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
- 2:54:59
Building an SEO Foundation Before Adding GEO
The speaker argues that SEO and generative engine optimization (GEO) overlap substantially, suggesting a strong foundation in one before layering on the other. He says current SEO demand is much higher and offers a personal forecast of rising GEO interest around September 2026, based on search trends he says he observed.
- 2:57:21
Sequencing Bug Fixes and Frontend Polish
Raises a practical question about completing phase one with a bug sweep before beginning a dedicated frontend-polish phase, and whether competitor data is sufficient to accelerate that work. The captions do not establish a final answer or implemented decision.
- 2:59:30
Consulting Service or Dashboard Product?
Compares offering SEO improvement as a consulting service with building a dashboard where customers sign in and scan their stores. A service is described as easier initially, while the speaker considers a dashboard potentially stronger long term but more code-intensive and dependent on careful competitor research.
- 3:04:51
Reaching Competitor Parity in Version One
A practical view that version one need not outperform immediately: matching competitors can itself be meaningful progress and create a base for a later version to surpass them.
- 3:06:10
Designing a Standalone GTM Dashboard
A review of an operator-facing metrics view proposes replacing the placeholder with a standalone GTM dashboard that defines what operators need to see, including cohort fill, aggregate revenue, traffic sources, and an “other” category for unexpected sources.
- 3:07:14
Evaluating Arabic Output in Content Tools
A discussion raises concerns that Arabic content quality may be weak or costly despite high expectations for the tool. It is presented as an evaluation concern requiring comparison of actual output, not as a documented test result.
- 3:12:05
Assessing the AI Customer-Service Market
An assessment of a platform offering round-the-clock customer service through AI agents. The speaker considers the broad business need attractive but warns that the market is highly saturated and competitive.
- 3:14:46
Integrating the GTM Dashboard into the Command Center
Luqman clarifies that he wants the GTM dashboard to appear as a section inside the command center, alongside existing Burnish sections, rather than as a separate dashboard embedded in the website. The discussion distinguishes where the interface is displayed from the secrets, passwords, or API keys that may still be required to retrieve its data.
- 3:20:47
Using One Coding Agent to Review Another Agent’s Work
Luqman describes a workflow in which Codex has worked on bug fixes while Claude reviews the work, pages, and visible controls as he shifts attention to GTM tasks. The passage presents this as practical agent-to-agent oversight, without claiming a completed test or final validation.
- 3:29:19
Does Token Usage Differ Between AI Models?
A discussion of how token consumption and efficiency may vary between AI models even for the same task. Experience-based opinions are offered about relative usage and output quality, while the speaker explicitly says the underlying mechanics are unknown and considers that a larger token count might still cost less.
- 3:35:27
Following a Deployment on Railway
During practical work, the speaker explains that they do not know how to run commands directly on Railway, then lets the deployment proceed and waits for the result. The captions do not establish that the deployment ultimately succeeded.
- 3:43:04
Preparing a Handover Prompt for a Coding Agent
After an agent stopped progressing on a task, a handover prompt was requested so it could receive the necessary context and resume from where it left off, without exposing the task or project identifiers.
- 3:53:58
Verifying an Application with an Agent in Chrome
Luqman explains that he asked the agent to resume verification in Chrome on a development store. In his account, the MCP-based setup lets the agent inspect the application, capture screenshots, interact with it, and recheck three earlier bugs. He describes the relatively new capability as highly useful to him while noting that it consumes many tokens.
- 3:58:54
Choosing a Large Monitor for Work
A practical opinion that a larger monitor provides more workspace and becomes comfortable with use, while returning to a smaller display may later feel restrictive.
- 3:59:25
Learning SEO and GEO Together
Advice to study search engine optimization and generative engine optimization together because the fields overlap and share related skills. The speaker expects GEO to grow, while emphasizing a foundation in marketing and the use of AI as a learning aid.
- 4:03:20
Using YouTube and AI to Learn a New Field
A suggestion to combine practical experiences from YouTube and public discussions with conversations in ChatGPT that explain a field and compare it with a related one; the conversational approach may be especially useful for people who prefer reading.
- 4:04:17
A Two-Agent Debugging Loop
A workflow in which Codex performs debugging and Claude reviews the result; if the fault remains, the reviewer sends specific feedback back to Codex for another attempt. This reflects the speaker’s experience, not a general proof that the approach succeeds.
- 4:09:55
A Unified Go-to-Market Metrics Dashboard
A proposal for a protected dashboard that uses APIs to consolidate key go-to-market metrics instead of checking several separate analytics and operations tools, making important measurements easier to track.
- 4:17:08
Protecting Secrets in a Public Repository
He flags that a public repository used by others is not an appropriate place for private API keys or an environment file, and describes the practical tension created by using the same project publicly and personally.
- 4:22:08
Browser-Assisted Debugging and Review with Claude
He explains a development workflow using Claude in Chrome, which he says was still in beta at the time. In his account, it can inspect the page, developer tools, and network activity, capture evidence of issues, and relay findings to Codex for fixes. He also describes using it for competitor research and identifying website technologies.
Got a question about something in this stream?
Ask in the group