand born at green threshold 65.3 still dormant — green gate not met
▲ +23.0 📈 Higher education's value unwinds, starting with the MBA ember watch 33.2 → 56.3
▲ +6.3 📈 Blue-collar trades boom and a rising premium on human service ember watch 43.6 → 49.9
▲ +31.2 🏛️ Federal research funding retreats from universities born at ember 31.2
▲ +17.2 🏛️ US compute export controls keep China off the AI frontier from 2025 dormant ember 10.0 → 27.2

Kill dates that landed since E223

1 hit · 1 partial · 2 miss — windows that closed after 2025-04-11 and up to 2025-04-19, auto-scored against price data and never hand-set. verdict · R · α

ideaverdictRαclosed
📈 Gamified betting apps take a rising share of the consumer wallet MISS -17.4% -25.0 2025-04-19
🏛️ Section 174 R&D amortization crushes small R&D-heavy companies MISS +2.3% -5.3 2025-04-19
🏛️ Counter-drone defense becomes the new procurement line HIT +82.5% +76.7 2025-04-12
📈 Shrinking global equity float lifts prices regardless of fundamentals PARTIAL +5.8% +0.0 2025-04-12

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

Chamath
Chamath
5 captures · 44% of movement · 1 idea born
+84.0 → net +84.0
Sacks
Sacks
3 captures · 22% of movement
+42.5 → net +42.5
Friedberg
Friedberg
3 captures · 18% of movement · 1 idea born
+24.8 / -10.1 → net +14.8
TI
Tim Dillon guest ×0.5
3 captures · 9% of movement
+12.9 / -4.8 → net +8.1
Jason
Jason
1 capture · 5% of movement
+10.1 → net +10.1

What got argued (7 ideas)

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

NEW NVDA 🏛️ Nvidia's Asia revenue is diverted China demand closed 0 CONTESTED ▲ +65.3 0.0 → 65.3

Chamath's allegation off the H20 ban: roughly 47% of Nvidia's revenue ships to China and China-adjacent countries, much of it to entities stood up in Singapore, Vietnam, Cambodia and Bhutan that act as way-stations into the mainland, while Nvidia has re-engineered a just-compliant part (A800/H800/H20) at every rule change. Sacks confirms legal and illegal evasion of the controls is a fact and says BIS needs more monitoring, inspection and enforcement. If the boundary lines tighten and enforcement lands, a far larger slice of Nvidia's revenue base is at risk than the disclosed $5.5B H20 charge.

plays NVDA ·primary SMH TSM evals 2026-04-19
Chamath
Chamath support ×3 explicit_prediction ▶ 14:20
About 47% of all of Nvidia's revenue goes to China and Chinese related countries. And I think when you peel back this onion, I think what you will find is a whole raft of companies that were stood up to buy these Nvidia GPUs to essentially act as a way station for China.
Sacks
Sacks support ×2 explicit_prediction POLICY-ADJACENT ▶ 18:33
So there have been attempts to set up shell companies to circumvent the export controls. And it is a very big problem.
Jason
Jason support ×2 sentiment ▶ 27:14
Taiwan and Vietnam do not have the need for that many domestically. They're obviously flipping them to someone, right?
NEW TMO 🏛️ Federal research funding retreats from universities closed 5 ▲ +31.2 0.0 → 31.2

The Harvard fight is the leading edge of a structural retreat of federal research dollars from elite universities: $2.2B frozen and $9B under review, NIH capping the indirect-cost rate that ran to 69% at Harvard, Friedberg arguing universities may not be the right venue for fundamental research at all, and Sacks arguing a non-compliant school simply gives up its federal funding. University lab budgets shrink, which hits the suppliers of instruments, reagents and preclinical services that sell into academic core labs.

plays TMO ·primary CRL ILMN evals 2026-04-19
Sacks
Sacks support ×1 sentiment POLICY-ADJACENT ▶ 40:14
No one is saying that they have to get rid of DEI. They just have to give up their federal funding the way that Bob Jones University did.
Friedberg
Friedberg support ×2 sentiment ▶ 48:38
So the question today that's being asked is, should we be doing fundamental scientific research at universities given that over time the administrative overhead has grown and they're basically creating administrative workloads and employing people without necessarily having a market incentive?
LOPE 📈 Higher education's value unwinds, starting with the MBA closed 28 CONTESTED ▲ +23.0 33.2 → 56.3
Chamath
Chamath support ×2 sentiment ▶ 49:55
It is an asset manager that may happen to have some educational things that they do on the side, which increasingly are not what we need it to do.
TI
Tim Dillon support ×2 explicit_prediction guest ×0.5 ▶ 1:00:56
I don't think it's a good idea for these kids to take out hundreds of thousands of dollars worth of loans to, you know, go away for four years, not have a solid plan and not execute and then graduate mired in debt without really a pathway to paying any of it back
Friedberg
Friedberg support ×3 explicit_prediction ▶ 1:03:04
So I do think like as AI kind of takes over education, I do expect that the workplace will change, Tim, and we will start to see more of this integration between education and workplace enabled by this AI-driven kind of, you know, development system, which is going to be radically different than what we have today.
NVDA 🏛️ US compute export controls keep China off the AI frontier from 2025 closed 12 CONTESTED ▲ +17.2 10.0 → 27.2
Sacks
Sacks support ×3 explicit_prediction POLICY-ADJACENT ▶ 7:57
It seems to me that at some point you have to say that some technologies are just too sensitive to be sold to China. And so then the question is, are you drawing the line in the right place?
Friedberg
Friedberg oppose ×2 explicit_prediction ▶ 11:20
So, Jay Cal, I think they're doing it either way. I think that they're going to invest and build their own EUV and chip manufacturing capacity either way.
TI
Tim Dillon oppose ×1 sentiment guest ×0.5 ▶ 13:06
I think we should keep them dependent. Keep selling it to them.
Chamath
Chamath support ×3 explicit_prediction ▶ 14:20
I think that the technology that they need is extremely non-trivial, and I do think that it actually slows them down quite a bit if they don't have access to it.
NVDA 🤖 DeepSeek R1 was distilled from OpenAI, not a clean $6M cost breakthrough closed 8 CONTESTED ▲ +10.4 23.2 → 33.6
Chamath
Chamath support ×2 explicit_prediction ▶ 14:20
At every turn and at every step of AI, they are at the same rate or one step ahead. And I suspect it's because that these ships are being used in very sophisticated ways behind the scenes.
PWR 📈 Blue-collar trades boom and a rising premium on human service closed 37 ▲ +6.3 43.6 → 49.9
TI
Tim Dillon support ×2 sentiment guest ×0.5 ▶ 19:47
you should destigmatize being an electrician, a plumber, a contractor, all these things that when I was told, when I was growing up, they point to a guy doing construction and go, you're going to do that if you don't do your homework. And then the people that did their homework are all bankrupt. And that construction guy is killing it.
KWEB 📈 Chinese tech stocks are 2025's best asset closed 17 CONTESTED ▲ +5.9 53.0 → 58.9
Chamath
Chamath support ×2 sentiment ▶ 45:43
So then you think about, OK, well, what is the implication of this? Well, the implication is obvious. These guys are inventing things. We're playing catch up.

Episode digest

written during extraction and stored in data/extractions/ep224.json — the auditable source of truth, including everything market-adjacent that did not earn a capture

The H20 ban is the whole market story. Sacks defends export controls on the merits with the ASML/H800/H20 history and argues the H20's memory bandwidth made it too good to ship, while Chamath flips the debate from 'do controls work' to 'is Nvidia complying' — a chart showing ~47% of Nvidia revenue going to China and China-adjacent countries, and an allegation that shell entities in Singapore, Vietnam, Cambodia and Bhutan are way-stations into the mainland, which is coined here as a new bearish-NVDA idea. Notably Chamath now says the controls genuinely slow China down, siding with the export-controls thesis he and Laffont/Naval previously fought; Friedberg takes the Gurley/Baker other side (China builds its own EUV either way, citing SMIC's 7nm Mate 60 part and a $37B 3nm program), and Tim Dillon's 'keep them dependent, keep selling it to them' is the same objection as a bit. The Harvard segment reinforced the higher-ed-unwind idea from three directions — Friedberg on AI-personalized learning plus Palantir's skip-college apprenticeship, Tim Dillon on the debt trap, Chamath on 'Harvard Corporation' as an asset manager — and produced a second new idea on federal research funding retreating from universities (NIH's indirect-cost cap against Harvard's 69% overhead, $2.2B frozen, $9B under review) with life-science tools as the read-through. Deliberately not captured: the DEI/Hollywood segment with Tim Dillon is culture-war commentary with no tradeable claim (his 'executives rediscovered the profit motive' line is retrospective, and the two obvious homes the orchestrator suggested, legacy-media-ad-model-terminal-decline-2024 and generative-video-remakes-content-economics-2024, are not in the era registry — closed windows); the Ukraine/martial-law exchange is politics only; Friedberg's mitochondrial-transfer science corner is a genuine new therapeutic modality but at academic-paper stage with no public play, so no idea; Tim Dillon's tariff take was ambivalent (Trump's instincts right, rollout chaotic, must avoid skyrocketing prices) with no directional claim, so no tariff attach; Chamath's China-rare-earth-retaliation aside was a news restatement. Conflicts to note: Sacks is the sitting AI/crypto czar defending a policy he administers (flagged sacks_policy, and his Harvard-funding mention capped at strength 1 as administration advocacy), and Chamath owns Groq, a direct beneficiary of Nvidia losing China. Clips/labels quarantined: SPEAKER_5 at 1:58 is an RFK Jr. bit, SPEAKER_7 at 1:18:56-1:20:13 is played Celebrity Jeopardy tape (Ken Jennings), and one Jason question at 52:38 landed under Tim Dillon's label — no captures taken from any of them.