Sam Altman on AI's future in banking

TL;DR

AI is rapidly accelerating, driving huge productivity gains and dramatically lowering the cost of knowledge work. It will reshape banking and many jobs by enabling new capabilities, while some tasks disappear and new roles emerge. The comparison to the transistor illustrates how a deep tech shift diffuses through society and products.

Transcript

Um we're looking really looking forward to all of the conversations that we're having today. Um especially for what we're going to take away as we're we're continuing to do our work within the inter agency and also with the Federal Reserve on on capital and as we're looking to the future of banking uh more generally. I also want to take a a moment ... Read More

Key Insights

• AI progress has accelerated and changed productivity expectations in finance and technology.
• AI can reduce the cost of knowledge work from high to very low, enabling faster implementation of tasks.
• AI systems already perform at expert levels in many fields, though long horizon tasks remain challenging.
• There is a large potential for productivity gains in software development and automation through AI.
• The cost per unit of intelligence has dropped significantly, suggesting widespread adoption across sectors.
• The analogy to the transistor highlights how AI could become a ubiquitous, invisible backbone for products and services.
• Future labor dynamics will include both job displacement and the creation of new roles, with human trust and interaction remaining important.
• There is an expectation of years of ongoing progress in AI research and deployment, with broad societal impact.

Questions & Answers

Q: What is the most important takeaway about AI's impact on productivity?

The most important takeaway is that AI is dramatically increasing productivity across many tasks, enabling people to accomplish in minutes what previously took hours or days, and this acceleration is expected to continue for years. This shift will transform how work gets done in banking and other industries, with broad implications for how teams operate and deliver services.

Q: How does AI currently affect software development according to the discussion?

AI is used to automate much of the software development process, making tasks that used to require extensive human effort much faster and cheaper. Programmers report higher productivity, and AI can perform substantial portions of coding and tasks that previously needed weeks of work, drastically reducing development cycles and cost.

Q: What historical analogy is used to describe AI's future impact?

The discussion uses the transistor as the best analogy. Like transistors that enabled the broader electronics revolution, AI could become a ubiquitous underpinning for products and services, diffusing widely across industries and driving large productivity gains without people thinking in terms of AI specifically.

Q: What does the speaker say about long horizon tasks for AI?

AI currently struggles with tasks that require long horizon planning or complex, multi-step reasoning akin to long-term human projects. While many capabilities are rapid and short-term, fully autonomous, long-horizon decision making remains a challenge, suggesting a phased adoption where AI handles many routine tasks while humans oversee more strategic work.

Q: What is said about the pace of AI progress in the near term?

The speaker notes that progress is likely to continue for many years, with multiple waves of improvement ahead. Even if research slows temporarily, the overall trajectory points to sustained advances and widespread impact on productivity, economies, and how services are delivered.

Q: How might AI influence labor markets according to the chat?

AI will both displace certain job classes and create new ones. The implication is that workers will use new tools to perform more in less time, changing the nature of many roles while requiring new skills and training. Trust, human interaction, and the ability to manage AI-assisted processes will remain important.

Q: What is said about the cost of AI-enabled knowledge work?

AI has driven down the cost of knowledge work significantly, making it cheaper to produce results that once required substantial human labor. The example discussed suggests that tasks that used to be expensive can now be completed with minimal compute costs, enabling faster delivery of outcomes and lower overall project costs.

Q: What is the overall outlook for AI deployment and impact?

The outlook emphasizes long-term, ongoing progress with widespread deployment across products and services. AI is expected to become a normal, integrated part of many offerings, improving capability, efficiency, and user experience, much like the ongoing diffusion of other transformative technologies in history. It will take time to realize the full scope of impact.

Summary & Key Takeaways

• AI progress is rapid and has wide-reaching implications for finance and regulation.

• AI dramatically reduces the cost and time of software tasks, enabling new capabilities in banking.

• The analogy to the transistor explains how AI will diffuse into everyday products and services, changing work and economy over time.