A Letter For Tyler Cowen & Dwarkesh Patel
A brilliant podcast from the Progress conference by Roots of Progress Institute
Last night I went on a run, in preparation for doing a 5k Trail run next weekend, and listened to this brilliant podcast. It was a refreshing listen, as both host (Dwarkesh) and Guest (Cowen) exhibit humility in the way they challenged and pushed each other with their points of view.
I've been haunted a bit today by something Tyler Cowen said about technology diffusion. It's that rare kind of insight that seems obvious in retrospect but fundamentally changes how you see the world going forward.
Tyler talked about how if you could transport [the economists] Ricardo or Malthus to our time, they wouldn't actually be that shocked by AI. They'd understand intuitively that while it would "extend the frontier," it wouldn't magically solve all our problems. There's something profound there about how we consistently overestimate the speed of change while underestimating its fundamental patterns.
It reminds me of conversations I have in the tech world, where there's often this implicit assumption that if we just make something good enough, adoption will naturally follow at lightning speed. But Tyler points out something subtle - that even if you can make a process twice as fast (like drug development going from 20 years to 10 years), you're still bound by the same human systems, regulations, and institutional inertia that shape how innovations actually spread through society.
In the tech world, where I spend much of my time working, has this tendency to overvalue raw intelligence and technological capability. We sometimes forget that even extraordinary breakthroughs face mundane but crucial limitations. It's like we're so focused on pushing the technological frontier that we forget about the complex dance of adoption, integration, and transformation that has to happen for technology to actually change lives.
What's striking about Tyler's perspective is that it's neither pessimistic nor dismissive of technological progress. Instead, it's a more nuanced and ultimately more useful way of thinking about how change happens. Some sectors will move quickly, while others - bound by regulation, institutional momentum, or human nature - will move more slowly. And that's not necessarily a bad thing.
I find myself applying this framework almost daily now. When I hear about a new breakthrough, instead of just asking "How powerful is this technology?", I find myself asking "What are the patterns of diffusion this will follow? What are the bottlenecks it will face? What are the human systems it needs to integrate with?"
Maybe the most valuable insight is that this slower, more constrained view of progress isn't actually a limitation - it's a feature of how human societies successfully absorb and benefit from change. The bottlenecks and constraints that sometimes frustrate us might actually be crucial to ensuring that progress is sustainable and beneficial.
The way that podcasts have democratized information is maybe one of the strongest forces to shape my own intelligence. I’m grateful for that, but also for the people like Dwarkesh and Tyler who have a unique ability to seed the information into these conversations. It’s not always just the information, it’s the way the conversation happens, the chemistry between host and guest, and the ability of each to challenge and push the convo.
With love and deep appreciation,
-Andrew