Building AI Apps Live from Cairo 🔴
Want to back the stream? The «داعم» membership on YouTube is EGP 60 a month: badges, emoji, and I answer your questions first on stream. Become a supporter
The gist
The session centers on building a low-maintenance stream hub and connecting it to YouTube so live status, schedules, and upcoming broadcasts update automatically, with a static fallback. Practical work covered YouTube Data API setup, Cloudflare variables, diagnosing an API-key restriction and an uncommitted file that broke deployment, then designing distinct live/offline cards, playlist sections, and a scrollable video carousel. Some offline-state priorities, day-versus-night organization, and a local refresh error remained unresolved. The discussion provisionally favored sending viewers to YouTube instead of embedding the player, especially for mobile usability and native engagement. It also offered personal, exploratory comparisons of Cursor, Antigravity, and Warp; task routing between Codex and Claude; parallel workflows and context handoffs; and coding-agent permission configuration that treated file deletion and GitHub pushes as sensitive exceptions. Terminal sandboxing and screenshot routing were criticized without a systematic final tool verdict. A separate command-center concept combined a timeline project map, subtask nodes, specialized agents, execution-attempt tracking, and automatic prompt assembly from project context and instruction files. Landing-page workstreams included Telegram community promotions, simplifying the email-capture area, and restructuring the “What I’m Building” section around audience needs, with several information-architecture options left undecided. On design, the discussion argued that visual references, precise spacing and motion specifications, and repeated human review matter more than broad labels such as “premium,” and that paid effects libraries are most valuable when they save substantial engineering time. Another experiment involved a goal-tracker-style visual counter, structured JSON handoffs between agents, and a proposed Playwright MCP browser connection; the captions do not confirm that the connection or implementation was completed. Additional questions covered SEO and traffic measurement for Egypt and MENA, the marketing and distribution gap facing AI-product builders, and general B2B-versus-B2C SaaS trade-offs, without documented final research findings or tests.
53 moments from the stream
- 2:36:00
When Are Effects Plugins Worth Buying?
The speaker examined why someone might buy a plugin if an agent can code the effect. The discussion’s conclusion was that simple visual effects may not justify a paid plugin, while libraries hiding substantial engineering can save considerable time through ready-made camera controls, performance monitoring, instancing, and level-of-detail support. Their value is time saved rather than otherwise impossible capability.
- 2:40:42
Previewing a Site and Sending Page Elements to an Agent
The speaker demonstrated a development workflow that opens a local site inside the editor and shows agent changes live. He also described selecting a page element and adding it directly to the agent chat, allowing a specific section to be referenced instead of relying only on prose or screenshots. The explanation continues beyond this page.
- 2:42:18
Troubleshooting Element and Screenshot Attachments in Antigravity
The speaker investigates which chat receives a selected page element, then says screenshots he previously could add to the terminal now appear to be added automatically to the agent. He presents this as a frustrating observed experience, not a conclusive test of the tool’s behavior.
- 2:45:10
A Personal Comparison of Cursor and VS Code
The speaker says he has tried VS Code and finds it acceptable, but is drawn to Cursor because of its adoption and his desire to try a new development environment. This is a personal preference rather than a comprehensive technical comparison.
- 2:47:58
Turning Visual Inspiration into a Precise Prompt
The speaker examines a striking website and tries to describe a visual interface element to an AI coding tool. He concludes that the request needs concise, highly specific visual requirements, especially when the tool cannot directly see the result, while also asking about V0 and whether plugins are necessary.
- 2:53:19
Trying Cursor Setup and Its Development Workflow
A hands-on trial of setting up Cursor, connecting GitHub, opening a project, running the development environment, and exploring its terminal, browser, and screenshot workflow. The initial impression favors this flow over another coding environment the speaker had used, though the assessment remains provisional.
- 2:58:20
Comparing AI Coding Tools Before Committing
The speaker considers suggested tools, bookmarks references for later research, and makes an early comparison among Cursor, Gemini, Claude, and other options. The choice is framed around reproducibility, ease of opening projects, and general usage limits, without reaching a final verdict.
- 3:02:11
A Command Center for Tasks and Coding Agents
The speaker demonstrates a command-center concept with a timeline-based project map, nodes that reveal subtasks, a hub for specialized agents, and a task board that records execution attempts. The system can automatically assemble a prompt preview according to the selected agent type, project context, instruction files, and guiding documents.
- 3:04:13
Prompting a Coding Agent to Design a Futuristic Visual Counter
The discussion covers refining a coding agent’s system prompt, asking it to research goal-tracker design references, and then create a more dynamic, futuristic effect for a numeric counter. It also proposes connecting Claude Code to the browser through Playwright MCP, restarting it to check whether it can see the website, and using structured JSON to reduce ambiguity between agents; the captions do not confirm that the connection or implementation was completed.
- 3:21:42
Request for an SEO and Traffic Measurement Strategy for Egypt and MENA
He requested a brief SEO strategy aimed at Egypt and the wider MENA region, along with exploring GA4 for website traffic measurement.
- 3:23:51
An Opinion on the B2B vs B2C SaaS Trade-off
He argued that B2B SaaS companies can achieve strong gross margins and that B2C SaaS startups may also perform well, but distribution and marketing are harder. This was his general opinion, not a demonstrated test result.
- 3:25:47
Why AI Product Builders Get Stuck on Marketing
He explained that AI has made product building easier, yet many young builders slow down when they must market what they created and bring it to market. He suggested there is a structured approach he wants to explain for addressing this marketing and distribution gap.
- 3:27:19
Coding Tool Preferences and Vibe Coding
He said moving to Cursor was one of the session’s best steps, noted that he would look into v0, and playfully described the workflow as “vibe coding.” These were personal impressions rather than a systematic tool comparison.
Got a question about something in this stream?
Ask in the group