Brockman announced that GPT-5.4 Pro has successfully solved a new Erdős problem, marking a significant milestone in AI capabilities. This achievement contrasts with past efforts, where extensive resources were needed to train models for competitive tasks. Brockman highlighted the potential of applying such advancements to fields like drug discovery, suggesting untapped opportunities. Additionally, Brockman hinted at improvements in personalization for large language models. Responding to critiques about the lack of "soul" in AI, he mentioned an upcoming model designed to address these concerns, inviting users to evaluate its enhancements post-release.