Mike Bal

@cotyledonlab @levelsio I appreciate you saying this because I was trying to find the words. Perplexity barely made a ripple with computer and almost every builder I know stopped using it months ago.

@levelsio MCP isn't dead it's a necessary protocol in regulated areas. What is hilarious is you taking some move by Perplexity as proof of anything 🤣 That's like Yahoo weighing in on the state of internet search

Thank god MCP is dead Just as useless of an idea as LLMs.txt was It's all dumb abstractions that AI doesn't need because AI's are as smart as humans so they can just use what was already there which is APIs

couldn’t have said it better.

@tetrisgm @aakashgupta @davidsbridal I think that’s a fake structure resulting from a lack of experience. Every pm should be ai enabled/ native in terms of how they work. Not every pm has had a chance to build with AI… but they will.

@aakashgupta @radres @TheMikeBal @davidsbridal Funnily enough I think you're selling the discipline short! AI PM is completely distinct from PM, in the same way AI Eng is from regular eng. I do think 90% of what you post isn't something that applies to regular PM orgs. But it would be for an AI org.

AI native PMs are lapping those who just use LLMs. Here’s how to build an AI PM operating system with @TheMikeBal, Head of AI and PM @davidsbridal: 8:41 - The OS mindset 15:22 - Key MCP integrations 28:26 - Demo: 30s design validation 48:52 - Manus vs Claude vs ChatGPT

@briancheong @aakashgupta Oddly enough I value the ones connected to the tools with the worst ux. For Claude teams we have Microsoft 365 and not looking at outlook and share point helps keep my blood pressure down.

@aakashgupta @TheMikeBal PM workflows are a perfect MCP use case because the data is spread across 10 tools and nobody wants to copy-paste between them. curious which integrations gave you the most leverage vs which ones were more novelty.

How @TheMikeBal uses Cursor and MCP to build an AI operating system that connects to all his tools Most PMs juggle dozens of disconnected tools.

Doesn’t look like @beehiiv sharing works…

Leading%20Product%20grew%20by%200%20subscribers%20today%20on%20@beehiiv.%20Subscribe%20below%20👇🏽%20https://www.leadingproduct.link/subscribe 0 new subscribers, yay!

Launching Intelligent Data Charting in @napkin_ai 🚀 Turn your data into visual stories – instantly. Now, you can: ✅ Upload messy or structured data and get multiple chart options ✅ Generate everything from simple bar charts to complex Sankey or Gantt diagrams ✅ Use annotations to call out insights and create visuals that explain, not just display data No more picking charts first and forcing data to fit. Napkin's AI reads your data, understands relationships, and shows you what works. Try it now at napkin.ai #napkinai #visualai #aitools #dataviz #ai

In the past few weeks I’ve been cold emailed by a 15-year-old and 16-year-old offering significant technical insights and contributions. We will see within our lifetime the rise of young supergeniuses that far surpass previous generations in both intellect and demonstrable accomplishments. With AI, anyone can self-teach any concept, science, discipline. Agents can implement and give the individual the power of a large team. It’s gonna be fun to watch the dramatic acceleration of human potential.

@Govikavaturi @aakashgupta Mcp isn’t a a legacy process. Hs e you found a better way to connect ai in a coherent way across shared tools?

@aakashgupta @TheMikeBal Today's enterprise processes are so old school. They have truly become legacy and irrelevant at this point. MCP is not the solution, those tools are the worst.

@aakashgupta Small but mighty!

I don't think most PMs realize how quickly "can you build with AI" is becoming a core performance review metric. @TheMikeBal's team is one quarter the size of comparable teams at David's Bridal. They ship at the same velocity. Sometimes faster. Here's what changed. Mike doesn't open JIRA to check if tickets closed. He asks from Cursor. "Check if this issue on this project was completed." Mike doesn't manually compare Figma designs to PRDs. He prompts "Find my Confluence doc about Feature X, load this Figma design, compare them." Mike doesn't export Clarity analytics to Excel and build pivot tables. He uploads the CSV to Cursor and asks "analyze drop-offs and create visualizations." Every one of these tasks used to take 30 minutes to 2 hours. Now they take 30 seconds to 10 minutes. The compounding effect is insane. At annual reviews, managers are starting to ask: How are you using AI in your workflow? What manual work have you eliminated? Can you show velocity improvements? The PMs who answer "I use ChatGPT for brainstorming" are getting mid ratings. The PMs who answer "I built a composable operating system that connects all my tools, I operate from Cursor and Claude Desktop, and I automated our competitor monitoring workflow" are getting promoted. This podcast covers the exact stack and workflows of an AI-native PM operating at that level. The tech stack is cheap. Claude Pro ($20/month), Cursor ($20/month), maybe Manus ($20/month) for research. $60/month total. The skill stack is what matters. Knowing when to use MCP connectors versus developer MCPs. Understanding how to structure prompts for design validation. Recognizing when to build versus when to buy. Most PMs are being evaluated on this in 2026 performance reviews and don't know it yet.

@yshaikh2001 @aakashgupta PMs thinking in prompts should actually reflect systems thinking IMO. Engineers are effectively prompting themselves as they think through the approach.

@aakashgupta @TheMikeBal If PMs think in prompts, engineers still think in systems. Curious who actually owns the hard tradeoffs.

Most PMs still think in tasks. Write a PRD. Update JIRA. Review designs. AI native PMs think differently. They think in prompts. .@TheMikeBal calls it "thinking in prompts" and admits it sounded silly at first. But once you integrate AI into daily work, it becomes natural. Instead of "I need to write a PRD," they think: "What do I need to get done? What are the steps? What are the best tools?" They treat AI as an extension of themselves. Breaking down tasks into instructions and steps. "If you're thinking about AI as like an extension of yourself, you're thinking like, what are the instructions? What are the steps? And if you're reflective, like you probably do that to yourself at some level anyway, you have your inner dialogue, right?" The shift is from task-oriented to process-oriented thinking. It's how you already think internally. AI native PMs just make that process explicit. Lesson: AI native PMs don't just use AI tools. They fundamentally restructure how they think about work, breaking every task into steps and instructions that AI can execute.

@Rich_Hillier @aakashgupta I think having systems that are built to talk to other systems is important but there are always work arounds

@aakashgupta @TheMikeBal Interesting approach to AI native PMs but the question I have is does this still work when the systems they work in aren’t aligned with this?

@foundrceo @aakashgupta Deep Focus > Busy Work

@aakashgupta @TheMikeBal this is so true, context switching is a killer. centralized systems are the move.

The average PM switches between 20+ tools daily. Jira for tickets. GitHub for code. Confluence for docs. Figma for designs. Gmail for communication. Each switch costs time. How AI native PMs eliminate context switching by building centralized operating systems with @TheMikeBal

@VibeCoderOfek @OrenE_IL @aakashgupta @davidsbridal @cursor_ai @claudeai Usually I have them working on different things or i have one building while one is testing depending on if their is a front end element or not. The new Cursor visual diff is pretty slick.

@TheMikeBal @OrenE_IL @aakashgupta @davidsbridal @cursor_ai @claudeai tandem setup is clutch cursor for inline flow, claude for big questions. i've been doing similar since desktop claude dropped; saves jumping tabs. how do you split tasks between the two without context bleed?

@anjanps5 @aakashgupta I would take it a step further and ask them to explain what tools they would use AND what prompts. See if they can think in systems, identify the most effective path, and write clearly.

@aakashgupta @TheMikeBal I have been asking for our PM candidates in interviews to submit their prompts along answers to any take home questions. It’s been very helpful to identify PMs who are using AI just as tools vs those that are fundamentally changing how they think

I use @cursor_ai + MCP to investigate/debug issues, understand diffrent code bases and applications, provide clear direction for implementation and to build cool sh*t.

@OrenE_IL @aakashgupta @davidsbridal I actually like building and developing in cursor because I can run @cursor_ai agent and @claudeai in tandem. When we recorded this Claude code wasn’t in the desktop app. So I also sometimes just ask questions about our code base

@aakashgupta @TheMikeBal @davidsbridal Need to look deeper - as the 2 tools (Cursor & Claude Desktop) are similar so why use both instead of 1?

What makes an AI native PM? .@TheMikeBal on thinking in prompts and building an operating system mindset Most PMs still work the old way. They need to check if a bug fix shipped.

Are AI-native PMs leaving LLM-only users in the dust? Yes and here's your One hour Masterclass to build an AI PM operating system, with expert tips from @TheMikeBal, Head of AI and PM at @davidsbridal.

@giudegio Thanks for sharing this btw!

Hands-on Lessons in Vibe Coding How to avoid disappointment, heartbreak, and embarrassment. leadingproduct.link/p/hands-on-les… via @themikebal