Microsoft’s Vision for an Internet Made for Agents With CTO Kevin Scott (Best of the Pod) (28 min)
ai-driven-innovation-economy
ai-governance-laws
ai-in-workforce-disruption
privacy-in-the-ai-era
- Release date: 2026-08-12
- Listen on Spotify: Open episode
- Episode description:
In 2025, Kevin Scott bet that the agentic web would be the next big thing in AI.The Microsoft CTO argued that for agents to be genuinely useful, they'd need to be able to take action on our behalf—which would mean giving them access to the same sprawl of tools, data, and systems that make up the internet. Today, that bet is starting to pay off, as the foundational infrastructure for the agentic web is now being built.On this week's AI & I, Dan Shipper revisits his conversation with Kevin. They discuss Microsoft's role in the agentic web, why openness doesn't have to come at the expense of security, and why programmers should stay curious about new tools rather than resist them on principle.If you found this episode interesting, please like, subscribe, comment, and share!To hear more from Dan Shipper:Subscribe to Every: https://every.to/subscribeFollow him on X: https://twitter.com/danshipperGo to https://attio.com/every and get 15% off your first year.Timestamps: 0:00 Start1:44 Introduction2:49 The race to close the "capability overhang"4:31 How agents will evolve into practical, useful tools6:48 The role Kevin sees Microsoft playing in the agent ecosystem12:05 How robust security measures can coexist with open ecosystems15:39 Kevin's philosophy on being a craftsman in the age of agents20:52 How the landscape of software development agents will evolve25:33 The future of agentic workflowsLinks to resources mentioned in the episode:Kevin Scott on X: https://twitter.com/kevin_scottModel Context Protocol (MCP): https://modelcontextprotocol.ioNLWeb: https://github.com/microsoft/NLWebGitHub Copilot: https://github.com/features/copilot
Summary
- 🤖 Agentic Web Emerges: Focus shifts from scaling laws to building protocols, memory, and tool access so agents can act autonomously across systems.
- 🛠️ Craft Preserved: Makers retain strong tool opinions; agents augment outcomes while choice in process remains personal and valued.
- 🔒 Security First: Open protocols like MCP need agent identities and entitlement systems to enable safe delegation at scale.
- 🌐 Open Standards Win: Permissionless innovation via simple composable protocols mirrors the web’s success and avoids org-chart shipping.
- 🚀 Capability Overhang: Models now exceed current product use; closing this gap through ecosystem plumbing is the immediate priority.
Insights
- Why is closing the capability overhang between model reasoning and product delivery the next big focus after scaling laws?
- Time: 2:17 – 5:53
- Answer: Scaling laws are now accepted as reliable, shifting attention to memory, tool use, and ecosystem protocols so agents can handle complex delegated tasks effectively.
- What security models will enable the agentic web to thrive like the original internet while protecting user entitlements and identities?
- Time: 9:42 – 12:01
- Answer: Discussion highlights the need for agent identities, permission requests, and open protocols like MCP to avoid Conway’s Law pitfalls. Robust yet simple security is essential for enterprise and personal use.
- How can permissionless innovation in open agent protocols accelerate useful AI adoption without central gatekeepers?
- How might the rise of AI agents transform the craft of software engineering without diminishing the passion makers feel for their tools and processes?