Dive into the cutting-edge worlds of technology and science with Dwarkesh Patel's meticulously researched interviews that bring together brilliant minds to unpack revolutionary ideas in AI, biotech, and beyond. Each episode delivers profound, thought-provoking conversations that challenge conventional wisdom and illuminate the paths shaping humanity's future.

Dwarkesh Podcast

https://x.com/jamesdillard/status/1879622099263967637

https://x.com/Yacqubismial/status/1879424582375080111

https://x.com/stephenjaoates/status/1879525336511304009

https://x.com/erratagnostic/status/1879364421094670761

https://x.com/g_rodriguez_r/status/1879642656973942957

https://x.com/Jackmc/status/1879505941298864423

https://x.com/StatisticsFTW/status/1879429373423652933

https://x.com/micahlerner/status/1879380584218399007

Having covered Chinese technology closely, the most important pattern is that military-civil fusion comes more naturally to AI than to any prior technology, because the same algorithms serve both commercial optimization and military targeting.

Military-Civil Fusion Is Easier in AI

The next big AI signal from China may not come from consumer chatbots, but from state-owned enterprises deploying foundation models across the physical economy—manufacturing, ports, and energy grids—in ways the US is not yet matching.

Physical Economy AI Deployment in China

Europe's comprehensive AI regulation, however well-intentioned, may lock the continent out of frontier AI development, leaving the United States and China as the only two real poles of advanced AI capability.

Europe Regulates Itself Out of AI's Frontier

Export controls on China mean nothing if Washington fails to pair them with aggressive investment in domestic fabs, advanced packaging, and chip talent. Denial alone is not a strategy; it must be coupled with building.

Controls Without Investment Will Fail

Advanced semiconductor export controls are redrawing the global map of AI more than any single policy since the Manhattan Project, creating twin technological blocs with separate supply chains and standards.

Export Controls as the New Manhattan Project

Are we already in a cold civil war over technological norms, with separate internet ecosystems, standards, and even values? Is a fully decoupled global internet now inevitable, or are there forces that still pull toward integration?

Is a Tech Cold War Already Arrived?

If China cannot access Nvidia's best chips, could that deprivation actually force the Chinese AI ecosystem to become more innovative in the long run—developing alternative architectures, software efficiency, or analog computing?

Does Scarcity Make China More Innovative?

Beyond AI and advanced chips, what are the most underrated fields in the US-China technological rivalry—biotech, quantum computing, energy storage, or something else—and where might a breakthrough genuinely shift the balance?

Underrated Arenas of US-China Tech Competition

When Chinese entities pre-stockpile chips, lithography equipment, and spare parts months before new US controls take effect, are the sanctions actually slowing their progress or just taxing it with higher costs and delays?

Are Stockpiles Blunting Sanctions?

Looking back from 2030, what is the question about US-China technological competition that we will most regret not asking ourselves five years earlier?

2030 Regret Premortem for US-China Tech

Contrary to expectations of rapid depreciation due to technological advancements, an H100 GPU is worth more today than three years ago because it can now run more advanced AI models that generate significantly higher value, such as GPT-5.4 compared to GPT-4. This dynamic highlights how compute value increases with model improvements in a capacity-constrained environment, benefiting early committers like OpenAI over more conservative players like Anthropic.

By 2030, the primary bottleneck for scaling AI compute will shift to ASML's EUV lithography tools, with production limits potentially capping cumulative AI chip capacity at around 200 gigawatts, challenging ambitious goals like Sam Altman's vision of deploying a gigawatt per week amid the ongoing AI boom.

Despite ambitious demands for massive AI compute scaling, such as Elon Musk's 100 gigawatts per year by 2028, the semiconductor supply chain lags far behind due to insufficient AGI-driven production ramps and architectural inefficiencies, making reliance on older 7nm processes inadequate for bridging the gap. This highlights a critical bottleneck in AI advancement, where performance differences between chip generations can reach 20x due to networking and data movement constraints, not just raw FLOPS.

Why is TSMC allocating the majority of its 3nm process capacity to Nvidia rather than diversifying among competitors like Google or Amazon to fracture the market and maintain leverage?

What prevents ASML from significantly increasing its production of EUV tools beyond 100 per year by 2030, despite the explosive demand for AI chips?

Given the compounding performance advantages of newer chip architectures, what fraction of the demanded AI compute could realistically be achieved by repurposing older 7nm fabs, and under what economic conditions would this become viable?

Get Jordan Schneider on Dwarkesh Podcast

The Case for This Conversation

Shape What Gets Discussed

You Might Also Enjoy