A few weeks ago, Stefan and I debated whether PMs need to become builders. We agreed. Then we had a heated WhatsApp argument and realized we disagree on something fundamental: Will agents make the PM role obsolete, or will they create the golden age of product management?
Stefan believes the 100-person software company will become a one-person company within three to five years. I think the work changes, but the core skill, understanding humans and their problems, stays. We decided to argue it out on the podcast.
Here is what we landed on.
The case for extinction
Stefan joined an AI-first startup three months ago. Everyone there uses AI for 95% of their work. The compounding effect is real: when the entire team operates this way, speed multiplies. When only one person does it, they are bottlenecked by everyone else.
His experience with OpenClaw pushed his thinking further. He gave it tasks he assumed it could not complete. It found creative solutions on its own. Research that used to take him hours, such as cross-referencing analytics tools, downloading CSVs, running analysis, now takes a single prompt.
His conclusion: if a solo developer can ship a working product from a WhatsApp chat, most of the coordination work PMs do will disappear. Companies that don’t adopt this speed will lose to companies that do. And those fast companies are small by definition.
The case for survival
I recently built DocReady, an app that helps Austrian doctors categorize tax documents. I built the entire product solo. Website, iOS app, web app, ads. The code part was impressive, sure. But it solved maybe 5% of the problem.
The hard part is acquisition and activation. How do I let doctors know this exists? How do I make them use a tax preparation app regularly (for a task they hate)?
One of my beta users told me she didn’t see why she should use DocReady instead of her Google Drive. She already had a system. This single conversation changed my product strategy. I realized I needed to extract structured data from documents, not just store them. That insight opens the door to Excel exports, spending dashboards, reminders, summaries.
No agent would have made that strategic observation from a user interview. Not yet.
Where we actually agree
The PM role as defined in 2020 is dead. The person who writes tickets, manages backlogs, and coordinates standups will be replaced. That work is pure overhead in a world where agents handle execution.
But the core skill, talking to customers, understanding their real problems, and translating that into product decisions, is more valuable than ever. Execution is cheap now. Context is expensive.
We also agree that customer success managers, implementation specialists, and support agents are equally well-equipped for this new world. They’ve had direct customer conversations for years. They just lacked the power to act on what they learned. Agents give them that power.
The PM title might survive. The PM job description won’t.
The golden age of tiny companies
Stefan made a point that stuck with me. If a single person can build and ship a hyper-specialized agent, say, one that handles support for Shopify stores selling sneakers, they could replace a company’s entire support team and their SaaS stack.
A company spending 100k a year on humans and tools might happily pay 50k for an agent that handles everything. That is SaaS, just built by one person instead of a hundred.
These companies won’t attract VC money. A 10 million ARR ceiling is boring for investors. But for a solo builder, 10 million a year is life-changing. And there will be thousands of these niche opportunities.
The interface changes everything
Stefan now does half his work through a WhatsApp chat with his agent. He vibe-codes, debugs deployments, and runs research. All from his phone. He took a three-hour bath and felt more productive than any day at his desk.
I use multiple AI subscriptions for different tasks. I brainstorm complex strategy through my phone. When I bought a house recently, I resolved hundreds of questions through an LLM. I felt more secure and informed than I ever expected.
The medium has changed. The screen-and-keyboard era is ending for many types of knowledge work. Agent monitoring might happen from a chat interface on your phone. People might not need to sit in front of a computer for their entire workday anymore.
What PMs should do right now
Forget most product frameworks. They were optimized for a world that no longer exists. Rebuilding an entire app was a crazy idea a year ago. Now it’s not anymore.
Stay at the frontier. If you know what the latest agent tools can do, you are already ahead of most people in tech. If you tried vibe coding once in early 2024 and gave up, try again. The gap between then and now is enormous.
Be the crazy person. Take risks. The likelihood of a positive outcome is higher than it was six months ago. The worst thing you can do is assume that the way you worked for the past ten years will keep working.
And most importantly: be curious. This is the first technology shift in years that is not just hype. Everyone can feel how their job is changing. The people who lean into that change will build the next generation of products.
It has never been this exciting to be in product. Even if the job title changes.
Links
Link to Podcast Episode
* 🎥 YouTube
* 🎧 Spotify
* 🎧 Apple Music
In case you want to reach out, please do so on LinkedIn:
* 🔥 Follow Hotfix: https://pod.link/the-hotfix-podcast
* 🔗 Follow Christoph: https://www.linkedin.com/in/christophbodenstein/
* 🔗 Follow Stefan: https://www.linkedin.com/in/stefanpernek/
