Keep pulling the thread on Marc Andreessen.
AI development is geographically concentrated, with the vast majority of foundational work occurring only in the United States and China.
The current wave of AI companies is growing customer revenue at an unprecedented rate, faster than any previous technology wave.
The price of AI is falling much faster than Moore's Law due to collapsing per-unit costs for all of its inputs.
The capabilities of large, leading-edge AI models are being replicated in smaller, more efficient models within 6 to 12 months of their release.
The new version of the Chinese open-source model Kimi has replicated the reasoning capabilities of GPT-5 and can run on one or two MacBooks.
Mark Andreessen predicts that AI chips will become cheap and plentiful within five years due to intense competition from companies like AMD, hyperscalers, and Chinese manufacturers challenging NVIDIA's market position.
The Chinese government has reportedly instructed the developers of the DeepSeek model to build its next version exclusively on Chinese-made chips, likely from Huawei, to bolster the domestic semiconductor ecosystem.
The U.S. policy landscape for AI has improved dramatically, shifting away from potential bans due to the growing awareness in Washington D.C. that the U.S. is in a competitive race with China.
The European Union's AI Act has significantly hindered AI development in Europe, causing major U.S. companies like Apple and Meta to withhold launching their leading-edge AI products there.
California's proposed AI bill, SB 1047, which was modeled on the EU AI Act and would have assigned downstream liability to open-source developers, was passed by the legislature but vetoed by the governor.
Elon Musk's company xAI was able to catch up to the state-of-the-art performance of OpenAI and Anthropic models in less than 12 months from a standing start.
Four different Chinese companies have now developed AI models that have effectively caught up to the state-of-the-art capabilities of Western models.