Another Power Outage... 🙄
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 landing-page development and a practical workflow combining coding agents with image and video generation. Reviews cover card layout, layering, gradients, contrast, synchronized motion, maintaining the hero section’s visual standard throughout the page, and the importance of later mobile optimization. Luqman explores using Nano Banana and Kling to turn generated images into looping backgrounds, multi-shot videos, and hover-triggered interactions; several experiments remained in generation or planning, so their final results were not established. The stream also explains a multi-agent workflow spanning codebase and Shopify-documentation research, implementation, security review, and validation. Luqman observes that launching many agents can consume token and usage allowances quickly, while a suggested usage-multiplying bug remained unverified. For Shopify, he outlines plans to manage approval and compliance risk through an early exemption request, support follow-up, checklists, and automated review, with the review and approval outcome still unresolved. Audience questions and broader discussions cover whether coding agents understand Arabic instructions, choosing Warp for terminal and backend work versus IDEs such as Cursor for UI/UX, perceived workflow differences between Claude and OpenAI, and the hardware and quality trade-offs of local models. The stream also addresses validating products before building, planning distribution, profitability, and scalability, beginning freelance work through Upwork or Fiverr, and developing a side project gradually while retaining stable income. These are the speaker’s experience-based opinions and recommendations, not guarantees of technical or commercial outcomes.
41 moments from the stream
- 1:22:28
Background Audio as a Focus Aid
The speaker describes sometimes using subtle background audio to support concentration while working, letting it fade from attention rather than treating it as entertainment in itself.
- 1:24:11
Plan Distribution Before Building the Product
The speaker recommends working backward on any project: decide how it will be marketed and understand the expected gross margin before building. In his view, distribution remains the main obstacle.
- 1:27:08
Improving Element Contrast Across a White Transition
While reviewing a design, the speaker notes that a hard, clean transition into a white section makes the lower line stand out less than the upper one. He asks how to improve the contrast, then rejects an experimental result as visually unsuitable.
- 1:29:36
Trying a Smaller Size
The speaker tries making an unspecified item smaller and considers the result slightly better.
- 1:32:46
From Chat Platform Features to Testing Business Viability
The speaker reviewed a browser-hosted chat platform with account creation, shareable contact codes, voice, and file attachments. He then argued—conditionally—that marketing it in Egypt or the wider Arab region could be difficult when variable usage costs are dollar-denominated, while a mostly fixed-cost structure may leave more room. He advised examining profitability, monetization, and scalability before building and using AI to challenge assumptions transparently, stressing that easier product creation does not make revenue generation easy.
- 1:40:14
Choosing a Visual Treatment for Product Cards
While continuing work on product-operations and commerce-intelligence cards, the speaker felt the design still lacked something despite being nominally complete. He considered either adding a video placeholder or generating a code-based visual effect before moving through the remaining cards.
- 1:43:54
Using Language Models for Coding and UI Design
The speaker explains that language models have become part of coding and terminal workflows, mentions using Claude and Codex, and offers the opinion that Codex is also useful for UI work.
- 1:45:05
An Interactive Card About Products Being Invisible to AI
The speaker proposes generating an image that communicates that products are invisible to AI, animating it, and placing it in a card whose effect activates only when the pointer hovers over it.
- 1:45:55
Local Models: Hardware and Quality Trade-offs
The speaker discusses local models as a trend worth watching and believes that running them locally may remove the need for an internet connection. He emphasizes their demanding hardware needs and says their current quality does not meet his preferred standard. Some model or tool names in this passage are unclear in the captions.
- 1:47:29
From Generated Image to Hover-Activated Video Loop
The speaker demonstrates an experimental workflow: generate a square image with Nano Banana, use the same image as the start and end frames in Kling to create a short silent loop, then plan to embed it in a website section that plays only on hover. Generation is started and the result is awaited, so this page does not establish the final outcome.
- 1:53:30
Rearranging a Web Section Before Adding Interactive Video
The discussion considered showing a video when a user hovers over a page element, then noted that motion alone was insufficient because the proposed video concept did not fit the section. The next step was to prompt the agent to place each of two cards on its own row before discussing video implementation.
- 1:58:15
An Alternative Path for People Who Prefer Technical Work to Management
In response to a question about having strong technical skills but little interest in management or marketing, the speaker suggested offering that expertise as a service through freelance platforms such as Upwork and Fiverr, with AI tools helping draft the professional profile and present the skills.
- 2:01:00
Redirecting a Coding Agent Mid-Task with /btw
The speaker explained that the /btw command can add guidance while a coding agent is still working, illustrating it with a request to make a layout match an earlier element. He then began formulating a visual treatment in which a specific card’s border would pop slightly.
- 2:03:51
Recreating Web Components with a Coding Agent
He discusses inspecting a polished page, copying a component’s structure, or attaching a screenshot to communicate the target to a coding agent. He notes that recreating such elements is possible but time-consuming, and that discovering some reference effects were videos encouraged him to devise his own effect instead of copying them literally.
- 2:06:48
Prioritizing Mobile Website Optimization
He explains that the phone layout has not yet been optimized because he is conceptualizing the desktop version first, while considering mobile more important in practice. Based on his experience, many sites and businesses receive substantially more mobile traffic, so desktop work should not eclipse the mobile experience.
- 2:07:47
Building a Side Project While Keeping Stable Income
He advises against immediately leaving a job to start a project without sufficient resources, recommending keeping a source of income while building the project gradually in available time. He adds that AI tools can make this path easier.
- 2:09:39
Maintaining the Hero Section’s Standard Throughout a Page
He reviews a section below the hero and feels its quality drops as soon as the user scrolls, making the page feel ordinary. His goal is to carry the hero’s standard of graphics, effects, animation, and presentation through to the footer, prompting a broad rethink of the section through a visual workflow involving Nano Banana and Kling.
- 2:13:54
Designing a Cinematic Multi-Shot Video from an Image
An experimental workflow moves from full-width cinematic visual concepts to generating an image and using it as the basis for a multi-shot video. The discussion explains that Kling can assign a prompt and duration to each shot in a custom sequence for visual storytelling, then configures several shots without confirming that the final generation succeeded.
- 2:24:59
Sequencing Development Milestones and Parallel Work
A roadmap review using swimlanes, with consideration of starting application development in parallel because the landing page, waitlist, and supporting content take time.
- 2:25:59
Getting Started with Remote Work on Upwork
General advice to explore Upwork based on one’s profession and experience, build a profile gradually, and verify that work is becoming dependable before relying on the platform fully.
Got a question about something in this stream?
Ask in the group