Claims

Informal discussions have already occurred between Constellation Energy and the NRC regarding Three Mile Island restart before any formal license application was filed, indicating that major players perform regulatory due diligence before signing massive contracts.

We need institutions we can trust for public health and national security; proliferating podcasts and newsletters are not sufficient infrastructure for navigating existential risks, and the conversation cannot be entirely crowdsourced to alternative media.

The ethos at All Souls has changed over time: previously, fellows had freedom to pursue intellectual interests; now there is more emphasis on monitoring what fellows are reading and whether they have published papers, suggesting that credentialism and bureaucratic oversight are gradually replacing the original institutional model.

Regulatory constraints on pension funds and insurance companies are deliberately pushing them into domestic bonds. France, for example, requires savers to hold life insurance policies (tax-advantaged) which funnel into domestic bonds, with regulatory restrictions limiting equity allocation to 20-30%, preventing diversification into gold or other assets. Similar patterns exist globally, creating structural capital control.

Gavk has been wary of the AI/semiconductor trade and lost money to colleagues who played it (his Asian opportunities fund manager had a monster year with semis), suggesting he lacks consistent edge in tech trends and is more confident in international opportunities where he's written extensively.

Korea recently implemented a policy offering capital gains-free treatment for citizens who repatriate foreign assets and buy domestic stocks, while implementing capital gains taxes on foreign stock investments. This tax policy is designed to encourage repatriation and has contributed to Korean stock outperformance as investors sell Microsoft/Nvidia to buy SK Hynix.

Japanese capital repatriation flows, unlike Mike Green's passive robot flows which move slowly and steadily, would likely be 'fast and furious.' This is because once the Japanese Government Pension Investment Fund (GPIF) initiates repatriation, it sends a green light to all other pension funds who copy GPIF to avoid risk of underperformance, creating cascading repatriation flows.

NIH (National Institutes of Health) has historically declined to fund protein and amino acid metabolism research because they see no direct disease relevance; this created a funding vacuum that commodity groups (dairy council, beef board) filled to support Layman's groundbreaking leucine research that revolutionized protein understanding

Commodity groups like the Dairy Council and Beef Board are subject to USDA oversight and cannot see publications until after peer review; commercial companies like Coca-Cola and Kellogg's have contractual veto power over publication, making commodity-group-funded research inherently more independent

The Pentagon was built as a State Department for Cold War Soviet containment and now represents 36% of global R&D spending, but this massive legacy institution is structurally obsolete for current geopolitical challenges and its very organizational weight prevents adaptation.

International institutions (UN, WTO, IMF, etc.) only have as much power as member states grant them; they are 'talk shops' and do not exercise meaningful governance; the most important institution moving forward will be the US Federal Reserve, not international organizations.

Any international organization that includes both China and India as members is functionally incapable of decision-making because they have different opinions on everything that matters; therefore, organizations like BRICS cannot produce binding outcomes.

Government efficiency varies significantly by location and institution (some states/governments are better-governed than others, some more bloated than others), suggesting governance can work better or worse depending on specific structures and incentives.

The speaker served on a committee to award blue plaques to historically significant figures and proposed that Basil Spence should have a plaque reading 'Architect and Vandal,' which would be completely accurate, but this was rejected because blue plaques are governed by a 19th-century notion of exemplary lives.

The art fund (presumably a major source of institutional art sponsorship) has been perverted and corrupted by modernism and the Turner Prize, which the speaker says he must work to expose by 'calling out the emperor's new clothes.'

The European Union faces the biggest risk: active regulatory hostility toward AI, lack of homegrown AI models, numerous laws making AI illegal or impossible, insufficient governance coordination due to quasi-federal structure, and inability to compete technologically—EU cannot progress on AI under current constraints.

Removing a single federal regulation can take 2-3 years due to administrative procedure, environmental review, and legal processes, making comprehensive regulatory reform a decade-long project even with aggressive political will.

The UK has less federalism than the US (simplifying infrastructure approvals) but faces a larger fiscal crisis, a talent crisis of unexplained origin, and administrative state constraints that impede the kind of growth Silicon Valley envisions despite theoretical governance advantages over the US.

DeepMind's timeline for AGI is 5-10 years, which is not long for institutions to be built to handle the transition; this is a relatively short time horizon for institutional adaptation.

Despite responsible behavior by major labs and market incentives, rogue actors (rogue nations, organizations, people using open source) may still build irresponsible systems; this could lead to a 'warning shot' incident that catalyzes international cooperation and standardization.

Society and economists should be spending more time thinking about post-AGI scenarios, but institutions are fragmented and insufficiently influential; geopolitical tensions make collaboration harder than on other existential issues like climate change.

AGI transition will require collaboration across institutions and nations; Hassabis sees himself as collaborative and hopes he can contribute to that collaboration from his current position.

The nation-state, despite predictions of its obsolescence in the 1990s-2000s, remains the primary institutional form organizing human political life; it remains the entity to which citizens turn for protection and governance (illustrated by the inadequacy of non-state actors in these roles).

The destruction of trust in public health institutions by figures like Dr. Anthony Fauci is something 'we never get over,' and when the next plague comes (and it will), people will actively resist public health measures because institutional trust is permanently broken.

Hassabis is worried that institutions and governments aren't moving quickly enough or collaborating internationally on AGI governance despite having only 5-10 year timelines before AGI, and existing international institutions seem fragmented and insufficiently influential for the scale of challenges ahead.

India's bureaucratic structure inherited from British administration makes industrial development slower than China's; therefore even if decoupling were possible, India cannot match Chinese speed, reinforcing that reshoring is not a viable alternative.

The Berkman Klein Center at Harvard brings together academics, industry experts, and tech professionals to address the biggest challenges posed by the internet through its Spring Speaker Series.

70% of U.S. housing land is zoned for single-family homes, preventing both new abundant housing supply and experimentation with communal/clan-based family forms that many young people desire.

A steady flow of new institutions (firms, voluntary associations, universities) is necessary to prevent institutional sclerosis; the United States was historically very good at generating new institutions until recently, when a noticeable decline in new company formation, new bank formation, and zero new universities founded in recent decades indicates the onset of a 'great degeneration.'

University corruption in science is harder to see than corruption in humanities because scientific misconduct is esoteric (cancer research not progressing, string theory making no empirical progress), so unlike the humanities where corruption is transparent, we cannot verify whether the sciences are actually serving their stated purpose.

Diversity hiring in universities has corrupted both humanities and sciences; the case of a neuro-scientist (Marc Tessier-Lavigne) at Stanford who engaged in multi-decade research fraud and stole tens of millions went largely unpunished, whereas a diversity administrator (Claudine Gay) was quickly fired for plagiarism.

The moderator notes that Institute economist colleagues won the Nobel Prize this year for work on institutions, and suggests that institutions scholars should work on governance questions relevant to AI—a suggestion Hassabis endorses.

Land records around the developing world are very poor, with cadastral records (land ownership records) being only partially accurate, and the median age of land records globally is approximately 40 years, meaning most places lack reliable documentation of actual land ownership.

Chinese government statistical officials only recently (2024) realized their population data had been wrong for 27 years because birth registration points were systematized differently (hospitals in developed countries vs. schools/taxes in China), doctors were inflating immunization numbers for payment, regional governments inflating school enrollment for subsidies, and COVID disrupted validation.

Brooke Rollins, USDA Secretary, had to deal with a bird flu crisis on her first day after Trump fired all top 1,600 federal personnel and all seasonal workers; she initially proposed vaccinating all birds (requiring 12 months per 3-million-bird flock), then pivoted to financing assistance when she realized the plan was infeasible—exemplifying how lack of institutional knowledge cripples policy execution.

Of the top 1,600 federal positions Trump fired, only 5% have been filled (by 'flunkies who don't know what they're doing'), another 5% are waiting on Senate confirmation, and the Senate is absorbed in a 17-bill bundle, so confirmation hasn't even begun—resulting in paying for a full government while not having one.

The Federal Reserve's arrogance is exemplified by spending $2.5 billion on renovation of their buildings while having 200 PhDs on staff, suggesting institutional detachment from reality and fiscal responsibility.

In elite university systems like Yale, there is simultaneous external prestige (everyone wants to get in) and internal misery (everyone wants to leave), creating a trap where the institution's external appeal masks internal dysfunction.

Most major institutions (media, civil service, political parties, corporations, education) have been captured, corrupted, or perverted and now project a worldview onto the public that the public does not share

The Institute of Applied Ecology of the North organizes scientific expeditions to the crater, with researchers needing to watch out for bears, mosquitos, and awakened ancient microorganisms in the region.

The University of Melbourne physics building has in its basement a large particle accelerator with several beam lines used for material science research directed toward quantum technology projects, and a cryogenic system for measuring material properties at very low temperatures, particularly the behavior of spins.

The speaker was offered a fellowship by the Royal Society (Wolfson visiting fellow) which supported her visit and activities in the UK for several weeks after Australia's borders opened following two years of closure.

'Adequate' security is completely subjective, but many companies reference the Gartner quadrant to provide cover for their liability—if they bought the tools Gartner recommended, they have defensible legal position even if breached, making Gartner recommendations function as liability shields rather than security drivers.

Ackman had calls with 35 heads of major scientific institutions and research organizations; almost all of them redirected their work from cancer to coronavirus research, illustrating the talent mobilization.

The Goodfellows program itself was enabled by pandemic virtualization, demonstrating that crisis-driven innovation can create value even when the crisis is destructive overall.

Green Mantle (Ferguson's organization) has successfully implemented distributed team model with regular multi-day offsites in pleasant locations, demonstrating that this approach works effectively in practice.

Hoover Institution and similar think tanks have benefited from pandemic virtualization because it enabled geographic access to experts who would previously visit only occasionally, expanding the effective talent pool.

Think tanks and institutions will face governance questions about disclosing AI contribution percentages (e.g., '83% AI') and whether to publish AI-contributed work, requiring new editorial policies.

Central banks are typically reactive agencies rather than proactive ones; they wait for crises to happen and try to solve them rather than getting ahead of the problem.

The popular narrative that the Bank for International Settlements is the central bank for central banks and controls them is a total misunderstanding; the BIS is actually a host for meetings providing secretariat support, similar to the OECD, with no independent decision-making authority.