This podcast ranks and discusses various 'vibe coding' applications like Cursor, Lovable, v0, and Windsurf, offering a tier list and advice for both technical and non-technical users in the evolving AI software development landscape.
Takeways• Cursor and Claude Sonnet 4.5 are top choices for technical users, with Cursor's agent providing a current edge.
• Non-technical users should adopt a patient mindset, recognizing that effective software development requires sustained effort beyond initial prompts.
• Choosing a vibe coding platform also involves 'betting on' the founding team and their vision for the tool's future.
The hosts provide a definitive, non-professional tier list of vibe coding apps, emphasizing that while all tools are amazing, their ranking considers factors beyond pure tech, like team trust and community support. The discussion highlights the differing target audiences for these tools—developers versus non-technical users—and stresses the importance of a patient mindset when building software with AI tools.
Tier List and Tool Overview
• 00:01:14 The discussion begins by categorizing various vibe coding apps, with Windsurf immediately placed in 'D tier' due to issues with the founding team, despite its technical capabilities. Cursor is highly rated, particularly for its strong community and extensive documentation, which is crucial for developer tools. Lovable and Bolt are positioned in the 'B tier' or 'A between B' for their abstraction of integrations, making them more user-friendly for non-technical individuals, though v0 is noted as potentially superior for prototyping due to its extensive integrations and Vercel ecosystem integration.
• 00:10:34 The conversation then shifts to mobile-focused vibe coding apps like Vibecode app and Rourke, which are deemed solid 'B tier' options given their relative newness and the growing trend of building consumer mobile applications. The hosts note that many of these mobile app platforms are likely built on Expo to support both Android and iOS, benefiting from React's robust ecosystem and financial backing.
• 00:14:51 The core AI models underpinning these tools are examined, with Claude Sonnet 4.5 identified as the current leading coding model. Codex, despite a rocky start, has significantly improved to be comparable to Claude, earning the 'most improved' award. The effectiveness of these tools also depends heavily on the 'agent' layer built on top of the language model, which handles file reading and writing, with Cursor's agent currently outperforming Claude Code's agent in some aspects.
• 00:21:42 For technical users, Cursor or Claude Code are recommended due to their advanced capabilities, assuming the user will heavily steer the development process. Non-technical users are advised to adopt a crucial mindset shift: accept that building functional, production-ready software takes time, planning, and testing, often more than just a few prompts. Users should persist through early failures and treat software development as a journey, similar to building a business or physical training, where experience and patience lead to better outcomes. Additionally, following the founders of these platforms is suggested as a way to align with a company's vision and trajectory.