At Stripe, engineers now collaborate on crafting the perfect prompt to guide AI agents. This new form of teamwork focuses on articulating the problem clearly and providing the right context, rather than co-writing code line-by-line. This can involve other engineers, data sources, or even other agents.

In a remote environment, immediate access to colleagues isn't always possible. A GPT loaded with context about your company and cofounders' thinking can act as a thought partner, helping you overcome the "blank slate" problem without scheduling a meeting.

Vercel CEO Guillermo Rauch demonstrates that production-ready AI prompting goes beyond simple feature requests. His prompt to v0 for a rating system also included crucial constraints, such as preventing abuse (rate limiting) and adhering to the existing design style, reflecting a production-first mindset.

Instead of switching between ChatGPT, Claude, and others, a multi-agent workflow lets users prompt once to receive and compare outputs from several LLMs simultaneously. This consolidates the AI user experience, saving time and eliminating 'LLM ping pong' to find the best response.

Anthropic's new "Agent Teams" feature moves beyond the single-agent paradigm by enabling users to deploy multiple AIs that work in parallel, share findings, and challenge each other. This represents a new way of working with AI, focusing on the orchestration and coordination of AI teams rather than just prompting a single model.

Block's CTO argues that LLMs are a wasted resource when they sit idle overnight and on weekends. He envisions a future where AI agents work continuously, proactively building features, running multiple experiments in parallel, and anticipating the needs of the human team so that new options are ready for review in the morning.

Tools like Claude CoWork preview a future where teams of AI agents collaborate on multi-faceted projects, like a product launch, simultaneously. This automates tactical entry-level tasks, elevating human workers to roles focused on high-level strategy, review, and orchestrating these AI "employees."

A hybrid approach to AI agent architecture is emerging. Use the most powerful, expensive cloud models like Claude for high-level reasoning and planning (the "CEO"). Then, delegate repetitive, high-volume execution tasks to cheaper, locally-run models (the "line workers").

Non-technical users are accustomed to a "prompt, wait, respond" cycle. Cowork's design encourages a new paradigm where users "hand off" significant work, let the AI run for hours, and check back on results, much like delegating to a human assistant.