Resources

AI-native code editor built on VS Code. Composer mode handles multi-file edits end to end. Approved for use with both clients and employer.

My take: My primary agent IDE for professional work. If you want to build real working prototypes as a PM, start here. A dedicated post on agentic IDE experience is coming.

Anthropic's agentic AI tool. I use the desktop app rather than the terminal CLI — more accessible, no technical setup needed. Employer-approved; personal use on a Claude Max subscription.

My take: This is how I use Claude. Not the chat interface, but as an agent that operates on files and handles multi-step tasks. The desktop app makes it approachable for anyone, not just developers.

A fully free agentic IDE: VS Code with Roo or Cline as the AI extension, connected to LM Studio running a local model. Everything runs on your own hardware, no subscription needed.

My take: The zero-cost path to a capable agentic setup. Useful when tools aren't employer-approved, for sensitive content, or when you want full control over the model. A dedicated walkthrough post is coming.

Google's AI assistant. Employer-approved. I use it for quick, one-off conversations and occasionally alongside NotebookLM for notebook-based work. Personal use via Google One AI Premium.

My take: A side tool, not a primary workspace. Fast for a quick answer or gut-check. I'm gradually shifting context work into local project folders instead.

Plain Markdown files organized into project-level folders. No dedicated app, no lock-in. Instructions, notes, and context live alongside the work they relate to.

My take: How well your agent IDE performs depends almost entirely on instruction quality and how you cascade context across a project. Local folders with clear structure are what make that possible.

Clayton Christensen's framework for understanding what customers truly hire a product to do. Separates the functional, social, and emotional dimensions of user need.

My take: More relevant in the AI era, not less. AI products that skip the JTBD layer build impressive demos that don't get adopted. Start here before writing a single prompt.

Donella Meadows' primer on stocks, flows, feedback loops, and leverage points. The mental model for understanding why complex systems behave the way they do.

My take: The single most transferable framework in my toolkit. Systems literacy is what separates product thinking from feature thinking. Required before building anything with AI at the core.

Ravi Mehta's framework for connecting company mission, product vision, product strategy, and product roadmap into a coherent hierarchy.

My take: I use this when a roadmap feels disconnected from strategy. It's a forcing function for finding where the chain broke.

Basecamp's free online book on product development. Appetite-based planning, fixed-time cycles, and betting on shaped work instead of managing a backlog.

My take: Even if you don't adopt it wholesale, the "appetite" concept is the best antidote to infinite backlog growth I've found. One idea worth the whole read.

Leopold Aschenbrenner's long-form essay on the trajectory of AI progress and what it implies for the next decade. Dense, opinionated, and essential for context.

My take: Required reading. Not for the predictions but for the analytical frame: treating AI timelines as a planning problem rather than a sci-fi scenario.

Benedict Evans' long-form analysis on technology, media, and where AI fits in the broader arc of the tech industry. Sharp macro perspective, updated regularly.

My take: Best for calibrating your worldview. His annual presentation is the one piece I forward to every PM I work with.

The gold standard for PM-specific tactical content. Data-driven benchmarks, practitioner interviews, and practical how-to guides.

My take: Worth the paid tier. The benchmark database — retention, activation, and growth metrics by stage and category — is invaluable for OKR conversations.

Gergely Orosz's newsletter for senior engineers and engineering managers. Increasingly covers AI's impact on the engineering org.

My take: I read this to stay calibrated on what AI actually looks like inside engineering teams. Essential for PMs working closely with eng.

Daily digest of the most important AI research, product launches, and industry news. Five minutes, high signal-to-noise.

My take: My morning catch-up. Good for staying current without spending an hour in RSS.

Slack community of 10,000+ PMs. Structured channels, AMAs with top practitioners, and peer-to-peer knowledge exchange.

My take: The best PM community I've found for getting experienced input on hard decisions quickly.

Global product management community with events, training, and content. Particularly strong for PM career development and conference learning.

My take: The conferences are worth attending at least once. The hallway conversations are the actual product.