Talkstore Begins Now! 🟢 Building to $100K
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 documents the public start of building Talkstore: a proposed Shopify-specific AI assistant through which merchants could manage products, storefront work, and commerce intelligence conversationally. Its capabilities and possible expansion beyond Shopify are presented as conditional directions, not validated outcomes. Practical work covered Shopify app and development-store setup, Supabase integration and Prisma migrations, secret handling, and repeated diagnosis of session, connection, preview, and Railway deployment problems. The preview and some setup steps worked, and the service later appeared active, but several database, network, build, and deployment issues remained intermittent or unresolved in the covered segments. This led to useful lessons about starting with logs, documenting a stubborn problem, seeking a fresh diagnostic perspective, and reviewing coding-agent compromises rather than accepting a result merely because it works temporarily. The stream also presents a multi-agent command center that separates research and task preparation from implementation, code, security, and compliance review, using dependencies, an adaptive timeline, and human approval gates. Luqman felt this structure accelerated and preserved the context of the work, while still questioning whether every agent was running and visibly logged. He emphasized that authentication, secrets, and nuanced decisions continue to require human intervention, and that task count is a poor measure of difficulty. Work later shifted toward pre-launch distribution: a simple but polished Astro waitlist site kept separate from the Shopify app, with email capture, analytics, social metadata, a teaser, and a distribution campaign. The first landing-page draft had missing hero content and styling or rendering failures, so it was treated as an unfinished iteration. Other useful discussions included audience questions about using AI to explore income paths from existing expertise, SaaS unit economics, choosing between ads and SEO/GEO, Arabic-market WhatsApp automation, and targeting external markets, plus local voice dictation, content automation, and Telegram community options. Stream lag and interruptions were also investigated without a conclusive cause, while music licensing and gym culture appeared as brief secondary discussions.
121 moments from the stream
- 2:16:19
Debugging an Unfamiliar Deployment Pipeline
The speaker discusses slow, iterative debugging of an apparently simple deployment problem and says unfamiliarity with Railway's pipeline limited his ability to help. He contrasts this with greater familiarity with Vercel and a less satisfying prior Cloudflare experience, noting that Railway was chosen on a recommendation suited to the project and possible later expansion.
- 2:17:52
Using a Fresh Session for a Stubborn Bug
The speaker suggests opening a fresh session or model when debugging stalls, seeking an independent perspective and a different problem-solving approach. Prior analysis was passed to another agent for explanation, with the explicit limitation that Docker was unavailable and an actual Docker deployment could not be run in that environment.
- 2:19:21
Checking Railway Environment Configuration
During the investigation, a suggestion emerged to verify that the Shopify app URL was set in Railway environment variables before redeploying. The captions support this only as a diagnostic suggestion, not a confirmed cause or successful fix.
- 2:24:28
Watching a Previous Crash Point After a Change
A new attempt is watched to see whether it passes the section where it had previously crashed. Running beyond the usual failure point is treated as a sign that something may have changed, but not as proof of a completed deployment.
- 2:26:24
Seeking Another Perspective on a Persistent Deployment Failure
After the Railway deployment remains stuck despite lengthy attempts with multiple agents, the speaker decides to seek a new perspective. He contrasts the difficulty with prior deployments on Vercel and Cloudflare and acknowledges staying focused on the issue despite other available work; the page records no final success.
- 2:33:30
Tracing a Deployment Failure at the Network Stage
Luqman follows a delayed Talkstore deployment and reports that the build finally completed but the process failed during a network check or handshake. He tries to understand the cause so it does not recur, then retries through a Git push while the precise cause remains unresolved.
- 2:33:58
Limits of an AI Operator on Detailed Technical Tasks
Luqman explains that an AI operator can accelerate work when the problem is understood, but is harder to guide through development and configuration details outside the user’s expertise. He also treats its schedule estimate cautiously because unexpected technical blockers, such as deployment issues, can change the timeline.
- 2:40:32
Clarifying a Task Before Starting Parallel Work
While waiting for deployment, Luqman moves to another task to use the time productively. For a collaborative task, he adds a note to the task-preparation command asking the agent to explain its understanding and clarify exactly what is required before proceeding.
- 2:43:47
Creating a Supabase Project and Choosing a Region
Creates a new Supabase project, considers the hosting region based on the current target audience and possible future expansion, and explains Supabase as a cloud database for application data, messages, and other information as needed.
- 2:48:14
Troubleshooting Deployment and Generating a Public Railway Domain
Tracks repeated deployment failures and the waiting cost of each attempt, then navigates Railway service settings to find public networking, distinguish it from private networking, and reach the option for generating a public domain.
- 2:55:51
Configuring Shopify Variables and Redeploying the Service
Adds the Shopify key, secret, app URL, and permission scopes to project variables, then investigates why deployment did not start automatically and proceeds toward a manual redeploy. The discussion remains uncertain about parts of the variable setup and how project-level values carry into the service.
- 3:01:34
Connecting Supabase and Running a Prisma Migration
Returns to the Supabase project to identify the database connection string, then asks the assistant to run the Prisma migration directly if the steps are straightforward. These captions do not confirm that the migration or deployment succeeded.
- 3:03:38
Human–Agent Collaboration and API Key Safety
The discussion explains that some setup tasks require a human operator to remain in the workflow, especially when handling API keys and environment variables. It warns against exposing secrets publicly because they could enable misuse of APIs or data access, noting that some AI assistants alert users when a sensitive key is shared.
- 3:05:34
Expanding Agent Activity Logs
While reviewing the Supabase connection, the speaker notes that the agent hub records work time and activity for roles such as the orchestrator and builder, but apparently not for several other roles including research, review, and security. He suggests expanding the logging to cover those activities as well.
- 3:08:03
Diagnosing a Stalled Deployment Instead of Guessing
After a prolonged deployment failure, the speaker criticizes continued guesswork despite using Codex and Claude, arguing that a fundamental setting may have been overlooked or the problem overcomplicated. He asks for the Railway configuration and error cause to be researched before another solution is proposed.
- 3:10:27
No-Code WhatsApp Automation for the Arab Market
The captions describe an Arabic-oriented platform for visually building WhatsApp conversation flows, sending bulk messages, and using AI for replies. The speaker considers the idea promising in the Arab market because many businesses rely heavily on WhatsApp, citing possible uses in clinics, real estate, restaurants and reservations, and e-commerce.
- 3:14:12
Reviewing Talkstore Tables in Supabase
After a task was submitted for review, the work moved to verifying the Supabase tables and checking the main production table, while deferring some identifier or connection cleanup until later.
- 3:19:20
Linking Shared Variables to Services
The diagnosis indicated that shared variables are unavailable to services until explicitly linked to them, an issue that caused delay and frustration during implementation.
- 3:22:00
Ads vs. SEO and GEO Based on Margin
The speaker discussed choosing a marketing channel according to available resources and profit margin: ads may be considered when margins allow, while SEO and GEO (Generative Engine Optimization, described here as SEO for AI) were presented as options when marketing funds are limited.
- 3:23:53
A Brief View on Meta Ads in MENA
Luqman briefly described Meta Ads as highly effective in the MENA region. The remark appears in a fragmented list about blogs, tutorials, and marketing channels, so this page does not establish his reasoning or conditions.
Got a question about something in this stream?
Ask in the group