Decoder with Nilay Patel

The DoorDash Problem: How AI browsers are a huge threat to Amazon

November 20, 2025

Key Takeaways Copied to clipboard!

  • The "DoorDash Problem" describes the existential threat AI agents pose to App Store-era service providers (like DoorDash, Uber, Airbnb) by abstracting the customer relationship and commoditizing the service provider into a mere database. 
  • The lawsuit between Amazon and Perplexity over Perplexity's Comet browser using an AI shopping agent marks the first major public battleground in the emerging war over who controls web browsing and economic experiences. 
  • While many CEOs of service companies (Lyft, ZocDoc, TaskRabbit) expressed confidence in their brand loyalty or operational complexity as a defense against AI agents, Amazon appears uniquely threatened due to its massive advertising revenue and reliance on being the default shopping destination. 

Segments

Defining the DoorDash Problem
Copied to clipboard!
(00:00:35)
  • Key Takeaway: The DoorDash Problem is the potential loss of customer relationship value for service providers when AI agents mediate transactions, eliminating monetization avenues like ads and loyalty programs.
  • Summary: The problem centers on AI interfaces sitting between service providers (like DoorDash) and the user, allowing agents to order services without engaging with the provider’s app or website. This bypasses crucial monetization layers such as user reviews, upsells, loyalty programs, and direct advertising. Companies from the App Store era, including Uber, Lyft, and Airbnb, face this threat to their direct customer ownership.
Amazon vs. Perplexity Lawsuit
Copied to clipboard!
(00:02:19)
  • Key Takeaway: Amazon initiated a lawsuit against Perplexity to block its AI browser Comet from shopping on Amazon.com, framing it as the first major front in the AI browser war.
  • Summary: Amazon’s lawsuit claims Perplexity violated the Computer Fraud and Abuse Act by using its AI shopping agent to make purchases, despite requests to stop. Perplexity countered by calling the action bullying and arguing that user agents should have the same permissions as human users. Amazon’s statement explicitly compares the situation to food delivery apps and online travel agencies, highlighting the core issue of respecting service provider participation decisions.
CEO Reactions to AI Disruption
Copied to clipboard!
(00:11:08)
  • Key Takeaway: Most service CEOs interviewed, unlike Amazon, expressed relative calm or confidence, believing strong brand relationships or operational complexity will protect them from full agent displacement.
  • Summary: Lyft’s CEO suggested customers would prefer branded services over ‘rando’ pickups, while ZocDoc’s CEO cited their 20-year lead in handling complex real-world edge cases as a competitive moat. TaskRabbit’s CEO dared Apple to build a competing service, emphasizing their cultivated network strength. This contrasts sharply with Amazon’s aggressive legal action.
Uber CEO’s Economic Strategy
Copied to clipboard!
(00:15:05)
  • Key Takeaway: Uber’s CEO advocates for embracing AI agents initially by charging zero to measure experience and incrementality before imposing significant take rates on cannibalistic agents.
  • Summary: Uber’s philosophy is to prioritize testing the user experience with AI agents before settling on economics, contrasting with companies that try to figure out the margin first. If an agent brings incremental business, a take rate (5% to 20%) might apply; if it is purely cannibalistic, the charge would be high. This approach mirrors Uber’s historical strategy of building the business before optimizing profit margins.
Amazon’s Unique Vulnerability
Copied to clipboard!
(00:23:48)
  • Key Takeaway: Amazon is reacting most aggressively because AI agents threaten its massive, high-margin advertising business and the stickiness of the Prime subscription, potentially reducing its retail offering to a commodity.
  • Summary: Amazon’s advertising revenue reached $17.7 billion in the last quarter, making it the third-largest ad business among tech giants. AI agents bypassing the Amazon interface would eliminate exposure to sponsored products and display ads. Furthermore, if agents simply find the lowest price across all retailers, the value proposition of the Prime subscription diminishes significantly.
Perplexity’s Defense and History
Copied to clipboard!
(00:27:00)
  • Key Takeaway: Perplexity justifies its actions by asserting that user agents are simply software labor acting on user permissions, directly challenging the notion that service providers can unilaterally block access.
  • Summary: Perplexity frames Amazon’s lawsuit as bullying that threatens internet innovation, arguing that a user agent should have the same access rights as the human user. The company has a history of being an ‘ask forgiveness, not permission’ player, having faced multiple lawsuits over scraping content for its AI summaries. Perplexity concludes that agentic shopping is the natural evolution of Amazon’s promise of low prices and fast delivery.