The evolution of large-scale language models (LLMs) like ChatGPT marks significant strides in understanding the world through human language. These systems, having integrated innovations such as web browsing and image generation, edge closer to true general-purpose artificial intelligence (AGI). However, realizing this potential calls for an understanding that transcends language, delving into mathematics and non-verbal realms like dance motion. The phrase from the Gospel of John, 'In the beginning was the Word (Logos),' hints that language derives its essence from the annals of natural history. As a mirror of humanity's intellectual legacy, language undeniably underpins the remarkable achievements of LLMs. Yet, for AGI to fulfill its promise, grasping mathematics and dance motion remains critical.
Understanding mathematics in its entirety is universally challenging. Defined by abstract thought and enriched with conceptual metaphors, mathematical thinking diverges sharply from linguistic reasoning. Its pivotal role in driving historical industrial progress, and central importance in AI and algorithm development, cannot be overstated. However, systems like ChatGPT find encoding basic mathematical constructs and emulating complex mathematical thought challenging; they fall short in offering refined expressions akin to Euler's formula or Schrödinger's equation. The pursuit of mathematical thinking, inherently distinct and more intricate than linguistic structures, necessitates the insights of extraordinary minds. AGIs adept in both language and mathematics will exhibit enhanced comprehension and problem-solving capabilities, potentially unveiling novel solutions to humanity's myriad challenges.
