Zephry
59 posts


So memory tax is effectively getting framed as Nvidia tax… If that’s the case, then the stock price action inevitably gets tied to the same framework as well.

CHINA CONSIDERS RESTRICTING OVERSEAS ACCESS TO CUTTING-EDGE AI MODELS China’s Ministry of Commerce has led meetings over the past month with major AI companies, including Alibaba, ByteDance, and Z.ai, to discuss measures that would restrict overseas access to cutting-edge AI models, including models that have not yet been released. The discussions reportedly include not only closed-source models but also open-weight models. However, the scope of application is still under debate, and the rules may ultimately apply only to future frontier models. Officials have also discussed designating the leakage or theft of proprietary AI technologies as a national security crime, with stronger penalties, as well as restricting the types of foreign capital that can invest in Chinese AI startups. The backdrop is the U.S. move to strengthen export controls on AI models, along with national security concerns over cutting-edge models that could possess advanced cyberattack capabilities. Chinese authorities are reportedly concerned that advanced U.S. cybersecurity AI models could be used to exploit vulnerabilities in Chinese software. Since the beginning of this year, China has continued to tighten measures to prevent AI technology from being transferred overseas. Authorities have investigated whether Chinese AI startups that relocated abroad violated export control laws, while also strengthening oversight of overseas transactions involving Chinese investors, technology, data, and national security concerns. Future regulations could take the form of a tiered framework based on technological capability. Basic open-source AI models may be managed through a filing system, high-performance models may be subject to security reviews, and the most sensitive frontier models may be banned from public release or restricted to use within China.

Chamath is one of the few people in the AI industry who seems to understand that AI products will be bought for their “profitability” and not their “intelligence”. This includes both enterprise and (surprisingly) consumer products. Nobody is going to pay 5,000% higher price for 2% more performance in applications where 95% of the value is captured by “good enough”. The power law doesn’t apply to farming, or road maintenance, or cooking dinner for a family of four, or a million other things. In some places the edge cases capture all the value, art, sport, software, etc. The power law applies in domains where edge cases are 99% of the value or 99% of the liability. But not everything in the world has this topology. This is why classifiers (a control layer) are required for product market fit. The classifier question is often as simple as “is this an edge case?” It wasn’t obvious whether very advanced and sophisticated models would be able to solve simple and easy problems cheaply… So far, they cannot. This makes the more advanced models less valuable, and they’re going to lose a lot of value (prompt traffic) to classifiers as a result. Will monolithic models evolve their own internal classifiers and gain the ability to answer low value questions with low cost responses? Who knows. But they can’t do that today and that changes the shape of what needs to be built. A few are using “adaptive models” which is the first attempt at solving this, but each user has a different classifier curve. So product market fit has a few orthogonal dimensions.













