TALKSTORE PIVOT
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 TalkStore’s public build after Shopify rejected its storefront domain: related code and references were removed, documentation and the roadmap were revised, and the app began repositioning around product operations. Proposed directions include an assistant that can act on Shopify tasks, understand product images and text, improve content in bulk, and potentially support GEO; these remain unproven ideas, while the product’s replacement advantage, defensibility, and market demand still require research and validation. In practical work, Luqman experiments with a command center that distributes tasks across Codex, Claude, and other tools while tracking dependencies, activity, and reviews. Unresolved issues include task-status synchronization, usage limits, and the cost of parallel agents; his tool comparisons are time-specific impressions rather than controlled tests. He also discusses separating engineering from debugging, QA, and manual pre-submission review, and establishing an adaptable billing foundation while plans and pricing remain provisional. Audience questions broaden the discussion to validating demand before building, software economics, distribution and B2B adoption, and dependence on platform rules. The main supported conclusion is that coordination, market validation, and distribution matter as much as build speed, while the product, workflows, and tool experiments remained works in progress rather than completed successes.
68 moments from the stream
- 1:02:20
An AI Assistant for Shopify Product Operations
The speaker reviews a thin product page and describes an assistant that could analyze images alongside text, improve, remove, or generate product images, create descriptions in bulk, and detect contradictions in product data. These are presented as proposed ways to reduce manual product-management work, not as tested results.
- 1:06:06
Synchronizing a Roadmap with Project Execution Boards
The speaker dictates a structured request to update the project manifesto and roadmap first, derive any additional tasks, and then revise the command center, swim lanes, and task board, including removing storefront scope that is no longer in the plan.
- 1:09:37
Custom Apps vs. Shopify App Store Apps
The speaker distinguishes building a custom app for one business from distributing an app through the Shopify App Store, where platform rules apply. He also broadly compares the concept with Shopify Sidekick while aiming for wider action-taking capabilities.
- 1:12:00
Persistence and Pivoting When a Product Direction Stalls
A personal view favoring steady work and experimentation, treating obstacles or rejection as reasons to pivot rather than stop. The speaker says AI has made building several things in parallel feel more practical, despite previously preferring to focus on one thing.
- 1:15:21
Removing the Storefront and Refocusing on Product Operations
A recap of completed work: sweeping the codebase to remove storefront-related code and references, then discussing additional domains and quickly brainstorming stronger product operations, with agents asked to update the swimlane to reflect the new decisions.
- 1:18:28
Testing Codex Inside the Command Center Workflow
The speaker plans to test a new workflow using Codex inside the Command Center cycle. He had previously explained task preparation and task start, but explicitly says the integrated workflow had not yet been tested at this point.
- 1:21:26
Distributing Work Across AI Coding Tools
The speaker discusses spreading development tasks across Codex, Claude, and tools such as Cursor to avoid relying on one tool and to manage usage limits intelligently. In his view, the value of paid tooling depends on the project’s clarity and its marketing and distribution needs; this is presented as an opinion, not a demonstrated result.
- 1:25:48
Trying Codex on a Command-Center Workflow
The speaker begins a first attempt to use Codex to modify a command-center workflow, including removing two unneeded lanes and then monitoring the agent’s status and speed. He says a successful deployment could help distribute work across tools, but the captions show an experiment in progress rather than a confirmed successful deployment.
- 1:32:02
Creating Logical Task Dependencies and Priorities
A request to inspect subtasks, connect current and later work through logical dependency and blocking relationships, and assign numerical values indicating execution order.
- 1:33:39
Trying Codex and Comparing Speed, Context, and Usage
During a practical trial, the speaker observed Codex agents, the creation of an instruction file, and the start of a task. He tentatively found execution faster and apparent usage lower than another tool, while reserving judgment until output quality was clear and acknowledging that the usage reading or his procedure might be mistaken.
- 1:38:24
Task-Status Synchronization Gaps in an Agent Dashboard
The speaker noticed that an agent had started working without the task changing to “in progress” in the agent dashboard. He described status synchronization as an implementation detail needed for a more reliable command center, while noting that the current version did not fully match his intended design.
- 1:44:55
Managing Multiple Coding Tasks in Parallel
The speaker demonstrates running several coding sessions and tasks concurrently with Codex, using a task board to track active and in-progress work. In his view, the workflow is useful but demands concentration to avoid losing track of tasks.
- 1:48:26
Comparing Context Usage and Reviewing GitHub Complaints
Drawing on his own experience and opinion, the speaker compares context consumption in Codex and Claude, then refers to a GitHub ticket dated March 23 containing user complaints about reaching usage limits quickly and delayed support. The year is not stated, and some ticket details are unclear in the captions.
- 1:51:47
Placeholder Values for a Pay-As-You-Go Structure
Questions covered the pay-as-you-go cap and when to show the subscription prompt. Luqman explained that the current numbers were placeholders pending a profitability review and asked for an implementation that would remain easy to revise later.
- 1:53:48
The Stream’s Premise: Building Products in Public
Answering an audience question, Luqman described the stream as a transparent, live view of building applications and workflows rather than only short tool comparisons, alongside an effort to form an Arabic vibe-coding community for exchanging ideas.
- 1:55:39
A Change-Friendly Billing Foundation
The discussion considered showing free-tier status on the build page and designing plan cards. Luqman asked for the leading recommendations to be applied as a solid billing foundation while stressing that pricing, offers, and the structure were still provisional and could change.
- 1:57:45
A Time-Specific Claude vs. Codex Trade-Off
Answering a beginner’s tool question, Luqman said a usage issue he had encountered with Claude since March 23 led him to recommend Codex temporarily until it was resolved, while still personally preferring Claude’s output quality. This was his time-specific opinion, not a controlled comparison.
- 1:59:38
Deferring Debugging to a Dedicated Workflow
After noticing a bug, Luqman chose not to debug it immediately because it required focused time and thought. He preferred to prepare a dedicated task and debugging workflow while continuing the project’s rapid pass.
- 2:03:29
Running Multiple Agents Despite Token-Usage Limits
The speaker describes using an IDE-like environment to run multiple agents in parallel terminals, allowing them to share context and review one another’s work. In his view, the workflow works well, but token consumption is a bottleneck; he says the issue had persisted since March 23, with no clear year given in the captions.
- 2:05:12
A Command Center for Large Projects, from Build to Distribution
The speaker recommends building a command center or “mission control” for a large project to clarify the roadmap, organize tasks, track what was completed or blocked, and check whether outputs meet the intended standards. He argues that its value extends beyond faster building to planning the project’s later marketing and distribution.
Got a question about something in this stream?
Ask in the group