WordPress community urged to increase AI literacy

During the WordCamp US keynote, discussions highlighted the need for the WordPress community to gain fluency in artificial intelligence technologies. Key areas for educational focus include understanding how large language models are trained through reinforcement learning from human feedback (RLHF) and the distinction between base models and post-training processes.

The presentation also addressed the rise of open-weight models, noting that researchers have found ways to modify or remove refusal behaviors in certain models. Additionally, the emergence of high-performing, low-cost open-weight models from Chinese laboratories, such as Qwen, DeepSeek, Kimi, and GLM, was identified as a significant trend in the field.