The Growth Podcast

Today’s episode Most companies are still debating whether PMs should ship code. OpenAI is already debating the best ways for PMs to ship code. They’re living in the future. The builder behind a lot of that harness engineering is Ryan Lopopolo. He wrote the OpenAI post on harness engineering and runs a frontier team where PMs, designers, and engineers all ship using the same system. The wild part for me? His PMs shipped around 100K lines of production code. Did they open the IDE? Hell no! Their coding happened through PRDs, tests, docs, and harness rules. The model did the typing. As someone who spent a decade in PM growth roles, I’ve seen how long it takes to move a feature from PRD in a doc to code in prod. For most companies, that latency is weeks. In Ryan’s world, it can be days, and the PM is inside the loop instead of watching from Jira. So I wanted to get to the bottom of this: * What does the harness look like when PMs can ship like that? * How do engineering teams set PMs up so they don’t ship slop? * What changes in the EPD trio when code is cheap, and validation is the bottleneck? That’s today’s episode, and I come with receipts as Ryan goes deep. ---- Check out the conversation on Apple, Spotify, and YouTube. Brought to you by: * Product Faculty - Get $550 off their #1 AI PM Certification with code AAKASH550C7 * Bolt - Ship AI-powered products 10x faster * Customer.io - Send smarter messages using your product data * Ariso - Ship AI agents and features faster, with fewer regressions * Pendo - The #1 software experience management platform ---- * If you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, Relay.app, Magic Patterns, Speechify, Bolt.new and Mobbin - become an annual subscriber ($150), and grab Aakash’s bundle. * If you want access to my AI PM customizations - PM OS, Job Search OS, and Prompt Library - become a founding subscriber ($250). ---- Key Takeaways:1. Code is a liability, not an asset - Every engineering org was built around the assumption that code is expensive to produce, validate, and deploy. Codex inverts this. Code is now the cheapest part of the stack and the constraint moves to how clearly you describe the problem.2. The new constraint is product decisions per week - With code generation effectively free and parallel, the bottleneck is no longer keystrokes. It is the quality of the brief, the clarity of the architectural boundaries, and the speed of verification.3. A billion tokens a day is the new floor - Ryan's claim is that if you are not running this volume you are negligent. The math comes out to roughly $2K to $3K per engineer per month, which is trivial against the headcount cost of human-only execution.4. A single PR can burn 350 million tokens - One refactor that would have taken Ryan three weeks ran on Codex for 60 hours straight across three days. He gave it two prompts total after the initial spec. The output matched what he would have produced himself.5. The harness is the actual product - Codex CLI is the surface. The harness is everything that gets the agent the right context at the right phase. Pre-work, messy middle, and close. Each phase needs different context, different tools, and different verification.6. agents.md is forcibly injected context - This file lives in the repository root and is always loaded into the agent's context. Use it for the operating model and the non-negotiable rules. Everything else gets pulled in dynamically because context is a hard, scarce resource.7. The painted-door technique works inside the codebase - Ryan's team enforces package boundaries so a designer can paint a fake UI on top of stubbed APIs. Real usage signal, no backend cost. This only works because the architecture refuses to permit a ball of mud.8. The PM's PRD can become a shipped PR in one week - In Ryan's setup, the PM wrote a markdown PRD, the team reviewed it in a Monday meeting, and a working feature shipped to customers by the following week with zero PM-to-engineer back-and-forth.9. The Monday morning roadmap starts with legibility - The first move is making the repository legible to the agent. Write the implicit team decisions down in a documentation tree. Use @-mention Codex to keep that tree updated whenever a Slack thread surfaces a new guardrail.10. One agent beats multi-agent handoffs - The lossy friction of agent-to-agent handoffs costs more than it saves. The right answer is one agent with full addressability over design, backend, and frontend, powered by a model good enough to hold the whole task in context. ---- Where to find Ryan Lapopolo * X * LinkedIn * OpenAI Related content Podcasts: * How to Run Evals in Claude Code with Aparna Dhinakaran * How to Build a Full AI Dev Team in Claude Code with Gabor Mayer * This CPO Uses Claude Code to Run His Entire Work Life with Dave Killeen Newsletters: * PM’s Guide to Claude with Pawel Huryn * How to Become a Builder PM with Mahesh Yadav * How to Build a Team OS in Claude Code with Hannah Stulberg ---- PS. Please subscribe on YouTube and follow on Apple & Spotify. It helps! If you want to advertise, email productgrowthppp at gmail. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.news.aakashg.com/subscribe

Today’s episode Many of the smartest AI teams I know are running their evals on Arize. Teams at Uber, Booking.com, Pepsi, and others. It’s become one of the most important skills for PMs. I already had on the CEO of Braintrust, Hamel Husain and Shreya Shankar, and Ankit Shukla. Today I’m adding to this knowledge base on evals with a masterclass on evals in Claude Code. Aparna Dhinakaran is the founder of Arize. She’s also their CPO. And she gives a masterclass in how to run all of your evals through Claude Code. So if you want to do AI evals like the best, like Uber, like Booking.com, check out this episode. For anyone in building in Claude Code, it’s a doozy. If a candidate did this in an interview, Aparna said she would hire them on the spot. ---- Check out the conversation on Apple, Spotify, and YouTube. Brought to you by: * Superhuman - The fastest email experience ever made * Sign up and get 1-month free of Superhuman Mail with my link: superhuman.com/akash (given by brand - Kartik) * Land PM Job - My 12-week AI PM + Job Search Course, first 10 enrollees get a FREE 30-min 1:1 consultation * Vanta - Automate your compliance. Close deals faster * Product Faculty - Get $550 off their #1 AI PM Certification with code AAKASH550C7 * Bolt - Ship AI-powered products 10x faster ---- If you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, Relay.app, Magic Patterns, Speechify, and Mobbin - grab Aakash’s bundle. Do you want to become an AI PM? I’ve created a course for you. Starts soon. ---- Key Takeaways: 1. Trace before you eval - A trace is the full step-by-step playback of what your agent did. Without it, you have no evidence base for evals. Every LLM call, every tool call, every intermediate output needs to be visible before you write a single eval. 2. A span is your unit of evaluation - A span is one discrete step inside a trace. Evals run at the span level, not the trace level. "Did this specific scoring step get the priority right?" is a more useful question than "was the whole run good?" 3. Instrumentation is now a one-command job - Claude Code's instrumentation skills can set up observability for your agent automatically. Arize Phoenix's skill looks at your codebase, identifies the LLM calls and tool calls, and wires them to the tracing layer. No engineering support required. 4. The vibe eval is a draft, not a verdict - An LLM can suggest what your evals should test by looking at your traces. That suggestion will not know your bug-first policy, your comp logic, or your definition of "critical." Treat it as v0 and refine against your actual judgment. 5. When evals fire, two things could be wrong - The agent produced a bad output. Or the eval is miscalibrated. Reading the flagged span yourself is the only way to know which one needs fixing. Both are normal. Both are good news. 6. Evals drift and need regular realignment - Your priorities change. Your bug policy changes. Your product changes. An eval calibrated to last quarter will start misfiring this quarter. Regular alignment to human feedback is maintenance, not a failure. 7. The self-improvement loop is already running at the best teams - Fetch all spans where evals fired. Group by failure category. Propose a specific prompt fix. Review and approve. Ship the new version. This loop runs on a schedule and requires a human at the approval step. 8. Enterprise PMs: start with one internal agent - Not a customer-facing product. An internal tool that takes four hours off your week. Once you have it, you will naturally want to trace it. That is when observability starts to matter to you personally. 9. The context graph is the enterprise unlock - Agents are only as useful as the context they have. Enterprise data lives in silos. The teams breaking through are building unified context layers that give one agent access to CRM, Gong, analytics, GitHub, and Slack. 10. Product taste is still the alpha - Code is cheap now. Shipping speed is table stakes. The PMs who pull ahead are the ones with the sharpest judgment about what to build, and the loops that make their agents better every day. ---- Related content Podcasts: * AI Evals with Hamel Husain and Shreya Shankar * Evals are the new PRD with Ankur Goyal * AI PM Crash Course with Aman Khan Newsletters: * AI Evals for PMs: Everything You Need to Know to Get Started in 2026 * Your Complete AI PM Course & Career Roadmaps * AI PM’s Guide to LLM Judges ---- PS. Please subscribe on YouTube and follow on Apple & Spotify. It helps! This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.news.aakashg.com/subscribe

Today’s episode The market is looking tough for non-technical PMs. Every single week, my comments look exactly the same: brilliant product managers who have the vision, specs, and roadmap in mind, but have zero coding skills. They want to build, and while thousands of technical resources exist online, they make a flawed assumption: that you already know how to code. So when I invited Andre Albuquerque on my podcast, I had to ask him to share his setup. Andre is the founder of Builders Camp, a product school with 4,000+ students across 30 countries, who runs five businesses with Claude Code and has never been a developer. Live on the episode, he built a fully functional product from scratch to show how easily a non-technical PM can go from 0 to 1. He also walked me through CLAUDE.md architecture, custom multi-agent skills, and the bridge between Lovable and Claude Code (which, by the way, not many people are talking about). If you have been putting off Claude Code because it feels too technical or intimidating to set up, this episode is absolutely for you. ---- Check out the conversation on Apple, Spotify, and YouTube. Brought to you by: * Customer.io — * Amplitude — The market-leader in product analytics * Bolt — Ship AI-powered products 10x faster * Arize — Ship AI agents and features faster, with fewer regressions * Product Faculty — Get $550 off their #1 AI PM Certification with code AAKASH550C7 ---- * If you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, Relay.app, Magic Patterns, Speechify, Bolt.new and Mobbin - become an annual subscriber ($150), and grab Aakash’s bundle. * If you want access to my AI PM customizations - PM OS, Job Search OS, and Prompt Library - become a founding subscriber ($250). ---- Key Takeaways: 1. Non-technical PMs are stuck in Jira, Linear, and PowerPoints - Most European PMs are still product owners in disguise, paper-shuffling between strategy and engineering teams. The way out is to actually start building, not to lobby for more autonomy. 2. Start with Lovable on a personal project - Build something for your family, your friends, yourself. The codebase does not need to be pretty. The point is the safety to make mistakes without breaking anything that matters. 3. The Lovable + Claude Code bridge nobody documents - Connect both tools to the same GitHub repo. Write code in Claude Code with all its depth. QA visually in Lovable with its hosted preview. Publish from Lovable's button. The perfect transition layer. 4. Lovable, Cursor, and Vercel are not competitors - Lovable bundles the IDE, the hosting, and the deployment in one product. Vercel exposes the hosting layer so you can run real branches with real preview URLs. Cursor is just an IDE with a generous free tier. 5. Cursor has a free debugging agent - When Claude Code breaks, open a Cursor agent and paste the error. The free agent unsticks you instead of leaving you stuck at step zero. 6. CLAUDE.md is your team's culture - Loaded automatically every session. The first rule should be "for every task, call the PM agent." When you notice yourself fixing the same issue twice, update CLAUDE.md so it never happens again. 7. The PM agent never writes code - The PM orchestrator's only job is to decide which other agent should handle the work. The researcher investigates. The designer proposes. The engineer architects. The implementer writes. 8. Do not copy famous people's skills wholesale - Going on LinkedIn and downloading 100 skills from product celebrities creates more confusion than value. Look at how your real team works. Write each role down as an agent. 9. Fix the agent, not the feature - When something ships wrong, do not patch the output. Identify which agent in the pipeline failed, update its instructions, and run the pipeline again. The next session inherits the fix. 10. The Monday morning move is exactly three steps - Get added as a collaborator on a low-risk repo. Pick the oldest ticket in the backlog. Push a branch and demo by Friday. ---- Related content Podcasts: * Claude Code and agents with Gabor Meyer * n8n, Claude Code, and OpenClaw with Mahesh Yadav * Claude Code with Hannah Stulberg Newsletters: * How to Build a Full AI Dev Team * How to Become a Builder PM * How to build a Team OS in Claude Code P.S. Reply with “CLAUDE” and I’ll send you Andre’s actual CLAUDE.md template. He said we could share it. PS 2. Please subscribe on YouTube and follow on Apple & Spotify. It helps! If you want to advertise, email productgrowthppp at gmail. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.news.aakashg.com/subscribe

Today’s episode When do you use Claude chat vs Cowork vs Code? No one has created a resource that helps you get the most out of the Claude ecosystem. Until now. I’ve brought back Pawel Huryn, the guest behind our most popular episode ever, the Complete Course on AI Product Management. Today we’re covering everything you need to know to get the most out of the Claude Ecosystem. Most PMs open Claude chat. Ask something. Get an answer. Close the tab. Tomorrow, same thing. Fresh context. Zero memory. The PM who tracked Anthropic’s 74 releases in 52 days stopped doing this entirely. He built a system where Claude organizes its own knowledge, extracts its own rules from data, promotes hypotheses when evidence confirms them, and demotes them when it does not. The system improves without him telling it what went wrong. I sat down with Pawel Huryn, creator of the Product Compass newsletter. He has defined 60+ PM skills, built a PM skills marketplace that hit 10,000 GitHub stars, and runs his entire content operation across Cowork, Claude Code, and Dispatch. In this episode, he walks through every screen live. Real files. Real agent workflows. Real self-improving knowledge bases. ---- Check out the conversation on Apple, Spotify, and YouTube. Brought to you by: * Bolt - Ship AI-powered products 10x faster * Amplitude - The market-leader in product analytics * Jira Product Discovery - Plan with purpose, ship with confidence * Product Faculty - Get $550 off their #1 AI PM Certification with code AAKASH550C7 * Land PM Job - 12-week experience to master getting a PM job ---- If you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, Relay.app, Magic Patterns, Speechify, and Mobbin - grab Aakash’s bundle. I’m accepting applications for my third LandPMJob cohort. Join Me. ---- Key Takeaways: 1. Stop using Claude Chat as your default. Cowork accesses real files, connects to Gmail and Slack via MCP, and runs parallel sub-agents. Chat does none of this. 2. Skills are the highest ROI investment. Install marketplace baselines, iterate 5-6 times with specific feedback, and Claude rewrites from first principles until 99% accuracy. 3. Progressive disclosure keeps context clean. Agent reads skill names and descriptions first. Loads full instructions only when the task matches. Hundreds of skills, minimal overhead. 4. Your CLAUDE.md should route, not store. Project structure and pointers only. Domain knowledge lives in separate files the agent loads on demand. 5. Build self-improving knowledge with three types. Rules are confirmed and applied by default. Hypotheses are tracked with evidence. Rejected patterns are kept to avoid retesting. 6. The three-line self-improving prompt works for any domain. Review rules before starting. Apply confirmed rules. Update after feedback. Testing, marketing, strategy, whatever. 7. Claude Code adds explorer view, hooks, subagents, and local MCP scoping. PMs need it once their system grows past 50 files. 8. Every Product Compass infographic was built in Claude Code. HTML generation, component library, iteration through conversation, PNG export. Zero code written by the human. 9. Use Agent Browser from Vercel instead of Chrome MCP. Chrome MCP screenshots every 0.5s and burns $100/hr. Agent Browser uses headless mode and is token-efficient. 10. Dispatch lets you run multiple tasks from your phone. Start an infographic, check emails, analyze competitors. Each runs as a separate thread. Your system works while you live. ---- Where to find Pawel Huryn * LinkedIn * Product Compass Newsletter * PM Skills Marketplace on GitHub * [Quadathon - starts May 9th](VERIFY - Quadathon URL) Related content Podcasts: * n8n Masterclass with Pawel Huryn * Claude Code PM OS with Dave Killeen * Claude Code Team OS with Carl Vellotti Newsletters: * The complete Claude Cowork guide * How to use Claude Code like a pro * Build your PM operating system ---- PS. Please subscribe on YouTube and follow on Apple & Spotify. It helps! If you want to advertise, email productgrowthppp at gmail. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.news.aakashg.com/subscribe