Building Unfair Advantages
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 follows the ongoing development of Burnish as a GEO application in the Shopify ecosystem. A central workflow uses Claude in Chrome to study competitor products and designs, then passes findings and review feedback to Codex for implementation and debugging, with human inspection still required. Luqman also discusses preparing the go-to-market foundation before spending on ads: analytics and attribution, landing-page, headline, and pricing experiments, lead qualification, and a unified metrics dashboard. He makes clear that many forecasts and experiments were not yet complete. Practical work exposes late-stage difficulties, including a mobile-performance regression after some changes, followed by a clearly measured improvement after a later deployment; whether that gain justified the design compromises remained unresolved. The work also covers deployment and Git-branch friction, pull-request checks, accessibility reviews, and Shopify requirements. Some deployments succeeded, while other technical questions remained open. Audience discussions broaden into multi-agent coordination and usage limits, learning SEO, GEO, and development, the growing importance of marketing and distribution as AI makes building easier, and choosing a familiar market. A substantial secondary thread covers content and audience strategy: batching short-form content or temporarily pausing it to focus on building, building in public, weighing Arabic against North American audiences, YouTube monetization and sponsorships, and livestreaming as both a focus aid and a possible distraction, alongside rest and burnout prevention. The stream documents an unfinished process of development, research, and market preparation rather than a completed launch, with several decisions still requiring measurement and review.
151 moments from the stream
- 8:10:03
Designing a clearer branches and worktrees workflow
He described the friction of Git branches and worktrees when work overlaps or a coding agent unexpectedly switches branches. He proposed a unified system for branches, commits, pull requests, and deployments that protects the main branch, reduces workflow confusion, and keeps attention on building the application.
- 8:21:57
Evolving a Git Workflow as a Project Grows
The speaker explains that a simple commit button stopped being sufficient as the project grew and became more interconnected. He plans to understand the workflow better before designing a clearer system for merges and pull requests.
- 8:23:31
Improving Text Contrast on a Privacy Page
The discussion attributes the privacy page's unreadable body text to white type on a pearl-colored background and suggests changing the font color as a simple fix.
- 8:26:07
Understanding GitHub Checks Before Merging a Pull Request
Luqman pauses before using squash and merge because some checks appear incomplete, and wants to understand the GitHub workflow that examines branches and pull requests before a push or merge.
- 8:27:34
Revisiting an Agent Task After an Unintended Change
Luqman revisits an earlier development task that was dropped after an AI agent made an unintended change and removed part of the work. He then considers assigning the follow-up to Codex after its current task, while GitHub surfaces an additional issue.
- 8:35:32
Merging a Pull Request After Checks Pass
The speaker discusses using squash and merge for a pull request, leaving another fix for a separate PR, and requesting a plain-language explanation of what happened and what the PR does. He then checks GitHub and waits for the checks to pass before merging, based on his understanding.
- 8:36:43
Adapting a Coding Agent’s Effort Level to the Project
The speaker says he changes Claude’s effort level according to the project and prompt, while preferring to keep Codex at a high-effort, high-speed setting. This is his personal preference, not a measured comparison.
- 8:44:37
Creating a Contextual Handoff Between Coding Agents
The speaker plans to finish the current task and then request a handoff prompt for another agent. He asks the agent to recover where the earlier Codex work stopped and identify the next step, preserving task continuity across agents.
- 8:47:00
Inventorying Active Branches Before Continuing a Merge
After referring to several design-polish branches, the speaker asks for an exact list of branches that remain active and have not been pushed or merged, clarifying the repository state before continuing.
- 9:00:30
Coordinating Claude and Codex Under Usage Limits
The host explains preserving work context by providing a handover prompt and the agent’s latest interactions, and describes the Claude–Codex combination as productive in his view. He also notes that usage limits can interrupt the workflow and that he adjusted Codex’s browser-access configuration to broaden its ability to perform tasks.
- 9:03:29
Reviewing Website Accessibility and Desktop/Mobile Performance
The host reviews an agent’s work on WCAG 2 website accessibility and reports that the check found no violations, while stressing that the visual result still needs inspection before acceptance. He also discusses page-load and mobile-performance improvements, then checks the site locally across desktop, mobile, and pages such as pricing, privacy, terms, and contact, identifying the footer as remaining work before opening a pull request.
Got a question about something in this stream?
Ask in the group