Que s'est-il passé
Airbnb CEO Brian Chesky stated in a TechCrunch interview that the current consumer AI landscape is failing because it lacks a dedicated operating system. Chesky argued that existing AI agents, such as Muse and Instinct, struggle to interact effectively with platforms like Airbnb because they lack a robust software developer kit (SDK) and kernel-level integration. He proposed that for AI to reach its potential, apps must transition into 'agentic' forms that are inherently interoperable, rather than relying on current chatbot interfaces that he believes are ill-suited for complex e-commerce and travel discovery.
In an interview with TechCrunch, Brian Chesky expressed skepticism regarding the current state of consumer AI, noting that his own experiences using agents like Muse and Instinct to book travel have been suboptimal. He argued that these tools lack the necessary software infrastructure to handle the nuances of platforms like Airbnb, which require complex features such as map integration, identity verification, and host messaging.
Chesky emphasized that the industry is currently missing a true 'AI operating system' that would allow components and agents to interact at the kernel level. He criticized the lack of a comprehensive SDK for AI agents, comparing the current state of AI development unfavorably to the early days of the App Store, which provided the necessary framework for developers to build functional, interoperable software.
The CEO outlined a vision where the Airbnb app itself evolves into an agent, with specific agents handling tasks like exploration and customer service. He suggested that these internal agents would eventually become interoperable with external agents through standards like MCP, effectively turning apps into 'apps with faces' that can communicate autonomously to complete user requests.
Détails de la source: techcrunch.com ↗
Pourquoi c'est important
Chesky’s critique highlights a significant friction point in the current AI industry: the lack of standardization for agent-to-agent communication. By calling for an 'AI operating system' rather than just a 'quarterback' agent, he identifies a structural limitation that prevents AI from seamlessly navigating complex, multi-step tasks across different platforms. This perspective challenges the prevailing Silicon Valley trend of building standalone chatbots, suggesting instead that the future of consumer AI depends on deeper, system-level integration that allows apps to function as interoperable agents. His comments underscore the ongoing struggle to move beyond simple text-based interactions toward functional, agent-driven workflows.
The core of Chesky’s argument is that the current 'chatbot' paradigm is a poor fit for e-commerce and travel, where users value browsing, inspiration, and collaborative planning. He believes that by forcing users into a linear, turn-based chat interface, companies are stripping away the 'library experience' of travel planning.
By advocating for an operating system approach, Chesky is highlighting the need for a shift from 'apps as data layers' to 'apps as agents.' This shift would theoretically allow for a more fluid user experience where an agent can perform tasks across multiple platforms without the user needing to manually switch between disparate applications.
The interview also touches on the role of design, with Chesky asserting that human designers are still better at creating intuitive interfaces than models. He suggests that the future of AI interfaces will be a hybrid of deterministic, pre-designed elements and generative components, rather than a total abandonment of traditional UI.
Mécanisme interactif : comment cela fonctionne réellement
Explorez de manière interactive la technologie sous-jacente à ce développement.
crm_get_transaction(id='4092').An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?
Que regarder ensuite
Airbnb plans to roll out voice-enabled agents this fall to assist with search and customer service. Additionally, the company is exploring 'multiplayer' AI interfaces designed for group trip planning, moving away from the single-user chatbot model. Observers should monitor whether Airbnb’s push for agent-friendly infrastructure and interoperability standards, such as the Model Context Protocol (MCP), gains traction with other major platforms or if the industry remains fragmented by competing 'quarterback' agent models.
Airbnb is actively developing 'multiplayer' AI, which aims to support group collaboration—a Chesky identifies as a major weakness in current single-user AI agents.
The company is preparing to launch voice agents this fall, which Chesky believes will eventually become the primary way users interact with computers, surpassing typing.
The industry's reliance on Apple and Google platforms remains a critical factor. Chesky noted that unless these platform holders build a new, agent-centric OS, the industry will likely remain stuck with 'apps that have agents inside them' rather than achieving a true transition to a universal agent ecosystem.