Inside the Core Engine
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 an in-development Shopify app intended to improve stores’ visibility on AI platforms, with a capability audit exposing gaps between the published plan matrix and actual implementation—especially APIs, webhooks, and multi-store agency support. Luqman demonstrates a review workflow combining specialist agents for security, compliance, and UX with hands-on testing, human-controlled merges, and automated deployment; mobile inspection revealed visual defects, confirming that the product was not finished. The discussion also covers free-versus-paid report design, A/B testing the placement of email capture, and an audit-led acquisition funnel whose channels should be measured before ad spending. Audience questions explore Claude, Codex, agent skills, and conversation-context compaction, with advice to master reliable tools instead of chasing every new model. On GEO, the speaker distinguishes it from SEO while leaving the effects of AI-generated articles, backlinks, and llms.txt unresolved because the evidence is not conclusive. The practical conclusion is to align public claims with real capabilities, understand integrations, and test the complete experience before Shopify submission; the completion date, best conversion flow, and need for external monitoring remain open.
109 moments from the stream
- 1:41:22
Ahrefs APIs and a Backlink Audit Idea
The speaker says he does not use the Ahrefs APIs and is considering a backlink-audit capability that is not currently in the application.
- 1:41:48
Backlinks Across SEO and GEO
The speaker considers backlinks valuable for SEO but stresses that SEO and GEO overlap only partly. He remains unsure whether backlinks help GEO directly or indirectly and says further research is needed.
- 1:44:19
Would Paid Prompt Tracking Add Friction?
A proposal asks customers to pay to specify prompts tracked across multiple engines, while explicitly questioning how it would work and whether it would add unnecessary friction. The captions do not preserve a clear response.
- 1:45:38
Evaluating llms.txt for GEO
The speaker describes trying an AI-oriented file and reports that ChatGPT flagged its self-promotional nature. Based on his research, he found no reliable evidence that llms.txt improves GEO, presenting this as his assessment rather than a settled fact.
- 1:46:53
Choosing Ad Channels by Cost per Lead
The discussion notes that both Google Ads and Facebook Ads can generate leads. The practical decision is to compare cost per lead rather than favor a platform by name, while recognizing that the best channel depends on what is being sold.
- 1:49:34
Testing Email Capture Before a Scan
After choosing to keep the current setup simple, the speaker proposes later testing email capture before a scan instead of beginning with only the store URL and collecting email in a later flow. The A/B test is deferred until the current setup is complete, followed by a request for a phased implementation plan that preserves the existing design where possible.
- 1:53:28
Success Benchmarks and Gradual Growth in GEO
The speaker argues that major success in generative engine optimization requires reaching the scale of market leaders, while stressing that direct comparison with much larger companies is unrealistic and progress happens incrementally.
- 1:54:50
Building Payment Plumbing Without a Manual Stripe Login
The speaker sets a clear implementation constraint: they cannot log into Stripe at that moment, so payment-plumbing work can proceed only if the required portion does not depend on their manual account login.
- 1:55:49
External Monitoring Compared with Sentry and Shopify Tooling
The speaker discusses a service that periodically checks sites and can alert by email or phone when they fail, distinguishing it from Sentry’s stronger focus on backend bugs and issues. They also review Shopify’s native visibility into webhook failures, response times, errors, and logs, deferring the decision on external monitoring until actual gaps become clearer.
- 1:56:58
Protecting a Shopify App’s Reputation at Launch
The speaker emphasizes close support for early users and the importance of earning good reviews, since negative feedback from early merchants can damage a Shopify app’s reputation from the outset.
- 2:02:13
End-to-End Uptime Monitoring for a Shopify App
He discusses periodically checking application services for outages, covering both frontend and backend rather than only the server. Since this is his first Shopify app, he plans to examine Shopify's built-in options before deciding whether a third-party monitoring service is needed, while stressing that monitoring is important.
- 2:03:35
Running an Automated Frontend Workflow
He briefly shows an automated-mode workflow operating on the frontend of a project whose name is unclear in the captions, then indicates that the task completed.
- 2:04:14
Impressions of Maximum-Effort AI Coding
He enthusiastically describes his personal experience using Claude with a high-effort setting and feels that the work quality improved. This is a subjective impression rather than a documented benchmark, with private usage and account details omitted.
- 2:11:19
A Personal Discussion of Claude and Codex
He reacts to a discussion about Claude’s strength and offers an informal comparison involving Codex, but the precise conclusion is unclear in the captions and is not based on a systematic test.
- 2:11:54
Codex or Claude for Building Applications?
The speaker says Codex can build strong applications from scratch, while his personal preference for Claude comes from greater familiarity with how to use it; he frames the choice as subjective.
- 2:12:35
A Custom Stream Screen and Preflight Sequence
The speaker explains that the screen in question is a simple branded visual, and that he had previously created a short preflight screen before launching the stream but later found it unnecessary.
- 2:16:03
Multi-Agent Milestone Review Before End-to-End Testing
The speaker describes an application milestone reviewed by four specialized agents, including compliance, agent-experience, and security reviews. Their findings are being fixed before an in-app verification walkthrough. In parallel, the top of the funnel is still changing and is intended to be tested end to end once ready.
- 2:20:14
Why Application Completion Is Hard to Estimate
The speaker says he cannot reliably estimate when the application will be finished: whenever completion seems close, more tasks appear. Some milestones have become outdated and are no longer worth completing, while the referenced QA section has seen little use in his view.
- 2:21:17
A Dashboard for Tracking the Marketing and Sales Funnel
The speaker presents a new dashboard section for tracking leads, spend, revenue, scan counts, and qualified leads. He also notes that the cohort strategy has been reworked or replaced in parts, mentions planned pricing for a full audit report, and remains undecided about whether spend will be shown in the dashboard.
- 2:21:57
A Unified Dashboard for App and Market Validation
A proposal for a database- and CRM-connected page that shows workflow and app potential, supporting market validation and evaluation of the advertising funnel before adding payments.
Got a question about something in this stream?
Ask in the group