美联储生产力与就业工作组会议
文章语言:
简
繁
EN
Share
Minutes
原文
会议摘要
Policymakers are urged to plan for superintelligence’s uncertain magnitude, scope, and timing, weighing broad productivity gains against labor disruption and the risk of heavy investment with disappointing returns. AI is automating underwriting, boosting engineering, and enabling autonomous agents that reshape hiring toward product and sales, while widening gaps between agile startups and incumbents. Cybersecurity risks rise as open-source models lower barriers for misuse, prompting calls to harden critical infrastructure. Innovation is increasingly individual-driven and must be institutionalized, with creativity as a key differentiator. The U.S. is seen as a leader in diffusion and competitiveness, though concentration of surplus among large labs and global competition remain concerns.
会议速览
The Productivity and Jobs Task Force Conference, led by a Federal Reserve Bank of San Francisco president and CEO, convenes to discuss AI's transformative potential in the job market, highlighting OpenAI's ChatGPT release as a pivotal moment in November 2022.
Speakers highlight AI's rapid, transformative growth, comparing it to historical innovations like electrification and the Industrial Revolution. They emphasize the need for understanding AI's impact on growth, productivity, inflation, and the labor market, advocating for collaborative dialogue among technologists, academics, and policymakers to distinguish between AI's promises and reality.
The dialogue explores the transformative potential of superintelligence on economic growth and idea generation, emphasizing its role in overcoming secular stagnation and enhancing productivity through nonrivalrous innovation.
The dialogue explores the transformative potential of superintelligence (SI) on economic growth and innovation, emphasizing the need for human oversight and ethical considerations. It highlights the importance of contingency planning in a rapidly evolving technological landscape, aiming to prepare for future challenges while leveraging the benefits of SI. The discussion underscores the role of policymakers in navigating the risks and opportunities presented by SI, ensuring its adoption supports sustainable economic development without compromising human values and responsibilities.
The dialogue delves into the potential economic outcomes of a new technology's impact, considering scenarios of increased productivity and economic gains, labor market disruptions, and disappointing returns. It emphasizes the need for careful consideration of these contingencies in monetary policy, highlighting the uncertainty and the importance of American leadership in navigating these challenges.
Reflecting on past collaborations and current challenges, the speaker envisions a future of strong economic growth, vibrant communities, and technological advancement. They express gratitude for the opportunity to work with serious thinkers to address SI goals, concluding with an optimistic outlook for America's dynamic economy and human flourishing.
A discussion on how leading technology company leaders are leveraging new technologies, impacting their businesses, and potentially the broader economy, questioning if these advancements are unique to tech sectors or can be widely adopted.
A firm leverages AI and machine learning to enhance engineering productivity, enabling the implementation of more predictive features. This results in improved risk assessment models, significantly boosting loan volume and profitability. The firm anticipates a future where AI facilitates the realization of niche ideas, fostering a balance between supply and demand in the market.
Discusses how AI is rapidly transforming businesses at the frontier, with new companies outpacing incumbents. Highlights reduced human oversight in feature development, accelerated hiring in sales and engineering, and the unprecedented growth of new businesses leveraging Stripe. Contrasts the agility of startups with the governance challenges faced by established firms, echoing historical technology adoption patterns.
The dialogue highlights the dual focus on AI's potential to democratize expertise globally and the enterprise's cautious approach due to high costs and profitability concerns, emphasizing the need for balancing societal benefits with business interests in AI adoption.
Discusses AI's strengths in processing large data, its limitations in handling edge cases, and its potential to democratize intelligence, emphasizing human judgment in critical areas like cybersecurity and research.
AI's capability to identify and exploit vulnerabilities poses significant risks to global payment systems and infrastructures. The cybersecurity field faces an asymmetric battle, with a scarcity of experts and the widespread availability of advanced hacking tools. Urgent measures are required to address the imbalance and protect against increasingly sophisticated threats.
The dialogue discusses the increasing frequency of cyber events, emphasizing the need for identifying weak links in interconnected systems, particularly in banking. It highlights the upcoming technology rebuild phase, leveraging AI for vulnerability detection and code refactoring. The conversation shifts to market structures, questioning whether large tech firms or technology users will dominate, and underscores the importance of upgrading technical infrastructure to avoid obsolescence and security risks.
The dialogue explores the future of AI and its potential to transform society into a cybernetic ecology, referencing literary visions and current productivity gains. It acknowledges the difficulty in predicting long-term market structures due to the speculative nature of AI's trajectory and the uncertainty surrounding its ultimate impact on human labor and surplus distribution.
Discusses how profound innovations originate from individuals embracing new tools, emphasizing the role of leaders in institutionalizing these advancements. Highlights the importance of creativity in the face of AI's limitations, predicting a shift towards more creative work. Addresses the impact of productivity gains on companies, focusing on budgeting and capital allocation amidst technological upheaval.
The dialogue emphasizes the critical importance of prioritization and risk management in the lead-up to Black Friday, highlighting the balance between embracing innovation for growth and ensuring system stability. It discusses the challenges of deciding which ideas to implement, the human judgment involved in risk-taking, and the potential rewards of successful deployments.
Speakers discuss the accelerating pace of decision-making in software development, highlighting the tension between productivity and code stability. They emphasize the need for managing AI-generated code to prevent bugs and vulnerabilities, warning that unregulated AI usage can introduce security risks.
The dialogue explores the US's leading role in AI technology diffusion, emphasizing the reskilling of the workforce and the potential for economic gains through enhanced efficiency and competitiveness. It contrasts the focus on model production with the diffusion's economic impact, suggesting the US is uniquely positioned to leverage AI for global economic advantage.
Discusses US dominance in AI innovation and productivity growth, contrasting with European economic divergence. Highlights skepticism over current productivity stats and debates future energy consumption potential, emphasizing US preeminence in AI development.
The dialogue explores the rapid advancement of AI technology, emphasizing its potential to revolutionize industries and economies. It highlights the importance of global leadership in AI development, noting the regulatory lag in Europe and China's competitive stance. The conversation underscores the need for investment in AI infrastructure and addresses concerns about energy consumption and market dynamics. Despite potential challenges, including cyber threats and economic shifts, the speakers remain optimistic about AI's role in enhancing creativity and societal progress, advocating for a gradual, adaptive approach to managing AI's integration into society.
Reflects on de Tocqueville's concept of soft despotism, explores the future of work in an AI-driven society, and considers strategies to prevent a complacent and unambitious social structure while envisioning a positive societal outcome.
The dialogue explores the potential for AI to disrupt industries, emphasizing the incumbents' early awareness and investment, which could solidify monopolistic control. It speculates on a future economy dominated by top corporations and the resilience of the creative long tail, while cautioning about the hollowing out of the middle tier and the role of agency in repopulating it.
要点回答
Q:What potential does AI have according to the speech?
A:According to the speech, AI has the potential to transform, and it is expanding at an unpredictable rate, making it difficult to predict its capabilities from one day to the next.
Q:How does the speaker describe the current transformational period?
A:The speaker describes the current period as transformative and different from anything she has experienced in her career, comparing it to transformative historical events like the Internet boom, electrification, or the Industrial Revolution.
Q:What is the role of the Productivity and Jobs Task Force as mentioned in the speech?
A:The Productivity and Jobs Task Force, led by Mary Daly and others, is tasked with examining the potential of AI and its effects on the economy. The group includes technologists, academics, businesses, and other leading thinkers.
Q:What are the opening remarks of Kevin Warsh about the conference and the task force?
A:Kevin Warsh commended the efforts of Mary Daly and the staff at the San Francisco Fed for organizing the conference on short notice. He expressed his first visit to a Reserve Bank as chairman was impressive, highlighting the expectations, reform orientation, and forward-looking spirit necessary for a changing world.
Q:What is the importance of the task force convened by Chairman Warsh?
A:The task force convened by Chairman Warsh is one of five he has convened as chairman. It serves to broaden thinking, foster robust discussions, and better equip the Federal Reserve for future policy challenges.
Q:How might superintelligence change economic policy and long-run growth?
A:Superintelligence might change economic policy by dramatically enhancing the rate and quality of idea generation, potentially allowing for sustained growth levels substantially higher than the long-run average. This could also impact the conventional view that slower population growth may reduce the number of new ideas.
Q:What is the significance of human contributions in the age of AI?
A:Despite the advancements AI can bring in terms of idea generation and innovation speed, humans are not replaced. Their roles involve judgment, morality, dignity, and responsibility, which are crucial in designing, building, and deciding the use of technology.
Q:What is the stance of the Federal Reserve regarding predictions about technology?
A:The Federal Reserve, including Kevin Warsh, does not speculate on future technology or make predictions. Instead, their job is to prepare for whatever comes next, which includes contingency planning for the fast-changing world.
Q:What are the economic policy considerations when it comes to AI according to the speech?
A:Economic policy considerations include being mindful of the risks associated with AI, especially in terms of vulnerabilities in critical infrastructures. The Federal Reserve aims to be neither advocates nor opponents of AI but to play a role in managing the new era of AI.
Q:What are the potential effects of widespread adoption of new technology on the economy?
A:The widespread adoption of new technology could significantly affect the economy with important implications for jobs and monetary policy.
Q:What are the key questions surrounding the impact of this new technology?
A:Key questions include whether the technology will lead to a rise in total factor productivity, if the democratization of specialized knowledge will raise standards of living, when this surge will start and taper off, and whether the US will lead in deploying this technology compared to other countries and adversaries.
Q:How should the central bank respond to the potential impacts of new technology?
A:The central bank has a responsibility to consider the range of economic outcomes that could occur based on the progress of new technology.
Q:How has the Federal Reserve's staff evaluated different economic scenarios?
A:The Federal Reserve's Board staff has begun evaluating various contingencies using stylized model-based frameworks to generate illustrative scenarios, which are not forecasts or predictions, to consider potential economic outcomes.
Q:What types of economic outcomes could arise from the progression of new technology?
A:Economic outcomes could range from persistently high real GDP growth and low unemployment to significant labor market disruption, with GDP growth and higher unemployment rates. There could also be a scenario where the economy disappoints with capital investment, falling stock market gains, and an economic slowdown.
Q:What potential disruptions could arise from new technology in the labor market?
A:New technology could lead to substantial labor market disruption as some jobs are displaced, potentially pushing the jobless rate above levels consistent with full employment, despite GDP growth.
Q:What are the expectations for American leadership in the context of technological advancements?
A:The expectation is that America will meet the challenges before it and lead to strong, durable economic growth and vibrant, secure communities with a healthy environment for future generations.
Q:What is the significance of having serious thinkers join the Federal Reserve?
A:It is a privilege to be a policymaker at a time of significant consequence, and it is encouraging to be joined by serious, sober, and responsible thinkers who are willing to contribute to the Fed and steer the direction of new technology goals.
Q:What insights will be gained from questioning panelists who are using new technology in their companies?
A:The panelists will share insights on the real happenings inside their companies and how they are using new technology, which could help understand broader implications for other companies and whether the technology will find its way more broadly.
Q:What is the impact of using machine learning on engineering productivity and loan production?
A:The use of machine learning has led to a fivefold increase in engineering productivity and a 10fold reduction in cost per feature. It has also resulted in the deployment of more features into production, translating into a 80% increase in performance and an additional two billion dollars per year in loans for the lender.
Q:How is Stripe leveraging new technology, and what is the effect on their team and operations?
A:Stripe is leveraging new technology to see a significant acceleration in the work of engineers and autonomous task identification by agents. This has led to end-to-end implementation and deployment of features with reduced human oversight compared to six months ago.
Q:What is the effect of AI on hiring and productivity in different departments?
A:AI is reducing hiring for general administrative functions, while there is an acceleration in sales due to diminishing future production costs, increased buyer interest, and higher productivity of salespeople.
Q:How has the use of new SaaS collaboration tools impacted company operations?
A:New SaaS collaboration tools have changed company operations, as seen when a large company's CEO was surprised by the collaborative and transparent way a smaller company, Stripe, was operating.
Q:What disparity exists between incumbent organizations and new companies with regards to new technology adoption?
A:There is a disparity and a mental model impedance mismatch between incumbent organizations and new companies regarding the adoption of new technology, with startups being more receptive to adopting new practices.
Q:How is the 2026 cohort of companies performing on Stripe compared to previous years?
A:The 2026 cohort of companies started on Stripe is outperforming any prior year cohort by a dramatic magnitude, with new businesses launched on Stripe increasing approximately twofold year over year.
Q:What insights does the Paul David 1990 paper offer in comparison to current AI adoption?
A:The Paul David 1990 paper discusses how benefits from electricity primarily accrued to new firms rather than incumbent firms adapting, and the current observation of new firms on Stripe is proceeding at an unprecedented pace in terms of AI adoption.
Q:What is the current adoption rate of AI among large companies?
A:Large companies, also known as the Fortune 450, are not changing their organization, mental models, processes, or governance quickly enough to adapt to new AI technologies.
Q:Why is there a focus on enterprise versions of new technology?
A:The focus on enterprise versions of new technology is primarily due to the high costs involved in building large data centers to train models and generate intelligence, which makes the economic viability of these technologies a significant consideration.
Q:What are the current applications of AI in enterprise operations?
A:AI is being used to process large amounts of data and retain large amounts of information, with applications in customer support and decision-making processes involving analytics and understanding. However, AI's capability to act independently on behalf of enterprises is still limited.
Q:What is the role of human judgment in the face of AI's increasing capabilities?
A:While AI is advancing in capabilities, particularly in areas like cybersecurity, human judgment is still needed for understanding and addressing edge cases and ensuring the safety of interconnected systems.
Q:How will AI democratize access to intelligence?
A:AI has the potential to democratize access to intelligence, providing society and humanity with vast amounts of intelligent resources and capabilities.
Q:What concerns are there regarding the security of payment systems and infrastructures?
A:There is concern about the security of payment systems and global infrastructures due to AI's ability to find vulnerabilities and bad code, and the current state of cybersecurity is at a disadvantage as the demand for understanding vulnerabilities exceeds the industry's capacity to provide assistance.
Q:What measures can be taken to address these concerns?
A:To address these concerns, it is crucial to ensure that interconnected systems understand their weakest links, particularly the banking system, and that measures are taken to safeguard these critical points in the infrastructure.
Q:What is the anticipated technology rebuild moment and its implications?
A:The anticipated technology rebuild moment refers to a period where there will be a necessity to go back and address past mistakes and improve on aspects that were previously overlooked. AI is expected to facilitate this process by helping to find vulnerabilities and refactor existing code. This rebuilding is crucial to avoid being in a worse position due to outdated technology, as evidenced by incidents like airline IT failures and Microsoft's recent issues.
Q:How does the market structure evolve after a period of technological development?
A:The market structure post-technological development is uncertain and may significantly change, with many speculating about whether the financial benefits will come from large labs or the users of these technologies. This transformation could represent a fundamental shift in economics, and the eventual outcome may depend on future developments and how the technological advancements are utilized.
Q:What is the vision for the future of AI as suggested by Daria Mote's essay?
A:Daria Mote's essay suggests a positive vision for AI's future, referencing a poem by Richard Brodigan where humans are freed from their labor and returned to a state of nature, with machines taking over the tasks. This vision implies a society where humans are no longer confined to repetitive tasks and can instead focus on more creative endeavors.
Q:What are the potential outcomes for market structure in the face of technological advancements?
A:The potential outcomes for market structure in the face of technological advancements are varied and uncertain. The vast majority of surplus is currently captured by the users of these technologies, leading to immense productivity gains. However, the exact long-term equilibrium is hard to predict and may evolve based on future developments and changes in the use of technology.
Q:What is the role of individuals and companies in driving technological transformation?
A:Individuals are at the forefront of driving technological transformation, with most profound innovations beginning with them. As leaders, it is their responsibility to institutionalize these gains within their companies. Companies play a crucial role in nurturing this transformation by institutionalizing patterns of genius and adapting to the evolving landscape of technological advancements.
Q:How is the landscape of AI and security vulnerabilities evolving?
A:The landscape of AI and security vulnerabilities is rapidly evolving, with AI serving as a tool to unearth latent bugs quickly. There is a temporary period of intense activity where all kinds of vulnerabilities are being found and fixed with AI support. Over time, this is expected to lead to the development of more robust software.
Q:In what way will creativity be affected by the technological revolution?
A:The technological revolution is expected to unleash a tremendous amount of creativity, as machines have yet to generate a brilliant idea independent of human origin. With increased efficiency and productivity in areas like fulfillment and price comparisons, creativity is anticipated to become the primary currency of the future, with individuals and companies pivoting towards more creative endeavors.
Q:What does the term 'upheaval' mean in the context of companies and their labor forces?
A:In the context of companies and their labor forces, 'upheaval' refers to a positive movement driven by the need to strategically allocate productivity gains. This process involves prioritization and risk management to determine the most appropriate investments or changes to be made, such as enhancing the software code base, while being mindful of introducing additional bugs or instability.
Q:What are the implications of AI's current capability to write code on the economy and security?
A:AI's current capability to write code has implications on the economy and security as it produces code that has introduced vulnerabilities which adversaries are exploiting. This has led to concerns regarding the robustness and security of AI systems.
Q:What are the implications of AI on the US economy and its trading partners?
A:The implications of AI on the US economy and its trading partners are related to the economic growth differential. The US is considered to be ahead in AI adoption and has the potential to achieve efficiency gains, margin gains, and global competitiveness. This could result in a stronger US economic position relative to trading partners.
Q:How is AI being diffused within companies and what is the significance of this?
A:AI is being diffused within companies through a gradual reskilling of the workforce, where every employee is experimenting with or observing the use of AI. This is significant as it indicates a shift towards a culture of curiosity and creativity in adopting new technology.
Q:What is the US's position in terms of AI production and global diffusion?
A:The US is seen as the center of AI production and a key location for intelligence production. While the full extent of the benefits from AI models built at this cost is yet to be seen, the US is likely to experience faster diffusion of AI and its associated benefits, leading to efficiency gains and global competitiveness.
Q:Is the US leading in AI adoption intensity and how does this compare to global adoption?
A:The US is leading in AI usage intensity, with the highest adoption rates observed. This leads to a competitive advantage as the US is ahead in the adoption and use of AI technology.
Q:Does the US hold a significant advantage over other countries in AI development?
A:The US holds a definitive preeminent position in AI development and usage, according to the discussions. There is a consensus that this lead is likely to grow, especially when considering the significant investments and advancements in AI happening within the country.
Q:What potential risks or challenges could emerge in the context of AI and its global impact?
A:Potential risks or challenges that could emerge with AI include regulatory lags, energy consumption concerns, and the impact of AI on employment and jobs. There is also a risk of 'soft despotism,' where the comfort provided by technology could stifle ambition and creativity. However, these challenges are seen as manageable and opportunities for positive societal transformation are emphasized.
Q:How might AI influence market competition and the future shape of the economy?
A:AI could influence market competition by potentially allowing incumbents to solidify their grip on the industry due to increased capital access. However, it could also foster a diverse and innovative economy with a multitude of small businesses. There is a concern that without proper adaptation, the market could evolve into a scenario dominated by a few large corporations, impacting pricing and consumer outcomes.

旧金山联邦储备银行
Follow





