Presenters

Source

The Vibe Coding Revolution: How AI is Reshaping Product Management in 2026 and Beyond 🚀

Remember the days of sticky notes, massive spreadsheets, and endless planning hell? If you’re a Product Manager who’s been around since 2024 (or earlier!), you likely do. Simon Kubica, co-founder and CEO of Alloy, takes us on a journey through the dramatic evolution of product development, culminating in the vibe coding revolution that’s transforming how we build products today.

From Sticky Notes to “Purple” Apps: A Brief History of Product Planning 🗓️

The “Stone Age” (2024 and Prior): The Reign of Spreadsheets and Sticky Notes 🗿

Simon paints a vivid picture of the past: Product Managers huddled in rooms, armed with sticky notes to map out roadmaps. The reality? Endless 500-line spreadsheets, tangled dependencies, PRDs galore, and the frustrating realization that only a fraction of the planned work made it to completion by the end of the quarter. It was a slow, often opaque process.

The “Dial-Up Era” (2025): The Rise of App Builders ☎️

Then came the app builder tools of 2025. While fun, fast, and often sporting a distinctive purple hue, they were like the dial-up internet of product management. They allowed for great communication of ideas and interactions, but they fell short of truly accelerating delivery. Prototyping felt disconnected from the actual product’s interface, and the core challenges of the old way persisted.

The Fireworks Moment: Vibe Coding Explodes in 2026! ✨

The last month or two have seen a monumental shift, a true “fireworks moment” for non-technical PMs and product leaders. The launch of Opus 4.6 in January has been a game-changer. Simon highlights the incredible capabilities of these new Anthropic models, enabling prototyping directly on your real application in the cloud.

Key Breakthroughs:

  • Asynchronous Agents: These AI agents can run in the background, triggered by Slack threads or customer requests.
  • From Prototype to Pull Request: The output isn’t just a static prototype; it can be a completed change ready for engineering review via a pull request.
  • Dial-Up to Broadband: 2026 marks the leap from clunky app builders to powerful tools that empower all product leaders to become full-blown builders.

The Accelerating Velocity: How AI is Supercharging Product Development ⚡

The evolution of these AI models isn’t just about better responses; it’s about their longevity.

  • Extended Run Times: Models like Opus 4.6 can now run for up to 15 hours (December to February saw this jump from 4.5 to 4.6). Simon shares his practice of kicking off changes overnight, allowing AI agents to refactor and improve the entire product while he sleeps.
  • Vertical Growth: This capability is on a “completely vertical growth curve,” projected to reach 30, 40, or even 50 hours in the near future.

This trend is reflected in adoption rates. Alloy, Simon’s company, is seeing “usage go vertical.” The sentiment is echoed by leaders like Toby at Shopify, who emphasizes that not leveraging these AI product agents means “leaving a bunch of value on the table.” The pace of change is so rapid that talks like this are constantly being updated!

Vibe Coding in Your Company: What 2026 Looks Like 🏢

So, how does this translate into your enterprise?

  • Product Managers as Builders: Tools like Cursor and environments like those at RAMP are empowering PMs and designers to directly contribute.
  • Internal “Background Agents”: Companies like RAMP are building their own AI agents. These systems allow everyone to make changes – from prototyping ideas and updating copy to adding new features.
  • From 0% to 70% PR Merged: RAMP’s internal agent has seen an astonishing 70% of its pull requests merged in just the last two to three months, a testament to the efficiency and quality of AI-driven development.

Is this just for startups? Absolutely not! Alloy sees adoption from startups all the way to Fortune 500 companies. The core shift is the product leader becoming a builder, a transformation accessible to all.

As Naval put it, “Vibe coding is the new product management,” and it represents the biggest workflow change in the history of PMing. 2026 is the year of the AI cloud playground, fundamentally altering how we build and what we do as PMs.

Best-in-Class Performance: Key Principles for AI Cloud Playgrounds 💎

To truly leverage these advancements, Simon outlines three core principles:

  1. Codebase Context: Your Single Source of Truth 💾

    • Moving Beyond Isolation: Instead of isolated systems, AI models should run on your product’s actual codebase.
    • Instant Pull Requests: This enables one-click pull requests for changes.
    • Leveraging Design Systems: AI can now utilize your existing design system, bridging the gap between Figma and your live product. By plugging codebase context into your cloud playground, AI can leverage years of engineering work.

  2. Leveraging SaaS Systems: Agents as Colleagues 🤝

    • The Mistake of Isolation: A common enterprise pitfall is spinning up agents without access to essential IT tools like Slack, Jira, or Google Workspace.
    • Self-Serve Solutions: Empowering agents with access to these tools allows them to self-serve.
    • Real-World Examples: Imagine an agent checking a Snowflake database for onboarding drop-off insights or pulling context from a Slack thread to fix a customer-reported bug. Without this access, agents are as limited as a human at a desk with no internet.

  3. Cloud Playgrounds: Collaboration Over Localhost ☁️

    • Beyond Local Development: Tools like Cloud Code and Cursor are powerful but can limit collaboration. Prototypes on a local machine are hard to share.
    • Unlocking Visibility: Cloud playgrounds make prototypes visible to the entire company, fostering human-to-human collaboration and unlocking the models’ full potential. Notion’s CEO shared that prototypes are accessible to everyone, leading to transformative insights.
    • The “Magic Superpower”: The key is to connect to your codebase, SaaS tools, and the rest of the company for truly magical performance and product velocity.

Three Ways to Evolve Your Product with AI Superpowers 💪

Simon then dives into practical applications:

1. Instantly Exploring Product Ideas 💡

The Old Way: Writing documents, waiting for reviews, chasing design resources. The 2025 Way: Granting codebase access and setting up local dev environments. The 2026 Vibe Coding Way: Simply dropping in a prompt!

Demo: Whiteboard View in Linear ✍️ Simon demonstrates how he, as a hypothetical PM for Linear, can instantly request a whiteboard view for issues.

  • Capture Context: The Alloy extension captures the current page’s context.
  • Cloud Playground: This context is loaded into the cloud playground, allowing exploration of ideas with full interactivity, including hover states and drag-and-drop functionality.
  • Prompting for Change: A prompt like “Let’s update this to be a whiteboard view instead of a list” is submitted.
  • Rapid Iteration: Within minutes, Alloy returns a functional whiteboard view, ready for interaction and further refinement.
  • Seamless Sharing: This prototype can be shared with a single click, made available to the company, or even pushed as a pull request to GitHub for engineering review. This transforms a multi-week process into a matter of minutes.

2. Visualizing Customer Feature Requests 📊

The Old Way: Feature requests languish in inboxes, met with “it’s on the roadmap” and eventual delays. The New Vibe Coding Way: Building features live with customers.

  • On-Call Collaboration: During a customer call, Simon types requests into Alloy. He can then share his screen, showing the customer a visualized solution, generating immense loyalty and admiration.
  • Automated Overnight Processing: Feature requests coming in from support systems at 3 AM can be automatically passed to the cloud playground. By 3:05 AM, AI agents have built a preview of the feature, ready for review upon waking.
  • Approving and Merging: For bug reports or feature requests related to the codebase, these can run autonomously, with humans simply approving and merging changes at “unbelievable velocity.”

Demo: Enriching Pipeline Cards 🃏 A customer requests more detailed metadata on pipeline cards.

  • Support Ticket to Playground: A ticket requesting more granularity is fed into the cloud playground.
  • AI Design Process: The agent enriches the cards with customer logos, contact details, and other metadata.
  • Human Review and Merge: The human reviewer can then share the prototype, confirm with the customer, and merge the change via GitHub/GitLab/Bitbucket. This streamlines the journey from ticket to implemented feature.

3. Quarterly Plan to Quarterly Playground 📈

The Old Way: Static spreadsheets and Jira boards. The New Vibe Coding Way: Turning plans into interactive prototypes.

  • Plug and Play: Simply connect your codebase, add your SaaS integrations (MCPs), and transform your roadmaps and product discovery into tangible, interactive prototypes.
  • From Insight to Action: This dramatically accelerates the path from understanding user needs to implementing solutions.

Take the Leap: Embrace the Vibe Coding Future! ✨

Simon’s message is clear: “It’s getting so easy, I think. Just do it.” Bring these capabilities to your teams, give vibe coding a try. It’s permissionless and easy to start, regardless of the tools you choose. The product management workflow is undergoing a seismic shift, and those who embrace it will lead the charge into an exciting, AI-augmented future. We can expect even more “crazy changes” next year!

Appendix