
The podcast analyzes the recent releases of Anthropic Opus 4.6 and OpenAI's 5.3 Codex, focusing on their implications for developers and AI practitioners. It highlights the increased context window of Opus 4.6 to one million tokens, which improves handling large codebases and complex financial tasks. The speakers discuss adaptive thinking in Opus 4.6, which automatically adjusts reasoning effort based on the task's complexity, and the default context gathering for actions, addressing previous user complaints. They also explore the shift towards using skills and evals over traditional RAG for ensuring accurate results. The conversation touches on the cost implications of using these advanced models and addresses concerns about the "coding is dead" narrative, emphasizing the continued importance of skilled developers who understand business impact.
Sign in to continue reading, translating and more.
Continue