Karpathy on DeepSeek-OCR paper: Are pixels better inputs to LLMs than text?
NeutralTechnology
Andrej Karpathy recently discussed the implications of the DeepSeek-OCR paper, which explores whether using pixels as inputs for large language models (LLMs) could be more effective than traditional text inputs. This conversation is significant as it could reshape how we think about data input in AI, potentially leading to advancements in machine learning and natural language processing.
— Curated by the World Pulse Now AI Editorial System