He coined 'vibe coding.' Now he says there's a 'growing gap' among AI users

Andrej Karpathy, a prominent figure in the AI community and former director at Tesla, has recently highlighted a concerning trend in the understanding of artificial intelligence capabilities among users. He suggests that a noticeable gap is emerging between ordinary users and those deeply engaged with advanced AI technologies. In a recent post on social media platform X, Karpathy described two distinct groups of AI users. The first group consists of casual users who have dabbled with free versions of ChatGPT and other early AI chatbots. These individuals may exhibit a limited understanding of AI’s full potential, often forming their opinions based on outdated experiences. In stark contrast, the second group comprises power users who invest in premium AI services like Claude Code and Codex. This demographic leverages advanced AI for high-stakes tasks such as programming, mathematical computations, and in-depth research. Karpathy emphasizes that these users appreciate the nuanced capabilities of state-of-the-art models, which are significantly more advanced than their free counterparts. Karpathy points out that many casual users often overlook the sophisticated developments in AI, focusing instead on its quirks and limitations. He referenced a viral moment where AI struggled to answer a simple question about transportation, illustrating how outdated models can misrepresent the technology’s current capabilities. As the AI landscape evolves, Karpathy notes that these two groups are often “speaking past each other,” leading to misunderstandings about what AI can achieve. OpenAI's CEO, Sam Altman, has also acknowledged the hesitance surrounding AI adoption, citing concerns among corporate leaders that could further widen this gap. With the growing interest in advanced AI tools, the hope remains that as more users engage with cutting-edge technologies, the divide in understanding may begin to close, fostering a more informed AI community.