Topic
Curated briefings and essays tagged with LLMs.
Latest
Lead
OpenAI’s fourth large language model (LLM), GPT-4, took an estimated 50 gigawatt-hours to train, or the equivalent of 5,000 American homes’ yearly power consumption. That was in 2023. Since then, the computational resources used to train frontier LLMs have only increased, though...
Also read
More briefings (2)
This article is part of our exclusive IEEE Journal Watch series in partnership with IEEE Xplore. As robots advance in terms of dexterity and other physical capabilities, it becomes more likely that humans may find themselves working alongside them. If that happens, how will robot...
Understanding the behavior of complex machine learning systems, particularly Large Language Models (LLMs), is a critical challenge in modern artificial intelligence. Interpretability research aims to make the...
Essays (2)
Industry
OpenAI has evolved from a non-profit research laboratory into the definitive platform company of the Generative AI era. By pivoting to a "capped-profit" model and forging a $13B partnership with Microsoft, it has commoditized Large Language Models (LLMs) through its GPT series. Its strategy now hinges on becoming the operating system for AI applications—the "iOS of AI"—despite facing significant leadership turbulence and mounting regulatory pressure.
Research
The next frontier of AI is "Multimodality"—the ability for a single model to natively understand look, listen, and speak. Unlike early systems which glued together separate models for image recognition and text generation, modern Multimodal Large Language Models (MLLMs) like GPT-4o and Gemini process all data types as tokens in a uniform "embedding space," allowing for seamless reasoning across senses.