The new product rule says collaboration destroys shipping speed

This article summarizes an episode of Aakash Gupta’s video series featuring Andre Albuquerque, founder of Builders Camp.

Andre Albuquerque, founder of Builders Camp / Photo credit: Andre Albuquerque

The modern tech industry is held back by rigid job titles and the trap of constant collaboration. Andre Albuquerque, founder of the product management training hub Builders Camp, argues that waiting on engineering departments to execute tasks creates massive bottlenecks.

To unlock true momentum, he proposes a new approach: stop collaborating during the execution phase, abandon traditional role silos, and let non-technical workers wield AI tools to push code directly to production.

Working alone using a simple set of tools

To move from testing ideas to actually launching them, Albuquerque argues that non-technical workers must seize total control over the codebase using a clear, four-step AI toolkit:

• Level 1: makes starting easy. Lovable is an AI tool that handles hosting, databases, and logins so a product manager can learn by launching a personal app where mistakes cost nothing and can be undone.
• Level 2 :connects visual screens and text code in one place. GitHub acts as the main folder. Claude Code edits the files with more control, while Lovable remains the visual checker and publish button.
• Level 3: adds systems for safe testing. Vercel turns each code version into a temporary web link to test ideas. Cursor becomes the favored code editor because it shows files and chat logs clearly.
• Level 4: adds AI assistants and a rulebook to manage results. Claude acts as a saved memory file. AI assistants divide research, design, planning, and coding so writing prompts becomes an organized process.

“If you look at a lot of companies that have non-technical PMs in teams,” Albuquerque argues, “they’re bureaucrats. They’re stuck in Jira, they’re stuck in Linear, they’re stuck in PowerPoints. They’re not actually building, they’re not pushing code, they’re dependent on their technical teams.”

He explains that syncing these tools allows non-technical workers to push flexible backend code while still tracking the visual evolution of their product in real time without constantly switching platforms.

Overcoming technical intimidation
Pointing to a screen showing live web links, Albuquerque notes that Vercel generates temporary preview branches for new ideas before merging any finalized changes to the main, user-facing product.

“When you’re non-technical working on some technical stuff,” Albuquerque says, “you want to find the comfort areas. It’s going to make you feel less fear and be more focused on doing stuff.”

Protecting the code with required checks

Empowering non-technical workers to write code introduces new risks. Treating AI mistakes as a system failure requires building rigid guardrails that catch errors long before they reach users.

Albuquerque warns, “The biggest fear a lot of people have is like, ‘Oh vibe coded code is bad code.’ It happens to a lot of people, especially when you’re non-technical. You don’t know what you’re asking. It’s very easy with very simple prompts to be creating a lot of slop.”

He shifts the burden of quality control entirely away from individual engineers and onto automated systems. He argues that once a great infrastructure is in place, a worker’s lack of technical skill stops mattering, as long as the security team does a flawless job protecting the production repository.

“When something ships wrong, your instinct would be ‘fix these things on the feature.’ That’s not the right mindset,” he says. “Identify what in the pipeline failed and led to a bad result, then change the agent, the skill, or your [Claude.md].”

Company rules control the speed of work