Send us a text
In this insightful episode, a seasoned developer shares their hands-on experience comparing various AI-powered web development tools to determine which solutions work best for different project requirements. After spending a month "in the trenches" testing these platforms, they provide a practical breakdown of the strengths and limitations of popular tools like Bolt, Lovable, and custom coding environments.
The episode begins with the speaker acknowledging the overwhelming number of AI web development tools flooding LinkedIn feeds, causing confusion for many who aren't sure which solution best suits their needs. Working at a production studio that needs to regularly test new applications, they've created a comprehensive comparison to help others navigate this rapidly evolving landscape.
At the heart of the discussion is a real-world test case: recreating Booking.com's homepage. The speaker demonstrates how they rebuilt the site with different approaches, comparing the results side-by-side. They distinguish between "one-shot effect" platforms like Bolt and Lovable, which produce impressive initial results quickly but become difficult to refine, versus a more traditional programming environment that offers greater control despite requiring more setup time.
Using visual examples, the speaker shows how Bolt produced better one-shot results than Lovable (though they acknowledge this might be due to using Bolt's paid version versus Lovable's free tier). The comparison extends to creating an Airbnb-style listing page, where the differences in layout refinement capabilities become evident. While Bolt created a functional page in just ten minutes, the custom solution took three hours but provided significantly more control over nuances and styling details.
The speaker's preferred custom setup includes Windsurf (similar to Cursor), Claude 3.7, and libraries like Three.js—providing a senior programmer's environment with complete control. They emphasize the importance of version control through GitHub for larger projects and the freedom to change libraries and add APIs as needed.
A key insight from the episode is the changing role of designers in this AI-powered landscape. Rather than creating mockups in Figma, designers are becoming "curators" who find exemplary references and iterate rapidly, prioritizing time-to-market over pixel perfection.
The speaker concludes with an impressive demonstration of building a custom video compression tool in just ten minutes, reinforcing their philosophy that mastering these development environments enables continuous creation and problem-solving without relying on pre-built solutions.
The episode offers practical advice for viewers: use Bolt or Lovable for quick projects where you can accept their limitations, but invest in learning custom environments like Cursor or Windsurf for longer-term projects requiring fine-tuned control. The speaker encourages viewers to embrace these tools now, noting that despite the learning curve requiring longer work days, the investment in understanding these technologies will pay dividends in future development capabilities.
Interested in joining the podcast? DM Behrad on LinkedIn:
https://www.linkedin.com/in/behradmirafshar/
This podcast is made by Bonanza Studios, Germany’s Premier Digital Design Studio:
https://www.bonanza-studios.com/
UX for AI