AI and I

Last year, there were 1 billion commits on GitHub. This year, Kyle Daigle expects that number to exceed 14 billion, a two-component explosion caused by more humans—and their agents—issuing pull requests. In March alone, 17 million pull requests on GitHub were created by agents.Daigle is the COO of GitHub and Microsoft’s chief marketing officer for developer products. He’s been at GitHub for 13 years, and is paying close attention to how AI is expanding the platform’s user base. Along with agents, legal, sales, and marketing professionals are building apps with the GitHub Copilot app. The line between developer and non-developer is disappearing.On this episode of AI & I, guest host Mike Taylor sat down with Daigle at Microsoft Build to discuss how GitHub is building infrastructure for an agent-native world: agentic code review, model routers that automatically select the right model for the task, and a philosophy that the most durable advantage in this market is developer choice.If you found this episode interesting, please like, subscribe, comment, and share!Want even more?To hear more from Mike Taylor:Subscribe to Every: https://every.to/subscribeFollow him on X: https://x.com/hammer_mtTimestamps for YouTube:00:00:52: Introduction00:03:27: The agentic PR flood00:04:33: GitHub's approach to helping open-source maintainers manage the surge00:06:15: What 14 billion commits means for code quality00:08:03: Moving from per-seat licensing to usage-based pricing00:09:45: Kyle's dual role as GitHub COO and Microsoft's chief marketing officer for developers00:13:03: Developer choice as competitive moat00:14:57: How to balance dogfooding your own tools with staying honest about the competition00:19:45: Hill climbing, frontier tuning, and solving the model-routing problem00:24:45: Kyle's agentic communication hackLinks to resources mentioned in the episode:Kyle Daigle on X: https://x.com/kdaigleMike Taylor on Every: https://every.to/@mike_2114Mike’s piece on building an AI version of Kyle Daigle: https://every.to/also-true-for-humans/i-interviewed-an-ai-version-of-github-s-coo-then-spoke-to-the-real-oneGitHub Copilot: https://github.com/features/copilot

Mike Krieger built one of the most consequential consumer apps of the last two decades as the cofounder of Instagram. He is now at the frontier of AI-native product development as head of Anthropic Labs, the team responsible for figuring out what the most capable AI models can do in the hands of real builders.When Krieger first got access to Fable 5 months before its public release, it was exciting and disorienting. “I feel like a total newbie again,” he remembers telling his team. The way he’d been thinking about productivity, strategy, and time management was out of date. The model had outpaced his workflows.Dan Shipper talked with Krieger for AI & I about what it looks like to build with a model as capable as Fable 5, including the new rhythms, challenges, and possibilities it reveals.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/danshipperGet started with Braintrust at https://www.braintrust.dev/ Timestamps:0:03 Introduction1:48 How Fable completely reshaped Mike's workflow4:48 When to use Sonnet versus Fable10:06 What the media tracker Mike built over a weekend reveals about agent-native architecture15:00 The cost to build has collapsed19:03 Is software engineering over?21:48 How Anthropic's engineering teams work today38:39 The mechanics of verification44:39 What people should use the model to build47:24 Dynamic workflowsLinks to resources mentioned in the episode:Mike Krieger on X: https://x.com/mikeykAnthropic Labs: https://www.anthropic.comClaude Code: https://claude.ai/codeEvery: https://every.toTimestamps:0:03 Introduction1:48 How Fable completely reshaped Mike's workflow4:48 When to use Sonnet vs. Fable10:06 What the media tracker Mike built over a weekend reveals about agent-native architecture15:00 The cost to build has collapsed19:03 Is software engineering over?21:48 How Anthropic's engineering teams work today38:39 The mechanics of verification44:39 What people should use the model to build47:24 Dynamic workflowsLinks to resources mentioned in the episode:Mike Krieger on X: https://x.com/mikeykAnthropic Labs: https://www.anthropic.comClaude Code: https://claude.ai/codeEvery: https://every.to

If your MCP server has dozens of tools, it's probably built wrong. You need tools that are specific and clear for each use case—but you also can't have too many. This creates an almost impossible tradeoff that most companies don't know how to solve.That's why we interviewed Alex Rattray, the founder and CEO of Stainless. Stainless builds APIs, SDKs, and MCP servers for companies like OpenAI and Anthropic. Alex has spent years mastering how to make software talk to software, and he came on the show to share what he knows. We get into MCP and the future of the AI-native internet. [Disclosure: Dan is a small investor in Stainless.]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/danshipperGet started with Braintrust at https://www.braintrust.dev/ Timestamps: 00:01:15 - Introduction 00:05:09 - APIs and MCP, the connectors of the new internet 00:11:00 - Why MCP exists 00:17:15 - Why MCP servers are hard to get right 00:20:24 - Design principles for reliable MCP servers 00:25:06 - Using MCP for business ops at Stainless 00:40:57 - Alex's take on the security model for MCP 00:44:42 - How one-off AI actions become permanent production softwareLinks to resources mentioned in the episode:Alex Rattray: Alex Rattray (@RattrayAlex), Alex RattrayStainless: https://www.stainless.com/

Most AI companies are racing to build bigger LLMs. Eve Bodnia thinks that's the wrong approach.Eve is the founder and CEO of Logical Intelligence, which is developing an alternative to the transformer-based models dominating the industry. Her argument: LLMs’ architecture makes them fundamentally unsuited for some mission-critical tasks. A system that generates output one token at a time, with no ability to inspect its own reasoning mid-process or guarantee its results, shouldn't be trusted to design chips, analyze financial data, or even fly a plane. Her alternative is the energy-based model (EBM), a form of AI rooted in the physics principle of energy minimization, not language prediction. Rather than guessing the next probable word, an EBM maps every possible outcome across a mathematical landscape, where likely states settle into valleys and improbable ones sit on peaks. Dan Shipper talked with Bodnia for AI & I about why she believes LLM progress is plateauing, what it means for AI to actually understand data rather than just pattern-match across it, and how her team is building toward formally verified code generated in plain English—no C++ required.If you found this episode interesting, please like, subscribe, comment, and share!Head to http://granola.ai/every and get 3 months free with the code EVERYTo hear more from Dan Shipper:Subscribe to Every: https://every.to/subscribe Follow him on X: https://twitter.com/danshipper Timestamps: 00:00:51 - Introduction00:02:09 - Why correctness and verifiability matter in AI00:09:33 - What an energy-based model is00:14:21 - How EBMs construct energy landscapes to understand data00:19:00 - Why modeling intelligence through language alone is a flawed approach00:26:54 - What it means for a model to "understand" data00:37:21 - How EBMs solve the vibe coding problem and enable formally verified code00:43:21 - Why LLM progress is plateauing00:49:54 - Mission-critical industries haven't adopted LLMs, and how EBMs could fill that gap

While walking to the office, our COO Brandon Gell had his AI agent call him and go over his emails in his inbox one by one. When he arrived, he opened Gmail and confirmed she'd done everything he'd asked. "My jaw is on the floor," he messaged me.That was the moment Every got serious about setting up each employee with their own agent. Today, it's a reality—and it has completely changed how we work.Dan Shipper talked to Every COO Brandon Gell and head of platform Willie Williams for Every's AI & I about what happens when everyone at a company gets their own AI sidekick. If you found this episode interesting, please like, subscribe, comment, and share!To hear more from Dan Shipper:Subscribe to Every: https://every.to/subscribe Follow him on X: https://twitter.com/danshipper Visit https://scl.ai/dialect to learn more about Dialect, a new system from Scale AI.Timestamps: 00:00 Introduction00:02:21 How Brandon built Zosia, an AI agent to run his household00:07:09 Brandon's aha moment re: using agents for work00:09:39 What happened when everyone on the team got their own agent00:12:42 How agents take on their owners' personalities, and why that matters inside an org00:23:51 Why it's important for agents to do work in public00:30:51 What we're still figuring out when it comes to agent behavior, including memory gaps, group chat etiquette, and the "ant death spiral" problem00:40:45 How we built Plus One, our hosted OpenClaw product00:47:27 The cultural shift required to make agents work at scale