Faith in God-like large language models is waning (www.economist.com)

The initial excitement around large language models (LLMs) like OpenAI's ChatGPT, once hailed as revolutionary breakthroughs akin to the launch of the iPhone in 2007, is now giving way to growing skepticism. Recent developments, particularly with OpenAI’s GPT-5, have generated comparatively little enthusiasm, signaling a slowdown in perceived innovation. This shift suggests that improvements are increasingly seen as incremental rather than transformative, much like the routine annual upgrades in smartphone technology. This tempered enthusiasm holds significant implications for the AI/ML community. It challenges the prevailing narrative of exponential AI progress and may encourage a more measured appraisal of LLM capabilities and their real-world impact. For tech companies like Apple, which has been relatively cautious in embracing AI, this could present an opportunity to catch up without pressure to lead every advance. Technically, while GPT-5 embodies incremental refinements—such as enhanced language understanding and efficiency—these are less headline-grabbing than previous leaps, pointing to a maturation phase for large-scale generative models rather than a new AI frontier.