Falling behind China on deep tech and chips
Ryan Graves warns about China's deep-tech push, Dylan Patel and Nathan Lambert explain chips, export controls and Taiwan, and Musk frames U.S.-China competition. Technology leadership is now a core question of national security.
What does America's AI lead depend on, according to Dylan Patel? Show answer
Nvidia's chip designs and TSMC's factories in Taiwan.
What did DeepSeek actually show? Show answer
That a capable Chinese lab can stay close to the frontier with fewer and weaker chips, through real engineering work. It did not leapfrog the U.S. overnight.
Why was DeepSeek's low cost figure misunderstood? Show answer
It covered only the final training run on about 2,000 chips, leaving out research, experiments and later training steps.
What does Nathan Lambert say export controls can and cannot do? Show answer
They cannot stop China from training models, but they limit how much AI China can run.
Why does Patel call export controls a bet on timing? Show answer
If AI is not transformative within five to ten years, controls help China long term, because it will subsidize and build its own chip industry.
What would it take to bring chipmaking fully to the U.S., per Patel? Show answer
About a decade and a trillion dollars.
Why is Taiwan so important to the technology race? Show answer
TSMC's leading-edge production and research are there, and even its Arizona fabs would stop working within a year or two without Taiwan.
What risk does Patel see in cutting China off from critical technology? Show answer
That it pushes China toward a blockade or takeover of Taiwan.
What sequence of bottlenecks does Elon Musk predict for AI? Show answer
A chip shortage now, then a shortage of voltage step-down transformers, then general electricity shortages.
What reassurance does Musk offer about China? Show answer
That China has historically been inward-focused and not acquisitive, thinking far more about itself than the outside world.
What is Ryan Graves's argument about funding? Show answer
China's state money can back long-shot science that takes a decade or more, which American venture capital tends to avoid.
Why does the article say falling behind overstates the case? Show answer
U.S. firms still lead in total compute, and the concern is that the lead rests on fragile foundations.