RIP. NEW MISSION. CC V2.0
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 redesigning a largely autonomous “Command Center” that routes coding tasks by size and dependency, delegates research, exploration, and implementation to agents, and introduces task- and milestone-level audits before human review. Luqman demonstrates unfinished practical work across Codex, Claude, and terminal workflows, weighing output quality against token consumption and usage limits while encountering real friction involving Supabase, Docker, MCP, and stalled tasks; the workflow is therefore not presented as a completed success. The discussion broadens to unit testing, lightweight feature specifications, competitor research before building, and the need to market products rather than stopping after implementation. He also outlines a Shopify tool concept for measuring and improving product visibility in AI answers through GEO, which he described at the time of the 2026 recording as an emerging field and forecast to grow, particularly toward late 2026, while leaving market viability, implementation complexity, and the GEO–AEO distinction unresolved. Drawing on experience that began with dropshipping in 2017, Luqman compares dropshipping with DTC and white labeling, e-commerce with app economics, and B2B with experimental B2C directions; he also offers opinions on SaaS margins and recurring revenue, reasons for choosing the Shopify ecosystem, and app marketing through defined audiences, acquisition channels, and a mix of organic and paid approaches, without presenting these views as guarantees or settled outcomes. Secondary discussions cover automated short clips and the risk of lost context, recorded streams as a work archive and community-learning channel, programmatic marketing videos with Remotion, thumbnail experiments, and copyright constraints. Audience questions add useful discussions about learning software development through practical AI-assisted projects, verifying the security and provenance of agent skills, estimating time, managing dependencies, model pricing and compute efficiency, and scheduling work around session limits. The qualified conclusion is that automation may reduce manual coordination but does not remove the need for oversight, debugging, and validation; Luqman argues that strong planning and context matter more than minimizing token use alone.
155 moments from the stream
- 2:06:15
What Is Generative Engine Optimization?
The speaker described Generative Engine Optimization (GEO) as an SEO-like concept for improving brand and product visibility across AI-powered search and assistant channels. He noted that some search behavior was shifting toward AI queries and described the field, at the time of recording, as still relatively new in Shopify.
- 2:09:28
Correcting a Supabase Port Value in an Environment File
The speaker said a Supabase startup issue was tied to how a port value was written in the environment file, and that restoring a different representation allowed the server to read the port.
- 2:11:16
Why Supabase MCP Helps Agents
The speaker says MCP makes Supabase work easier by letting an agent execute SQL within it on his behalf, shortening the workflow, while noting that he had not yet been using the dedicated Supabase skill.
- 2:13:42
Discovering and Installing the Supabase Skill
While exploring the Supabase dashboard and its Connect section, the speaker found and installed a dedicated skill. He expected it to help agents understand Supabase when they become slow or confused about its workflow. The discussion also covered the installer's warning to review skills before granting full agent permissions, with first-party availability treated as a trust signal rather than a demonstrated guarantee.
- 2:17:12
Skill Plugins for Brainstorming
The speaker noted that a skills marketplace exists and said the Superpowers plugin was especially useful to him for brainstorming.
- 2:19:58
Phishing Caution and Agent-Skill Sources
After phishing was mentioned, the speaker explained that he generally avoids downloading skills and plugins indiscriminately from GitHub, preferring to build his own tools or use sources he considers trustworthy. He then pointed to dedicated marketplaces for discovering agent skills, while the name of one such marketplace remains unclear in the captions.
- 2:21:18
Finding Agent Skills and Checking Their Sources
Skills can be browsed, searched, and sorted, but abundance does not imply trustworthiness; users should check who published a skill and what it contains before adopting it. Building a custom skill is also possible, and prebuilt skills are not essential to every workflow.
- 2:23:36
Moving from AI Product Building to Marketing
The goal is not merely to build applications, but to take them to market and repeat the build-and-market cycle. Marketing experience can be an advantage while learning to build because AI tools may assist with implementation and automation without fully replacing market understanding.
- 2:27:20
Crowded Market or New Category?
Existing competitors do not necessarily make an idea unattractive; they can demonstrate demand and an opportunity to win market share. A solution with no precedent is a double-edged proposition because the lack of competitors may also mean absent demand and a need to educate the audience first.
- 2:28:38
Designing an Orchestrator and Sub-Agent Workflow
The discussion considers an orchestrator that assigns work to sub-agents and reviews their output, either delegating autonomously or following a predefined workflow. The described approach ties deployment to task complexity, keeps simple work with the orchestrator, and adds research, exploration, and post-build auditing roles for larger tasks to preserve cleaner context and clearer control.
- 2:31:19
Failure as Experience That Enables a Better Next Step
The speaker frames failure not as a negative endpoint but as experience that supports a better next step; unsuccessful attempts can be useful when they produce learning and forward movement.
- 2:33:50
Installing an App and Trying Its Shopify Preview
The attempt follows instructions to refresh the app, open it from the admin preview, visit Shopify Partners, and install the application. The captions do not confirm that the preview ultimately worked.
- 2:38:48
The Muted-Microphone Meeting Mistake
The speaker realizes he had been talking while muted and notes that the same mistake commonly happens in online meetings such as Google Meet.
- 2:39:29
Tracing a Supabase-Related Server Error and MCP Setup
While using an auditor agent to review a work milestone, the speaker suspects—but does not confirm—that the server error is related to Supabase tables. He also explains that the installed MCP skill requires restarting the session before it can be read.
- 2:43:36
Choosing Another Coding Tool When a Task Stalls
The speaker practically compares AI-assisted coding environments and says he feels more comfortable with Warp despite liking other interfaces. When a small task remained stuck in Codex, he chose to try another tool because subsequent work depended on its completion.
- 2:45:42
Reviewing Task Dependencies for Greater Flexibility
The speaker requests an analysis of the links and dependencies among 32 tasks, aiming to remove only dependencies that are not genuinely useful, preserve accurate ones, and suggest ways to make execution more flexible.
- 2:46:39
Switching Tools or Models After a Persistent Stall
After the attempt remains stalled, the speaker decides to close it, start a fresh session, run the server himself, and return to a more comfortable environment. This illustrates a practical response to tool friction, without evidence in the captions that the alternative solved the problem.
- 2:58:26
Learning to Build Software Practically with AI
Responding to a question about progressing from beginner to professional software building, the speaker acknowledged a personal bias against relying on courses alone. He favored learning the fundamentals, understanding how to use AI, then building projects from scratch and learning through problems over time. He also suggested YouTube material and discussion with an AI-focused community as supporting resources, presenting this as personal advice rather than a guaranteed path.
- 3:01:53
Programming in the AI Era
The speaker argues that the key skill now is not writing every line manually but learning to work effectively with AI, while acknowledging disagreement around that view.
- 3:05:55
Color-Coded Agent Workflow States
The speaker briefly explains that task colors represent different working modes, including one requiring collaboration with the agent and others associated with agent-led or automated work; some mode names are unclear in the captions.
Got a question about something in this stream?
Ask in the group