Why Scaling Large Language Models Alone May Not Lead to Artificial General Intelligence

Artificial General Intelligence (AGI) remains one of the most ambitious goals in arti-ficial intelligence research. Recent advances in transformer-based large language modelshave demonstrated powerful emergent capabilities, leading to debates regarding whetherscaling alone may eventually produce AGI. However, despite impressive performance acrossreasoning and language tasks, current systems still exhibit major limitations in long-termplanning, memory persistence, grounded understanding, and autonomous reasoning. Thispaper analyzes the architectural limitations of modern large language models and exploresadditional components that may be necessary for achieving AGI.