The Next Decade Belongs to Product Managers

AI is collapsing the cost of execution while inflating the premium on direction. The scarce resource is no longer coding hours. It is clarity, sequencing, and outcome ownership.

That is the core PM skillset. And it gets amplified when PMs are technical enough to prototype, instrument, and ship.

The Signals Are Clear

Software is being re-priced as a database + rails layer

I wrote about this in Most SaaS is dying. The unbundling is not uniform. Generic workflow SaaS is vulnerable. Regulated rails, deep incumbency, network effects, and hard switching costs persist.

WSJ echoes the market version: AI may not kill software, but it can puncture growth narratives, compress multiples, and force a shift away from seat-based expansion. On top of that, just today we saw many more articles from WSJ, with claimed "winners" of the AI race. Between Google leaning hard into its AI winner Status, and Open AI unveiling a product for building AI Co-Workers, and Anthropic's Clawds, and "Long Running Agents", it might be hard to tell who the winner is. I think it's Product Managers, Product Designers, and Solutions Architects who are winning the most.

Frankly, value migrates upward into the workflow orchestration layer. Where PMs already live.

Long-running agents compress the build-test-learn loop

WSJ describes org-level workflow compression: prototypes replacing persuasion docs, faster merging of internal data with web research and customer feedback, role boundaries blurring. But your CEO likely can't do this on their own, unless they're technical.

This is my From Idea to MVP playbook made inevitable. PRD → prototype → instrument → iterate. AI eliminates the lag between each step.

A Meta PM with no technical background says "vibe coding" gave him superpowers. He went from coordinator to builder when AI collapsed implementation costs. Again, a builder, with a business mindset, became super-powered. But it doesn't mean just "anyone" can do it.

Capital markets are repricing right now

WSJ's KKR piece shows investor-lens signals. Public and private market players are explicitly discussing AI-driven disruption risk to software investments.

The people paid to price risk are repricing the old moats. They are starting to take this new modernization more seriously, and it's about time. But don't go throwing out your engineers just yet.

Why Product Managers Win This Era

PMs are built for the new bottlenecks

AI reduces the friction of making. The constraint shifts to:

• Problem selection (what is worth doing)
• Constraint navigation (security, compliance, data shape, integration reality)
• Outcome definition (what done means, how we measure it)
• Sequencing (what to do now vs later)
• Tradeoff arbitration (user vs business vs technical approach)
• Adoption design (getting humans and orgs to actually use the thing)

Those are PM muscles. Not just engineering muscles.

Technical PMs are the apex predator

Because they can:

• Prototype without waiting
• Sanity-check architecture claims
• Instrument success metrics
• Ship increments that de-risk the roadmap

McKinsey calls this "the agentic organization" where cycles compress, decisions accelerate, and leaders shift from reporting mechanics to interpreting signals.

That is the environment where PM-style outcome ownership becomes central.

"Unstoppable" Has One Condition

My Collaboration is a myth argument is a selection mechanism. Winners will be the ones who can produce shipped artifacts. Not vibes, meetings, or pitch energy.

So the condition is: PMs who can ship become disproportionately powerful. PMs who only coordinate humans in slow cycles get automated first.

Reforge frames this cleanly: AI automates low-value PM work and amplifies differentiators. PM does not disappear. It moves up the stack into strategy, judgment, sequencing, and outcomes.

The New Operating Model

Replace requirements with tests

A PRD becomes a hypothesis sheet plus instrumentation plan. Not a novel.

Make the agent your default first engineer

Use AI to generate prototypes, integration spikes, migration scripts, dashboards. Then review like a staff engineer.

Run roadmaps as outcome portfolios

AI makes feature throughput cheap. Strategy becomes the moat.

Own the workflow layer

A16z notes the shift to outcome-based pricing. When pricing moves from seats to usage and outcomes, organizations need tighter definition of value delivered and measurement.

That is PM terrain. Especially for technical PMs who can instrument and iterate.

The Steelman Counterpoint Still Supports PM Dominance

Even if AI dents growth more than existence, someone still has to redesign products and business models around:

• Usage and outcome pricing
• Agent layers
• New UI/UX patterns (delegation, review, auditability)
• Data and integration defensibility

ITPro calls this "Outcome as Agentic Solution" - the shift from tool access to vendor/agent delivers outcomes. The governance and accountability implications are PM-shaped work.

That someone is still product leadership. Just closer to systems and outcomes than ever.

We have been sitting in the difficult middle between business needs, user needs, and technical approach for years. We are the movers and shakers who have been begging teams to hear each other out. We know business, discovery, and delivery.

Now we are unstoppable.

For the love of God please don't listen to any non-technical "AI Consultants"

They are grifters, who feel they have been unblocked by AI. They shit out lazy prototypes using Loveable, but they have no clue how the tech works. They don't know compliance, security, safety, or billing. They don't know UX, or roadmapping, or measuring results of OKRs. They just repeat the jargon they hear on LinkedIn and crap out junk. They want to trick you into thinking the market is changing in their favor. It is not. Grifters have always been bandwagon hoppers, but you don't want to listen to these people. You need your technical lead, your design lead, and your analyst using AI. You yourself, as an executive could be using AI. But do not throw the baby out with the bathwater over a Loveable prototype. It's not production ready, and you will end up licking wounds, or worse, on the front page of the WSJ for yeeting out a prototype and calling it "ready".

Listen to your experts, while you all accelerate.