It's Over For Paid Media

What It Covers

Neil Patel and Eric Siu react to Meta's $2B acquisition of Manus AI and debate whether paid media is dying. They cover AI's role in ad automation, the convergence of marketing roles, Block's 40% staff cuts, and a new pay-for-performance AI deployment business model Eric is building.

Key Questions Answered

• •Paid Media Role Convergence: Button-pushing media buyers face elimination, but those who expand into AI-assisted creative production remain valuable. The shift requires learning creative strategy at scale — not just ad platform mechanics. Taylor Holiday's framing captures this: head of growth, media buyer, and creative strategist converge into a single "profit engineer" role.
• •Creative Over Automation: The real competitive edge in paid media is not how ads are bought but what creative is tested. AI accelerates experimentation cycles, but the winning formula still requires human-generated, untested creative concepts. Once a high-converting angle is found, the strategy is to double down aggressively rather than continue broad testing.
• •Platform AI Misalignment: Meta and Google AI ad tools optimize for platform revenue, not advertiser ROI. A concrete example: unconstrained AI frequently over-allocates budget to Gmail ads, which convert poorly. Keeping humans in the loop to override platform defaults protects against this systematic conflict of interest between advertiser goals and platform incentives.
• •Hiring for Obsessive Attention: Filter candidates by assigning detail-dependent tasks without providing all necessary resources — such as a design brief without a style guide — then observe whether they request missing assets. Candidates who self-direct, match brand specs unprompted, and arrive at interviews having read company books demonstrate the bias-to-action trait that predicts performance.
• •Pay-for-Performance AI Deployment: Building reusable AI agent scaffolding — outbound email, phone call generation, end-to-end sales fulfillment — makes pay-for-performance models viable where they previously were not. Hard people costs made performance deals risky before; agent infrastructure reduces that cost base, enabling the model to scale across multiple verticals with the same underlying build.