Interconnects

Interviewing Arvind Narayanan on making sense of AI hype

Oct 17, 2024

Arvind Narayanan, a computer science professor at Princeton and director of the Center for Information Technology Policy, delves into the realities of AI amidst the hype. He discusses the pitfalls of AI policy, emphasizing the need for harm-focused research. The conversation covers the risks of open-source foundation models, critiques of traditional AI in risk prediction, and the implications of scaling laws. Narayanan also sheds light on the balance between innovation and societal impact, highlighting the necessary collaboration between researchers and policymakers.

54:21

Creator website

Episode guests

Arvind Narayanan

AI Summary

AI Chapters

Episode notes

Podcast summary created with Snipd AI

Quick takeaways

The podcast emphasizes the importance of a proactive approach in addressing potential harms from AI to ensure societal benefits are realized.

Discussions highlight challenges in AI policy, particularly in distinguishing between generative and traditional predictive AI to inform effective regulation.

The complexities of scaling AI are examined, stressing that improved model performance does not guarantee practical applicability in real-world scenarios.

Deep dives

Optimism Amid Criticism

Despite being labeled as an AI critic, the speaker expresses a profound optimism about technology, particularly in the realm of artificial intelligence. They emphasize that some AI applications, like foundation models and generative AI, have significant potential for positive societal impacts in the long run. However, they underscore the need for researchers and technologists to proactively address the potential harms posed by these technologies rather than simply reacting to them. This proactive stance is seen as essential for guiding policy and ensuring that the benefits of AI are realized while minimizing possible negative consequences.

Intro

3min

Navigating AI's Societal Challenges

13min

AI in Coding and Research Dynamics

19min

The Balance of Innovation and Social Diffusion

2min

Exploring the Gap Between AI Advancements and AGI

3min

Navigating Privacy and Accountability in AI

13min

Skepticism Towards Traditional AI and Its Moral Implications

2min

Arvind Narayanan is a leading voice disambiguating what AI does and does not do. His work, with Sayash Kapoor at AI Snake Oil, is one of the few beacons of reasons in a AI media ecosystem with quite a few bad Apples. Arvind is a professor of computer science at Princeton University and the director of the Center for Information Technology Policy. You can learn more about Arvind and his work on his website, X, or Google Scholar.

This episode is all in on figuring out what current LLMs do and don’t do. We cover AGI, agents, scaling laws, autonomous scientists, and past failings of AI (i.e. those that came before generative AI took off). We also briefly touch on how all of this informs AI policy, and what academics can do to decide on what to work on to generate better outcomes for technology.

Transcript and full show notes: https://www.interconnects.ai/p/interviewing-arvind-narayanan

Chapters

* [00:00:00] Introduction

* [00:01:54] Balancing being an AI critic while recognizing AI's potential

* [00:04:57] Challenges in AI policy discussions

* [00:08:47] Open source foundation models and their risks

* [00:15:35] Personal use cases for generative AI

* [00:22:19] CORE-Bench and evaluating AI scientists

* [00:25:35] Agents and artificial general intelligence (AGI)

* [00:33:12] Scaling laws and AI progress

* [00:37:41] Applications of AI outside of tech

* [00:39:10] Career lessons in technology and AI research