Claude Opus 4.8 just dropped, and host Emily Laird is kicking the tires on Anthropic’s new heavyweight model for coding, reasoning, and long-haul AI work. This episode breaks down inference, agentic workflows, million-token context, effort control, and why “reliable AI” may be the new arms race. Think less magic chatbot, more suspiciously calm senior engineer who finally says, “Actually, this plan is broken.”