Dwarkesh Patel on AI misalignment
From AGI Thought Experiment with Dwarkesh Patel, Noah Smith, and Erik Torenberg | From the a16z Podcast · · Econ 102 with Noah Smith
“I think the main thing I worry about is the AI playing us off each other rather than us playing the AIs off off each other. More so like the way that the East India Company was able to play different provinces in India off of each other and to ultimately to the once at some point you realize okay like they control India. Um and so you could have a scenario like okay think about the concistadors right a couple hundred people show up to your border and they take over an empire of 10 million people and this happened not like once it happened like two to three times.”
On , Dwarkesh Patel, CEO and Founder at The Dwarkesh Podcast, spoke about AI misalignment during AGI Thought Experiment with Dwarkesh Patel, Noah Smith, and Erik Torenberg | From the a16z Podcast on Econ 102 with Noah Smith.
About Dwarkesh Patel
Dwarkesh Patel, founder and host of The Dwarkesh Podcast, has been a frequent guest on other programs and published episodes with researchers and executives. On Triggernometry, Patel discussed the potential societal effects of artificial intelligence, stating that he finds the prospect of mass job displacement "scary" and that AI could make authoritarian surveillance far more efficient because "a lot of the reasons that government has not been as authoritarian as it has in the past is that it just physically not been possible." He also said that while he is "a very libertarian person by inclination," he believes the dynamic of capital replacing labor "justifies a huge amount of redistribution." Regarding AI sentience, Patel said he "genuinely doesn't know" whether current systems are sentient, and argued that future AI systems will need to have "their own values" and that a "constitutional convention" should be held to define those values. Patel has also hosted guests including former Google DeepMind researcher Eric Jang, who discussed rebuilding AlphaGo and the lessons it offers for self-play and reinforcement learning; Harvard geneticist David Reich, who presented new findings showing accelerated natural selection during the Bronze Age; Nvidia CEO Jensen Huang, who defended Nvidia's moat by stating that "the transformation from electrons to tokens is such an incredible journey" and is "hard to completely commoditize"; and research fellow Michael Nielsen, with whom Patel explored how scientific progress is recognized and how that question applies to AI-driven discovery. Patel has described the improvement of AI models as "very fast" and observed a "huge discrepancy between what people are seeing in Silicon Valley and what people are observing outside."