Preparing for Ad Launch
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
A wide-ranging practical stream combining a Burnish website and mobile review with discussions about product launches and AI-assisted work. Luqman examined content density, spacing, motion, input and CTA clarity, and visual identity, while the mobile audit, landing page, and some template regressions remained unfinished. On launch strategy, he argued against letting perfection delay shipping, but the product, lead funnel, email sequences, and ad creatives still required preparation and testing; conversion costs and projected outcomes therefore remained unknown. He also discussed lead qualification, install-to-paid measurement, entering the Shopify ecosystem after competitor research, and balancing report depth against speed. In audience Q&A, he shared personal views on Claude, Codex, Gemini, context compaction, multiple agents, and CLI workflows, stressing that tool choice is workflow-dependent and that these were not systematic comparisons. He warned against exposing API keys or selecting a cheaper provider without checking privacy and reliability. He advocated incremental learning through building rather than theory alone and framed AI as an amplifier of existing skills, not a replacement for human judgment. The conversation also covered education, freelancing, B2B versus B2C, unit economics, focused work and breaks, plus brief practical suggestions for preparing chicken meals.
159 moments from the stream
- 2:59:37
Reassessing Product Direction After an Overlapping Feature
The speaker says he prefers executing defined directions during a livestream and doing strategic thinking and brainstorming off-stream. He recalls an experience in early April when he paused to reassess after Shopify released a feature he considered very similar to his idea, then returns to the website’s general objective of generating leads.
- 3:01:20
Running an AI Tool Locally or on the Web
Answering an audience question, the speaker recommended downloading and running the tool locally, explaining that its web version relies on cloud execution. The tool’s identity is unclear in the captions.
- 3:06:07
A First Vibe-Coding Project
The speaker said he has long experience building projects and businesses, but believes this is his first project developed through a vibe-coding approach.
- 3:07:41
Balancing Report Speed and Analytical Depth
The speaker compared his product with competing tools that appear faster and may be free, suggesting their speed could reflect shallower analysis. He wants to provide a deeper report while recognizing that some customers prefer a quick, compact result before deciding to install.
- 3:08:52
Measuring the Install-to-Paid Conversion Funnel
The speaker explained that funnel economics depend on cost per install, followed by a seven- or fourteen-day free-trial journey and the install-to-paid conversion rate. He emphasized that the landing page should encourage installation, while the cited example figures were illustrative rather than confirmed forecasts.
- 3:10:04
Planning Facebook and Shopify Campaign Paths
The speaker discussed sending one Facebook campaign to the funnel page and another directly to Shopify, tentatively expecting the direct route to perform better. He also mentioned testing ads within Shopify, though he appeared uncertain when he could not see them during the live search.
- 3:12:59
Define the Field Before Asking for Experience-Based Advice
Responding to a broad question about lessons from experience, the speaker said priorities vary by field and suggested that studying marketing may matter more than some other skills, so the question should first be narrowed to a specific area.
- 3:14:50
Learn AI-Assisted Coding Through Daily Building
The speaker said his vibe-coding experience began gradually in September 2025, followed by a break and a later return with greater focus. In his view, daily practice, direct work with AI, and continual building can teach a great deal without necessarily relying on courses.
- 3:16:15
Remove the Duplicate Email Signup on Mobile
During an interface review, the speaker noticed two nearby email signup prompts for AI-search product updates. The recommendation was to remove the redundant prompt on mobile while keeping it on desktop.
- 3:23:51
Finishing the Mobile Version Before Switching Workstreams
A practical update on a mobile version that still has issues to resolve. Working on two priorities at once caused distraction, so the stated plan was to finish the mobile version before focusing on the other workstream.
- 3:25:44
Automatic vs. Manual Compaction in Long Coding Sessions
An exploratory discussion of context compaction in long coding sessions: compaction summarizes session history to save space while attempting to preserve recent changes and plans, while auto-compaction was described as a command that lets Claude choose when to compact instead of waiting for a manual trigger. The speaker found the idea useful but explicitly said he had not tried it and continued with manual compaction to preserve context before the next step.
- 3:33:52
When Is Manual Context Compaction Better?
The speaker discusses the risk of automatic context compaction during active debugging, where precise details or important context may be lost, and prefers choosing the timing manually.
- 3:35:55
Uncertain Forecasts About a New Model Release
A brief discussion of probability-based forecasts for a new AI model release, rejecting confidence in a release the following week and leaving the timing explicitly uncertain.
- 3:37:23
Denoising Homepage Content for Mobile
The task is to review the homepage first, identify elements that need compaction or denoising, and make mobile copy concise, direct, and punchy before proposing changes.
- 3:48:08
An Opinion on Building Tools for the AI Ecosystem
Responding to a question about simple work that supports AI, the speaker offered an opinion—attributed to himself and experts—that tools or businesses serving AI platforms may occupy a promising space. He cited AI-oriented development tools in general, while his claim that failure becomes difficult remains an optimistic opinion, not a guarantee.
- 3:49:11
Terminal Use and Sandboxing in Cursor
The speaker discussed using a terminal inside Cursor and said that Cursor and another tool whose name is unclear had introduced terminal sandboxing. He dated his observation to around mid-February after initially mentioning March, then suggested that using Cursor directly would likely provide a better experience; some technical details remain unclear in the captions.
- 3:50:46
Start with Codex Inside Your Existing Field
The speaker suggested that a beginner install Codex and experiment with building something small in their existing field, such as development operations or network support, using dialogue with the tool to explore what is possible. He framed AI as a way to amplify existing skills and multiply their impact.
- 3:55:02
Using Featurebase for Customer Support
The speaker says he subscribes to Featurebase and uses it for customer support. Further operational details are unclear in the available captions.
- 3:58:04
Using AI as a Work Partner Before Rushing to Build a Team
The speaker recommends trying AI as an experienced work partner: explain your skills and goal, ask it for guidance, and use it as a supporting team before deciding to bring in a human partner. He presents this as practical advice while acknowledging that teams and partnerships can have both advantages and drawbacks.
- 4:02:35
Choosing Between Claude and Codex Based on Budget
Responding to a question about an alternative coding tool, the speaker says Claude can be used and gives Codex as his personal recommendation when the budget is limited, without presenting a comparative test or documented performance result.
Got a question about something in this stream?
Ask in the group