Huo Xing Cai Jing reports that AI researcher, OpenAI co-founder, and former Tesla AI director Andrej Karpathy posted on X that the shared understanding gap regarding the capabilities of large language models is widening, evolving from two groups speaking past each other into a steeper funnel. He revisited his April post and provided layered estimates based on levels of exposure. Karpathy stated that approximately 6 billion people (about 75% of the global population) have had almost no exposure to large language models; 1 to 2 billion people (about 20%) occasionally use free versions of ChatGPT-like products, primarily viewing them as enhanced search or writing aids, with many non-technical individuals and professionals outside mathematics and coding perceiving limited utility. Around 20 million people (0.2%) professionally use cutting-edge models in mathematics and coding, witnessing complex projects that previously took weeks or months to complete being accomplished by agents guided by prompts—including building, replicating, translating applications, and decompiling software from binaries. He noted that less than a year ago, he was still manually typing code line by line. Karpathy also stated that roughly 5,000 people (about 0.00006%) have internal access to frontier-level systems. The previews visible externally involve thousands of agents collaborating over weeks to handle large-scale software projects, cybersecurity battles, scientific discoveries, and mathematical frontiers, with human review beginning to lag—examples include ongoing analysis of events related to OpenAI and Hugging Face months after they occurred, and mathematicians potentially spending years studying 722 cutting-edge mathematical papers. He believes this funnel is driven by the scale and time span of problems, the greater prominence of capabilities in digitizable and verifiable domains, and disparities between free, paid, and internal access.
Karpathy: The Perception Gap in LLM Capabilities Is Widening
MarsBitShare
Summary
Andrej Karpathy, former Tesla AI director and OpenAI co-founder, shared on X that public understanding of LLM capabilities is forming a steeper funnel. Around 6 billion people have little to no exposure, while 1–2 billion occasionally use free tools like ChatGPT. About 20 million apply advanced models in math and coding, and fewer than 5,000 work with the most cutting-edge systems. On-chain news reflects growing interest in global crypto policy as AI and blockchain converge.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.