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
- 0:07:28
Sequential vs. Agent Workflows in AI Coding Tools
Answering a question about new models and options, Luqman describes experimenting with Claude, Codex, and another tool whose name is unclear. In his experience, a sequential workflow is slower but concentrates the model on one task and may use fewer tokens, while deploying many agents can consume tokens quickly. This is presented as his personal assessment.
- 0:09:30
Arabic Chat Being Automatically Translated into English
Luqman notices that Arabic chat messages are being translated automatically into English, says he does not know why it is happening, and finds it inconvenient.
- 0:10:21
The Burnish Dashboard as the Application Hub
The Burnish dashboard was described as the final piece connecting the application and as its current homepage. A full code and context review was requested before assessing the dashboard’s current state.
- 0:13:11
Balancing AI Tools and Workflow Additions
The discussion expressed comfort with Claude while identifying token limits as a constraint, and deferred adding Obsidian and memory-oriented ideas because of their learning curve and the priority of finishing the current work. Early experiments were mentioned, but no completed codebase integration was claimed.
- 0:16:19
How Does an Advertising-Based Website Make Money?
The speaker tried to understand an unidentified website’s monetization model by requesting a concise explanation and a practical example in plain language. The discussion indicated that revenue comes from an advertiser and that a highlighted word may become an advertising line, while the mechanism for creating and serving those ads remained unresolved in the captions.
- 0:21:10
Choosing an Interface for AI-Assisted Coding
The speaker says the choice between a CLI and a desktop app depends on background, workflow, and available constraints. A comfortable CLI remains viable, while the speaker personally prefers the desktop experience and describes Codex as a strong quality-oriented option.
- 0:23:07
Start Product Ideas with Problems Around You
The speaker recommends observing real problems in one’s own life and surroundings, then considering ways to solve them—even when the problem is small. In his view, AI can suggest ideas, but it is not a strong substitute for human observation and thinking during this creative stage.
- 0:25:48
Design a Dashboard Around the Next Step
The excerpt explains that a dashboard should direct merchants to the application’s key pages and reduce uncertainty about what to do next. The source moves on before the explanation is completed.
- 0:26:45
Turning a Recurring Everyday Need into a Software Idea
The speaker suggests that X may surface ideas, then illustrates the method with someone who wants to manage a workout program without pen and paper; that need could inspire software that generates a refreshed plan. The core approach is to define the problem and desired outcome before imagining the solution.
- 0:28:38
Strengths and Limits of “Command Center”
In the speaker’s assessment, “Command Center” is useful for initiating the basic structure of an idea or project, but less suited to detailed work requiring repeated AI dialogue and deeper implementation. He says adapting it for that use may be possible.
- 0:31:46
An Autonomous Commerce Agent with Merchant-Visible Activity
The application combines content and topic discovery with an agent designed to perform ongoing work for a merchant. Trust is treated as a prerequisite for broader autonomy, so the autopilot interface shows completed work, planned actions, and an activity log. Structured business knowledge and guidance also shape how the agent writes and optimizes products.
- 0:35:06
Reviewing AI Product Optimizations Before Publishing
The demonstration shows a flow that scores product-page elements, generates improvements to titles, descriptions, and structured data, and lets the merchant review or regenerate changes before publishing. Deeper business context is described as a way to strengthen the agent’s suggestions, while the demo was running outside the faster production environment.
- 0:38:17
A Qualified View on Using Remotion for Video Generation
Answering a question about video tools, the speaker describes Remotion as a promising option in a video-generation workflow, while clearly qualifying that the recommendation is not based on extensive personal use and is partly informed by others’ demonstrations and recommendations.
- 0:39:48
A Setup Panel for the Post-Onboarding First Impression
The discussion requests a mockup for the home page merchants reach after onboarding, including a setup panel that guides important next steps. Because the dashboard forms the first product impression, its layout and flow should be evaluated before implementation.
- 0:41:04
Balancing Product Quality Against Feature-Scope Expansion
The speaker reflects on how difficult it is to stop adding features when the product still feels short of the desired quality bar for users. The observation captures a practical tension between delivering sufficient value, continuing to improve the product, and preventing each new idea from opening another major development track.
- 0:47:44
Configuring the App Embed and Discoverability Settings
The work identifies two important setup areas: enabling the app embed and reviewing discoverability and crawler settings. This suggests that setup should guide merchants through activating required integrations rather than merely presenting a generic dashboard.
- 0:48:23
Start AI-Assisted Coding by Building Tools for Yourself
Responding to a beginner, the speaker advises against making income the immediate goal when experience is limited. He suggests using Codex to build useful things for oneself first; practicing on real personal needs can broaden ideas, after which tools for communities or clients—and possible monetization—can be considered.
- 0:51:23
Validating AI Search Access Permissions
A research task was proposed to verify whether enabling the AI search access tab requires theme write permission, determine whether the existing scopes are sufficient, and examine alternatives if the additional permission is necessary.
- 0:55:48
Why an AI Coding Agent’s Evaluation Is Not Enough
The speaker describes iteratively auditing and fixing a storefront with an AI model. Although its report eventually showed an apparently perfect assessment, a brief manual check exposed a bug. He attributes this to possible mismatches between the model’s interpretation and the intended behavior, arguing that the human remains chiefly responsible for clarifying intent, debugging, and validating the result.
- 1:02:32
Are AI crawler permissions redundant in Shopify?
A review of whether the AI search and crawler-access page provides anything Shopify does not already handle, including a request to check all eight crawlers and verify any real difference before removing the feature.
Got a question about something in this stream?
Ask in the group