The First Numbers Are Coming In
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 documents preparations for a Shopify app intended to improve merchants’ product visibility in AI answers, with practical work on help-center-grounded support, onboarding emails, scan-flow clarity, progress states, and pre-merge review. Much of the session focuses on launch readiness: developing static and video ads with tools such as Codex, Remotion, and Higgsfield, then using audience feedback to prevent the creative from being mistaken for a shoe ad and to explain AEO within seconds. Luqman discusses testing multiple angles and formats across Meta and Shopify rather than predicting CTR, CPC, or CPI without data; comparing a direct-install link with a scan-and-email funnel; and using predefined success criteria and a unified analytics dashboard. Ad underdelivery, attribution, incomplete data integrations, target-market selection, pricing, and campaign structure remain unresolved or subject to testing. Other discussions cover reaching Gmail’s Primary tab, the ethics of reviews and UGC, avoiding unsupported marketing promises, and personal—not benchmarked—comparisons of AI tools. Audience questions also explore learning programming and marketing through practice, repository-level specialist agents, changing software roles, and building skills or projects with limited resources. No launch success is claimed: the session is presented as an evolving work record whose market results will require weeks of measurement.
171 moments from the stream
- 2:46:55
How to Set a Digital Product Marketing Budget
The speaker explains that there is no single suitable budget for every digital product. The product, break-even point and target audience must first be defined, and audience geography can materially change the appropriate budget, so the question needs narrowing before a budget can be estimated.
- 2:47:56
Coding-Agent Speed and the Changing Work Cycle
The speaker compares Claude Code and GPT on efficiency and resource use based on his current experience, then speculates about how much faster agents could reshape debugging, application review, and media generation. The timing examples are personal observations and projections, not documented benchmarks.
- 2:51:04
What Is the Point of a Checkbox That Changes Nothing?
The speaker questions the practical value of a checkbox when checking or leaving it unchecked does not change the outcome, contrasting functional purpose with mere compliance. The underlying legal requirement remains unresolved in the discussion.
- 2:52:02
AI Is Reshaping Software Roles, Not Simply Ending the Field
The speaker argues that AI tools have rapidly moved from chat toward acting inside computers, so some programming tasks may shrink while new roles emerge. He cites “Forward Deployed Engineer” as a role being discussed, while framing his near-term outlook as personal and uncertain.
- 2:55:09
Codex Implements Interface Elements Over a Video
The speaker points to a practical example in which Codex created interface elements placed over a video in a project he was working on. The project name is not clear enough to identify, and the page does not document a formal test of the result.
- 2:55:37
Limits of Overly Cautious Assistants in Compliance Questions
The speaker describes Claude as overly cautious around legal-style questions and challenges its rationale for adding a user option when the outcome is unchanged either way. This is a critique of assistant behavior, not legal advice or a legal conclusion.
- 2:56:54
Software Engineers’ Divide Over AI’s Impact
The speaker distinguishes between engineers he sees as engaged with the shift and others whom he believes dismiss its scale, responding sarcastically to a claim that recent years brought no meaningful change. This is a personal view within a professional debate, not a representative measure of engineers’ opinions.
- 3:01:41
Re-examining Assumptions About Software Engineering’s Future
The speaker says his own unspecified field has ‘disappeared’ and argues that people who reject his assessment of software engineering’s future should consider the views of multiple technology figures rather than rely solely on personal certainty, while keeping this attributed to him rather than presenting it as settled fact.
- 3:05:05
Respectful Discussion and Responses to Insults
The speaker says respectful discussion can help clarify different viewpoints, but adds that he responds in kind when disagreement becomes a personal insult, before mocking the other person’s position.
- 3:06:49
Valuing Programmers’ Expertise and Mutual Learning
The speaker says he respects software engineers and computer science practitioners and has learned from their way of thinking. He invites people with programming backgrounds into discussions for mutual learning and acknowledges that their programming expertise may exceed his own.
- 3:08:15
Making the Free-Audit Call to Action More Prominent
He asks for a UI revision that makes the free-audit section clearer and more concise, considers placing it near the report-delivery field, and requests a preview of the result.
- 3:11:33
Building Specialized Agents Inside a Code Repository
Answering an audience question, he explains that Codex provides agents and that repository-level agents can be configured to use Codex or Claude for focused work such as auditing, security, compliance, and code consistency. He recommends defining the desired responsibility, then asking the AI tool to help design the specialized agents.
- 3:12:40
Replay Availability and Content-Channel Strategy
A viewer suggests making past streams available through Telegram for people who want to revisit them later. The streamer says he will explore the idea while expressing a preference for keeping specialized live content on YouTube and sharing short excerpts on social platforms.
- 3:14:28
Integrating a Customer-Service Bot with a Help Center
He demonstrates a customer-service bot connected to a help center so it can answer client questions, while allowing a person to step in when needed and receive message alerts on a phone. He also says booking is working and that some debugging remains, although he believes the experience is ready.
- 3:16:05
Learning AI-Assisted Programming by Starting Immediately
He says he did not set aside a separate period of theory before starting; instead, he began working immediately and learned extensively through practice. This is his personal experience, not a general estimate of how long others will need.
- 3:18:24
Learn Codex by Building Something You Need
The speaker recommends starting with a clear idea for something personally useful, then using Codex and learning by building it on your own machine; hands-on work gradually reveals how the development process works.
- 3:21:09
The Stream as a Record of an Evolving Learning Journey
The speaker describes the successive streams as a record of his attempt to develop alongside rapidly changing AI, noting that his knowledge and workflows have changed since the early recordings and may keep evolving.
- 3:22:07
The Value of Programming Experience in AI Communities
The discussion highlights how people with programming and game-development experience can enrich an AI community, encouraging them to participate so others can benefit from their technical perspective.
- 3:23:11
Reviewing Identity Clarity in Ad Creatives
While reviewing ad creatives, the speaker observes that some could be mistaken for footwear ads rather than AI-related advertising. He mentions that one design could be revised and notes the presence of five AI-engine logos.
- 3:24:24
When an AI Model Gets Stuck, Revisit Your Reasoning and Prompt
Asked about model limits, the speaker says complete dead ends have become less common in his experience than in late 2025. He recounts resolving a difficult issue after considering that his own approach might be the problem and changing both his reasoning and prompt; this is his experience, not a guaranteed general rule.
Got a question about something in this stream?
Ask in the group