May 2023

May 2023 is anchored by Geoffrey Hinton's resignation from Google and his warning of 'danger ahead' — a moment that lends mainstream legitimacy to long-held AI safety concerns and reframes the conversation from speculative to urgent. Technical education content surges, with explainers on transformer models, RLHF, and autonomous agents meeting growing demand for foundational literacy, while LangChain's JavaScript expansion and generative AI's impact on developer workflows mark the tooling ecosystem maturing rapidly. The safety and governance thread deepens on multiple fronts: Australia's AI Ethics Principles and CSIRO framework gain coverage, DeepMind launches an early warning system for novel AI risks, and LessWrong's AI takeover risk essay circulates alongside a new EU algorithmic transparency initiative. Prompt injection extends into PDF resumes, copyright disputes continue with a text archaeology of ChatGPT's training data, and Yejin Choi's TED Talk on AI being 'incredibly smart and shockingly stupid' captures the persistent ambivalence underlying a month of intense, unresolved debate.