The emergence of GPT-4 has been the biggest news in the AI world in recent weeks, with ChatGPT plugins bringing augmented language models to the masses and new language model tricks being discovered. However, an open letter calling for a stop on giant AI experiments has caused concern around AI ethics and existential risk theories. Meanwhile, there have been several research breakthroughs with large language models, including their ability to learn within the prompt and embed in self-improvement loops. Additionally, language models are being applied beyond language to decision-making contexts, with challenges around the dataset gap in reinforcement learning.

source update: Trends in AI — April 2023 // GPT-4, New Prompting Tricks… – Towards AI


There are no comments yet.

Leave a comment