# Chatbot Software Begins to Face Fundamental Limitations ![rw-book-cover](https://www.quantamagazine.org/wp-content/themes/quanta2024/frontend/images/favicon.png) ## Metadata - Author: [[Anil Ananthaswamy]] - Full Title: Chatbot Software Begins to Face Fundamental Limitations - Category: #articles - Summary: Large language models (LLMs) like ChatGPT struggle with complex reasoning tasks, showing fundamental limitations in their ability to solve problems that require compositional thinking. Research indicates that these models typically approximate solutions based on their training data, which can lead to incorrect answers. Despite efforts to improve their performance, LLMs are constrained by inherent mathematical boundaries, making some tasks beyond their capabilities. - URL: https://www.quantamagazine.org/chatbot-software-begins-to-face-fundamental-limitations-20250131/ ## Highlights - Take basic multiplication. Standard LLMs, such as ChatGPT and GPT-4, fail badly at it. In early 2023 when Dziri’s team asked GPT-4 to multiply two three-digit numbers, it initially succeeded only 59% of the time. When it multiplied two four-digit numbers, accuracy fell to just 4%. ([View Highlight](https://read.readwise.io/read/01jjzsqqfye6n8j40cw9mkq6hy))