Wall Street Week | Larry Summers on the Fed, Argentina Elections, Hinton on AI, Trump’s H-1B Fee

Key Concepts

• Federal Reserve Monetary Policy: Interest rate decisions, inflation targets, data dependence, balance sheet reduction.
• US-China Relations: Trade, technology competition, AI, export controls, fentanyl cooperation.
• Argentine Economy: President Milei's reforms, inflation, fiscal spending, currency support, swap lines, FX intervention.
• Artificial Intelligence (AI): Risks, safety, existential threats, superintelligence, job displacement, regulatory challenges, international collaboration.
• H-1B Visa Program: Skilled immigration, fee increases, impact on innovation, startups, universities, and global competitiveness.

Federal Reserve and Economic Outlook

Larry Summers, a contributor from Harvard, discusses the Federal Reserve's recent decision to hold interest rates steady, marking the second consecutive meeting without a cut. Summers expresses his approval of this decision, stating it was the "right thing to do." He argues that inflation remains further from its target than unemployment, and the Fed has a more durable impact on inflation. Summers suggests that given current deficits and AI spending, the neutral rate of interest might be at or below the current level, making further rate cuts unnecessary. He commends Chair Powell's signal of a return to "data dependence" and "agnosticism about what was gonna happen next."

Summers acknowledges the risk of an economic slowdown but believes a 50-basis-point cut six weeks later would be insignificant compared to the potential loss of Fed credibility on inflation. He highlights the challenges of massive deficits, political pressure, international uncertainties, and rising inflation expectations as reasons for the Fed's cautious approach.

Regarding the availability of government data, Summers admits it's a disadvantage but notes the proliferation of real-time indicators, such as MIT's Million Price Project, which mitigate the problem. He considers politicization and budget deficits to be far greater challenges for the Fed than data limitations.

Summers critiques Chair Powell's comment about tariffs potentially bringing inflation closer to the 2% goal. He argues that removing tariffs from the numbers doesn't negate the fact that consumers are spending more on tariffed goods, thus less on others. He likens this selective data analysis to the "transitory inflation" idea from 2021, calling it "cherry picking."

The presence of dissents in the Fed's recent decision is seen by Summers as reflecting both the confusing economic picture and a genuine debate among economists. He dismisses the dissent favoring rate hikes as a "politically aberrant moment" due to the appointee's temporary status and the administration's rhetoric against the Fed.

Summers is not overly concerned about the Fed's decision to stop the roll-off of its balance sheet on December 1st, stating that the precise size of the balance sheet is less critical in an era of interest on reserves. He views the Fed reducing its balance sheet as a positive for the country as a whole, leading to less short-term federal government obligation.

US-China Relations and Technology

The meeting between President Trump and President Xi Jinping in South Korea is discussed. Summers emphasizes that the most significant outcome was the avoidance of a "massive confrontation and economic conflict." While acknowledging potential benefits for farmers through soybean sales, he asserts that long-term prosperity will not hinge on these sales. The critical issues remain technology and competition in artificial intelligence. Summers notes that these areas do not appear to have evolved significantly from the meeting. He credits the potential cooperation on fentanyl as a significant social issue.

Summers highlights the difficulty of managing export controls on advanced microchips between the US and China, balancing national defense with technological development. He criticizes President Trump's approach of relaxing controls in exchange for revenue sharing, deeming it inconsistent with American traditions and more characteristic of "deals capitalism" than "rules-based capitalism."

Argentina's Economic Reforms and US Support

President Milei's electoral victory in Argentina is presented as a strong mandate for his economic reforms. Hans Humes, CEO of Greylock Capital, notes that the country understands what has hindered its progress for decades and is willing to endure personal pain. He anticipates seeing how the transition unfolds and what new messaging emerges if Peronism prevails.

Humes reports significant financial gains for the US due to the election, with Argentine bonds and debt ratings increasing. He previously warned of potential social resistance to Milei's reforms, citing a pattern where initial market enthusiasm is followed by difficulties. He shares feedback from middle-class Argentines who feel the reforms are similar to past austerity measures, leading to disillusionment.

David Kim, a textile manufacturer in Buenos Aires, describes the long-standing challenge of runaway inflation in Argentina, which has been a constant since the 1970s. He explains the difficulty of managing costs and pricing when inflation is extremely high, sometimes requiring multiple price increases per month. Kim expresses that business owners in Argentina deserve recognition for their resilience in such a challenging environment.

Fabio Natalucci, CEO of the Andersen Institute for Finance and Economics, details the US support for Argentina, including a $20 billion swap line and potential FX intervention. He mentions reports of outright peso purchases and dollar sales by the Treasury. Natalucci explains how swap lines work, involving the exchange of currencies between central banks at a specified rate, with a commitment to swap back. He differentiates this from outright purchases, where the intervening country takes on the risk of currency devaluation.

Natalucci provides historical examples of US intervention in currency markets, primarily with advanced economies like Japan and European nations. He notes that the US Treasury's intervention in Argentina is a different exercise, involving direct market purchases.

Regarding potential profits from currency intervention, Natalucci explains that gains are only realized if the acquired currency is sold at a higher value. He points out that the peso's appreciation post-election was temporary, and holding pesos could lead to a return to pre-election levels. Swap lines, with their fixed exchange rates, do not involve gains or losses.

Natalucci clarifies that the primary purpose of the swap line and currency acquisition was to support President Milei and the Argentine government, not to generate profit. He suggests the intervention aimed to relieve pressure on the peso and prevent devaluation. He describes Milei's "shock therapy" of fiscal spending cuts and devaluation, followed by a controlled, slow depreciation of the peso.

He explains that interventions in the spot market are short-lived unless fundamental forces driving the exchange rate are addressed. He draws parallels to historical exchange rate-based stabilization plans in Chile and Argentina, where central banks eventually ran out of foreign currency. The US intervention is seen as a response to the Central Bank of Argentina potentially running out of dollars.

Natalucci questions the systemic importance of Argentina to US financial stability, noting that trade with Argentina is relatively small compared to other Latin American partners. He contrasts this intervention with previous rescue packages that were loans and combined with multilateral development bank support, totaling over $40 billion. He highlights the success of the Mexican package in 1997, which allowed for quick repayment and access to capital markets. The crucial question for Argentina is whether the US intervention will be sufficient to reverse the fundamental forces driving peso depreciation.

David Kim expresses hope that President Milei's policies will bring stable prices and growth. While acknowledging the reduction in inflation from 150% to 40%, he points to high bank interest rates exceeding inflation as an issue. He urges politicians to consider the difficult situation faced by everyone and calls for more than just inflation reduction for industrial companies.

Artificial Intelligence Risks and H-1B Visa Program

Geoffrey Hinton, a Nobel laureate and "godfather of AI," discusses the escalating risks associated with artificial intelligence. He draws an analogy to an alien invasion fleet, stating that humanity would be scared and act if faced with such a threat, and that AI development is akin to constructing these "aliens" that will be smarter than humans. He emphasizes the need to think "very, very hard" about coexistence.

Hinton expresses that companies like Anthropic and Google take AI safety "fairly seriously," but acknowledges the commercial competition. He believes Meta is "not particularly responsible" and that OpenAI, despite its founding principles, is becoming less responsible, with its best safety researchers leaving.

He criticizes the focus on the "race to become dominant" over the survival of humanity and society, particularly concerning massive unemployment. Hinton sees a potential for international collaboration on preventing AI from taking over, as no country desires this outcome. However, he believes the current model of control, where humans are the "boss" and AI is the "executive assistant," is flawed. He proposes a model of a baby controlling a mother as a more plausible way to coexist with superintelligence, accepting that humans would be the "babies."

Hinton states that the US is still slightly ahead of China in generative AI but that China is rapidly catching up due to its large number of skilled scientists and engineers. He suggests that attacking basic research funding and good research universities, as he perceives President Trump to be doing, is akin to working for Xi Jinping and will have long-term detrimental effects on US innovation.

He estimates that a trillion dollars or more has been invested in AI since his last conversation. Hinton worries that the primary way companies will make money is by replacing jobs, as this is the most direct path to increased profitability. He questions whether this time will be different from past technological revolutions, where new jobs were created, and expresses doubt that AI-driven job replacement will lead to new employment opportunities.

Hinton reflects on his role in AI's creation, acknowledging its potential for tremendous good in healthcare and education, but emphasizes that the negative consequences stem from how society is organized. He suggests that a "Chernobyl for AI" or a "Cuban Missile Crisis" moment might be needed to create a sense of urgency for safety and resource allocation.

He believes that the current market enthusiasm for AI, driven by significant investment, could potentially slow down development if it proves to be a bubble. However, he also notes that AI based on neural nets already works well and is useful for applications like reading medical images, making it unlikely to disappear.

Hinton explains that his Nobel Prize has increased his influence and ability to speak out. He distinguishes between two types of AI risks: those from bad actors exploiting AI's power (fake videos, cyber attacks, viruses) and the risk of AI itself becoming the "bad actor" by surpassing human intelligence. He describes the superior communication and learning capabilities of multiple AI copies, likening it to 10,000 people learning simultaneously.

He confirms that Canada was a significant hub for early neural network research due to sustained, curiosity-driven funding, which attracted top graduate students. He dismisses the idea of integrating symbolic logic into neural networks, comparing it to using electric motors to inject gasoline into an engine. He believes that progress in AI, such as "chain of thought" reasoning, demonstrates that AI thinks in natural language, not symbolic logic.

Hinton stresses the need for public understanding of AI risks to counterbalance the influence of big AI companies lobbying against regulations. He believes that public pressure is crucial for politicians to enact necessary regulations, particularly concerning shorter-term misuse risks and the longer-term threat of AI takeover.

He reiterates that AI's current effectiveness and ongoing improvement make it unlikely to be a bubble.

H-1B Visa Program and Skilled Immigration

The discussion shifts to the H-1B visa program and the proposed $100,000 fee. The program, established in 1990, allows highly skilled foreign workers into the US, with an annual cap of 65,000 visas (plus 20,000 for graduate degree holders). Demand consistently exceeds supply, leading to a lottery system.

Britta Glennon, an assistant professor at Penn's Wharton School, explains that H-1B visas are employer-sponsored and tied to a specific firm. She notes that demand has outstripped supply since the late 1990s, particularly with the growth of Silicon Valley.

The high demand has created opportunities for outsourcing and staffing companies, which accounted for nearly half of H-1B visas in 2023. Todd Schulte of FWD.US criticizes the system's susceptibility to "bad actors" and attributes the issues to Congress's failure to update the immigration system.

Jure Leskovec, a computer science professor and startup founder, argues that the lottery system disadvantages small startups that cannot afford to sponsor multiple visas and play the numbers game. He believes this hinders their ability to grow and innovate.

The proposed $100,000 fee is seen as potentially prohibitive for small organizations and startups, which are often the drivers of innovation. The concern is that this will stifle the birth and progress of innovative companies, impacting the economy. The difficulty of scaling a recruitment process based on such a fee is highlighted, with implications for both small and large companies, as well as research universities. The potential for scientific innovation to shift to other countries is raised.

Walmart, a major H-1B user, has announced it will no longer sponsor applicants due to the fee. NVIDIA CEO Jensen Huang, however, praised the announcement. The Department of Homeland Security (DHS) has clarified that college graduates on student visas and certain foreign workers already in the US will not have to pay the fee, though confusion remains about its annual or one-time nature and applicability.

Alternative approaches to skilled immigration are discussed, including points-based systems used by countries like Canada, Australia, and New Zealand, which allocate points for qualifications, language proficiency, and high-demand fields. An auction system where firms bid on H-1B visas is also proposed as a market-driven mechanism, contrasting with the government's arbitrary fee.

The potential for the US to lose skilled talent to other countries due to high costs is emphasized. Vijay Ravi, an H-1B recipient who returned to India, advises against coming to the US due to immigration stress, despite the potential for a good life. He highlights the trade-off between quality of life and immigration challenges.

The impact of insufficient access to highly skilled people is discussed, with AI talent in the US dropping from 60% to 40%. The ability to hire quickly and scale is seen as crucial for US competitiveness.

The segment concludes with a reminder of Geoffrey Hinton's upcoming discussion on two different threats from AI.