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 stream centers on time-boxed polish for a terminal-themed website: reviewing visual hierarchy, reducing clutter, and iteratively refining a “code rain” effect—its motion, density, visibility, and content—through precise feedback, screenshots, and component references. Luqman argues that coding agents need specific direction and should understand a project before editing it, then explores an Explorer–Planner–Builder workflow with automatic context assembly, prompt previews, and verification before handoff. The value of a fully automated pipeline, exploration quality, CLI task resumption, and unexpected behavior in plan mode remain unresolved. Other discussions cover Nightbot and social-distribution automation, an unfinished YouTube API integration for live status and schedules, and pre-launch marketing through a clear landing page and waitlist while distinguishing a global product market from an Arabic knowledge audience. Audience topics include practical streaming advice, choosing web or mobile development according to the project, reading AI benchmarks cautiously, and brief questions on core AI concepts. Preferences involving Claude, Codex, Cursor, and Warp are presented as personal judgments, and the stream does not establish that every proposed improvement or test was completed.
63 moments from the stream
- 0:04:33
Livestreaming for Review and Learning While Staying Focused
The speaker explains that recording work sessions helps him review his process, notice progress, revisit earlier ideas, share knowledge, and build a community for mutual learning. He also notes that chat questions can be distracting, so maintaining focus on the workflow remains a priority.
- 0:06:00
Chat Automation and Testing Nightbot Commands
The speaker tests Nightbot automation to reduce the effort of monitoring chat while working. He initially thinks it is not working, then confirms that it is active, reviews an initial set of chat commands and platform links, and adjusts small details.
- 0:07:09
Optimizing YouTube API Quota Usage
The speaker briefly outlines a website-polishing agenda and says YouTube API quota was being consumed inefficiently, which he fixed off-stream. This page does not show the fix or its tests.
- 0:10:50
Benefits of Live Streaming and the Challenges of Audience Growth
The speaker says streaming while working can feel difficult without an audience, and chat can become distracting, but it suits people who enjoy thinking aloud, preserves a record of progress, and may help build an audience. He also discusses automating short-form distribution across TikTok, Facebook, Instagram, and YouTube to reduce manual work and encourage Arabic knowledge sharing.
- 0:13:56
Automating Stream Status and Schedules with the YouTube API
The speaker describes a section intended to change its thumbnail when the channel goes live and show upcoming YouTube streams when offline. He says it is currently affected by YouTube API usage limits, then reviews the consumption of a Telegram bot integration without exposing private figures.
- 0:15:38
What You Actually Need to Start Streaming
In a discussion about streaming setup, the speaker says an advanced camera is not essential; what matters more is knowing what you are doing and who you want to reach.
- 0:16:51
Splitting the Development Workflow Between Cursor and Warp
The speaker discusses a terminal-sandboxing change he attributes to Antigravity on February 26, 2026, and interprets it as encouraging greater use of its built-in agent, especially when sharing terminal screenshots. Because he prefers multiple terminals and working with Claude Code and Codex, he explains using Cursor for UI and UX work—citing in-IDE previews, element selection, and screenshot sharing—while keeping Warp for backend and core-project work. He notes that he is still early in learning Cursor.
- 0:21:07
Recreating UI Elements from a Reference Site with AI
Discusses an experiment recreating a call-to-action button and then a harder visual element from a reference site. The agent received element context and a screenshot through Playwright MCP; the initial result was considered acceptable, with small differences remaining in animation behavior.
- 0:22:35
Why Coding Agents Need Specific Direction
The speaker says Claude told him it could implement an interface when the desired result was described precisely, while requests such as ‘improve the design’ remained ambiguous. He therefore sees a brainstorming stage as useful when the general idea is known but the intended visual solution is not yet defined.
- 0:23:47
Using Component Libraries as References for Agents
Shows a workflow for browsing directories of interface components such as charts and progress bars, then copying their code or passing a reference to a coding agent with instructions to adapt colors and details. The approach turns a preferred component into clearer implementation input.
- 0:25:37
Time-Boxing Polish for a Terminal-Style Website
The speaker explains that frontend work is not his primary focus, so he wants to finish the terminal-style website that day after trying a few touches, leaving only limited polish for later.
- 0:27:23
An AI Review of Website Visual Identity and Layout
The agent is asked to understand the site and focus on its interface, layout, and visual character. In the critique captured by the captions, the concept reads as a dark terminal-style command center, but the hero is overloaded, some colors and backgrounds clash, sections feel repetitive, and the information hierarchy is flat. These are the agent’s observations, not user-testing results.
- 0:30:37
Planning a Landing-Page Refresh with a Vibe-Coding Aesthetic
The discussion outlines an optimization plan and possible broader refactor of a landing page while retaining suitable interface language. One proposed direction is a subtle hero background with continuously falling code lines, lightly inspired by The Matrix to signal a vibe-coding aesthetic, alongside interest in a visual effect for the pricing section.
- 0:34:12
Troubleshooting Voice Dictation in a Cursor Terminal
While recording a message, Wispr Flow temporarily stopped working in a Cursor terminal. Troubleshooting included repeated tests, considering whether privacy or security settings might be blocking it, and considering a restart; one attempt indicated successful transcription, but the underlying cause was not clearly established.
- 0:41:45
Reducing Visual Competition and Polishing a UI with AI Tools
He identified competing visual elements, explored code-rain or Matrix-style backgrounds, and reviewed UI components. He criticized a tool for implementing changes after being asked only to refine the plan, and described Codex with a lightweight model as useful for polish work.
- 0:45:02
Balancing Website Quality with Time and Practical Goals
He argued that a working website does not need excessive visual polish, and that reaching a satisfactory baseline matters more than spending additional hours refining it before the main work begins.
- 0:46:37
Starting Marketing Before the App Is Finished
He outlined a distribution and pre-launch marketing campaign that begins while the app is still being built. The plan includes a landing page explaining the product and its value, a short promotional video, a waitlist, and an incentive for some sign-ups to become early users and generate leads.
- 0:49:00
Separating the Product Audience from the Arabic Content Audience
He distinguished the global audience for a Shopify app from the YouTube channel’s aim of serving as a knowledge hub for Arabic-speaking viewers, noting that the product market and content audience do not necessarily match.
- 0:51:42
Visually Inspecting a UI Before Editing It
The assistant was told to respect a planning-only request instead of making unrequested changes, then visually inspect the site and capture each section before fixing it. Issues included a logo hidden behind the navigation, an above-the-fold section that missed the intended direction, and a continuously moving horizontal presentation for some tool items.
- 0:53:27
Why YouTube Ad Rates Vary Across Markets
The speaker offered the general view that YouTube advertising returns in some MENA markets may exceed those in Egypt, linking higher CPMs to a market’s economic strength and purchasing power.
Got a question about something in this stream?
Ask in the group