China vs US AI and Energy

The Future of Code Review: Stop Reviewing Line-by-Line, Start Governing AI Agents - Itamar Friedman

Topic 8 of 11

Published May 4, 2026 · Clip 48:11–55:26 (7:15)

Itamar frames the China-vs-US LLM race largely as open-source vs closed-source, noting NVIDIA's Nemotron as a promising US open-source effort that Qodo tested (Supernova 3) with surprisingly strong, efficient results. He argues Chinese foundation labs like Alibaba have business moats (cloud, like Google/Gemini) pushing them toward open source, while OpenAI/Anthropic's moat is the LLM itself. He predicts the two capability curves will close the gap and run together rather than cross, but sees energy as the decisive frontier where China holds an advantage. He shares an S-curve view of innovation — intervals between breakthroughs shrinking — and argues an energy S-curve is needed for AI to start creating AI, equivalent to running billions of human brains for a century.