Update + Checkpoint
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 developing Burnish, a Shopify application intended to measure and improve merchants’ visibility in AI answers through content and automation. Practical work covered the dashboard, settings, post-install onboarding, V1 information architecture, and surfacing autopilot activity while keeping proposed changes subject to merchant review. The UI/UX phase was described as complete for this stage, but full debugging and validation remained ahead; proposed next steps included testing the journey from fresh data, defining an objective readiness rubric, and verifying features incrementally because defects become harder to diagnose as dependencies accumulate. In audience discussions, Luqman offered qualified views on Codex, Claude, Remotion, multi-model review, and coding benchmarks, stressing that agent scores and benchmarks do not replace human validation and that model and workflow choices should balance quality, cost, and usage constraints. The stream also explored finding product ideas in real everyday problems, combining SEO with GEO, positioning Burnish within Shopify, and designing dashboards around a clear next action. Questions about AI-crawler permissions, parts of the settings behavior, and the tradeoff between deep automation and fast responses remained unresolved or in need of further verification.
64 moments from the stream
- 2:28:00
Turning QA Review into a Measurable Workflow
While parallel sessions worked on the dashboard home page and settings, preparation began for a QA walkthrough starting with the storefront before moving to other areas. The goal was for a QA assistant to understand the product from available evidence, suggest what should be verified, and establish a more objective readiness rubric instead of relying on an opaque score.
- 2:36:23
Understanding Codex Usage-Limit Resets
The stream discussed a Codex feature understood to provide a periodic usage-limit reset, with both parties potentially receiving an additional reset when a new user joins through an invitation. This was the participants’ understanding during the discussion, not an independent verification of current terms.
- 2:42:05
Reviewing Onboarding Through the Code
The speaker described onboarding as one of the most important customer experiences, but could not inspect it visually at the time. He therefore asked for the implementation to be studied in code, its current behavior documented, and the process understood in depth before deciding next steps.
- 2:50:44
Scoping Post-Install Onboarding Before Debugging
Clarifies that onboarding means the wizard shown after installation and before the user reaches the homepage, not the pre-install scan funnel. It also separates work requiring direct review from work suitable for supporting sessions, while waiting for the dashboard and settings to merge and deploy before debugging continues.
- 2:52:21
A Public Timeline for Understandable Project Progress
Discusses a public roadmap or timeline, possibly using swimlanes, so visitors can understand what is being worked on and how far the project has progressed. Backend work, debugging, and quality assurance are acknowledged as difficult to represent accurately, so a simpler and clearer progress view is deferred until the current work is finished.
- 3:00:18
Preparing a Deep Review Session for Onboarding
Requests a clear handoff for a fresh session that starts with the underlying purpose, develops an in-depth understanding of the application's pages and their roles, and then dissects the onboarding process through structured questioning. The named areas include storefront visibility, articles, intelligence, and settings.
- 3:06:12
Feedback on a Code-Built Interface in Remotion
The speaker praised a first attempt at building the design in code with Remotion and suggested adjusting the header and obscured elements, while stressing consistent layout and colors across the app’s pages.
- 3:10:25
Responding to a Model’s Possible Removal from Subscription Plans
Answering an audience question, the speaker said a model’s removal from subscription plans would not be a major issue because alternatives are in use, and speculated—without certainty—that it might return through an API near year-end.
- 3:11:41
An Internal Rules Document for Guiding Agents
The discussion described using a model to prepare an internal document containing rules and working principles, which other agents could consult while carrying out tasks to check that they follow the guidance.
- 3:21:05
Positioning Burnish Between Deep Automation and Fast Responses
The speaker argues that reducing Burnish to a model query and a one- or two-line answer could weaken its value, because automation and differentiated capabilities are central to its positioning. At the same time, he worries that a slow session could make users leave, so he does not reject the proposal but suggests exploring alternatives that preserve both depth and responsiveness.
- 3:26:33
Combining SEO and GEO in a Content Strategy
The discussion keeps SEO articles as the foundation while adding GEO to help each client appear for relevant questions. This can be supported through article content and guidance for home and product pages, with the observation that traditional SEO clients are already asking about visibility in AI answers even when they do not know the term GEO.
- 3:27:39
A Shopify-First Strategy in a Competitive GEO Market
The speaker describes Western GEO competition as strong but not impossible and sees a specialized Shopify app as a possible differentiation wedge before later expansion. He also offers the qualified market opinion that Arabic and Gulf markets may be less crowded and could have demand for these services.
- 3:28:08
Moving From Visibility Measurement to Improvement
The speaker explains that the application is intended not only to measure merchants’ visibility across AI platforms, but also to help improve that visibility through research and other mechanisms whose wording is not fully clear in the captions. A recorded presentation is described as a way to explain the concept in detail.
- 3:29:49
Dashboard UI/UX Completed, Debugging Next
The speaker announces completion of the dashboard’s final page, treating the application’s UI and UX work as finished for this stage. The next phase is debugging, which he characterizes as the more difficult part.
- 3:31:50
Direct AI Provider Queries and the Quality–Cost Tradeoff
For queries made on a client’s behalf, the speaker prefers calling each AI platform’s own API rather than routing through OpenRouter, citing accuracy and data-retention considerations while acknowledging that the details are complex. He also discusses choosing a lower-cost model when the quality difference is small, especially for a B2B product, while preserving adequate quality and sustainable product economics. The model names in the captions are unclear and are therefore not asserted.
- 3:37:16
Choosing a Model by Quality, Cost, and Gross Margin
He explains that choosing a model or configuration should begin with product pricing, market expectations, offered value, and gross-margin calculations. He recommends comparing output quality directly across options while accounting for cost rather than relying on impressions alone.
- 3:41:00
Strengthening the Growth Loop and Correcting the CTA
During a UI review, he asks to improve the growth-loop section, explore different wording directions, strengthen the call to action, and correct a CTA he considers misleading.
- 3:47:07
Reviewing the Value Proposition in Onboarding
He reviews a card near the bottom of the onboarding flow that presents the value proposition or unique selling proposition, asking to inspect its current content before revising it.
- 3:48:43
Why Bugs Emerge as AI-Generated Features Accumulate
Answering an audience question, he explains that an AI tool may implement an isolated feature correctly, while bugs emerge as features, variables, and interdependencies accumulate. He attributes this partly to limited system-wide context rather than failure in every individual addition.
- 3:50:50
Keeping the Value Proposition Current as a Product Evolves
He notes that an otherwise rich application description had become outdated because it no longer reflected later capabilities, and proposes summarizing the product’s largest aspects and values in five main points.
Got a question about something in this stream?
Ask in the group