This news podcast covers recent developments in the AI industry. The episode begins with headlines summarizing a surprising new Gemini AI capability (processing two visual inputs simultaneously), Google's move to offer free AI in Gmail and Docs (while raising Workspace prices), and Meta's internal focus on beating OpenAI, revealed in court documents. The main segment then dives into eight predictions for how AI agents will improve in 2025, four from Anthropic's Chief Scientist and four from the host's observations from conducting agent readiness audits. Specifically, improvements are expected in tool usage, contextual understanding, coding assistance, safety measures, data quality, multi-agent systems, evaluation infrastructure, and ROI tracking. The podcast concludes by encouraging listeners to engage in the discussion and consider agent readiness audits.