YouTubeStanford CS146S: How always-on agents work
The clearest end-to-end architecture of an always-on agent, from its builder.
Topic archive
YouTubeThe clearest end-to-end architecture of an always-on agent, from its builder.
YouTubeThe bottleneck never dies, it moves — and Ramp automates each new one in place.
YouTubeGranola's co-founder demos the meeting stack his team actually runs.
YouTubeAnthropic's Claude Code insider on where agentic coding goes next.
YouTubeOpenAI product leads on shipping at model speed: plan 90 days, not years.
YouTubeA GitHub Next research engineer rebuilds design practice for the agent era.
YouTubePrompting is the floor; harness engineering is the ceiling.

A clear reverse-engineer of Instinct's git-Markdown agent memory.
YouTubeUnfiltered product judgment from Snap and Discord's ex-product chief.
YouTubeA working alternative to open-ended agents: AI employees with jobs and budgets.
YouTubeThe durable advantage is a set of compounding loops, not a single funnel.