The host demonstrates building a new online events feature for a meetup app using a spec-driven, modular approach with GPT-5 and Cloud Code, emphasizing thorough testing, validation, and user privacy. They also discuss an upcoming “Built with Claude” competition, share insights on various AI models, interact with viewers about AI development challenges, and plan to deploy the app publicly while encouraging community participation.
In this video, the host begins by greeting viewers and discussing plans to build a new feature for a meetup app using Cloud Code and GPT-5. The goal is to create an online events page where registered users can create live stream events, allowing others to sign up if they want to share demos or topics. The approach involves a spec-driven, modular development process with small incremental steps, including schema creation, API development, input validation, and thorough testing before moving on to the next phase. The host emphasizes the use of a detailed Product Requirements Document (PRD) generated with GPT-5 to guide the development.
The host also mentions an upcoming competition on Reddit called “Built with Claude,” hosted by Anthropic, where participants can build AI-powered projects using Claude or Cloud Code to win prizes. They plan to participate in the contest during the next week’s stream and encourage others to join. Additionally, the host discusses experimenting with new AI models like Sonic and Nano Banana for various tasks such as image editing and code generation, sharing initial impressions about their speed and quality compared to other models.
Throughout the stream, the host interacts with viewers, answering questions about AI tools, real-time voice applications, and the differences between Cloud Code and Gemini CLI. They share insights on the challenges of AI development, particularly around security and validation, and highlight the benefits of building custom tools tailored to specific needs rather than relying on generic SaaS products. The conversation also touches on the evolving AI landscape, including the potential for running models locally and the importance of modular, test-driven development workflows.
The host demonstrates the progress on the online events feature by showing the creation of live stream events, managing event data, and testing the user interface. They address some technical issues encountered during development, such as build errors and slow performance, and explain how they resolved them. The feature allows event creators to keep contact information private while letting others view event details and sign up. The host plans to deploy the app publicly for viewers to test and provide feedback.
In closing, the host reflects on the success of the spec-driven, modular approach using GPT-5 and Cloud Code, noting that it enabled smooth development with continuous testing and validation. They thank viewers for their participation and announce upcoming streams, including a giveaway for a Cloud Code Pro subscription and the planned build for the Claude competition. The video ends with a positive outlook on AI-assisted development and community engagement in building innovative projects.