Working on Merchant Facing UI
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 preparations for a Shopify app submission, prioritizing completion of the move from Polaris React to Polaris Web Components and visual and functional QA of the merchant interface before returning to website, report, and funnel polish. Practical work also covered GA4, Google Tag Manager, and PostHog configuration; verifying that changes reached production through GitHub and Cloudflare; reviewing Supabase table organization; and troubleshooting MCP access—without treating visible identifiers alone as proof of end-to-end tracking. Luqman drew lessons from accumulated Git branches and worktrees and from coding-agent completion claims that failed manual inspection, arguing for regular integration and direct verification. He also discussed automated build-and-audit agent loops, his personal impressions of Codex and Claude, and audience questions about deployment, model choice, project languages, client acquisition, programmatic SEO, Kanban boards, and Hermes. Open questions remained around some Cloudflare and MCP failures, the best database organization, the value of particular agent tools, and the final submission date.
104 moments from the stream
- 3:37:50
Command Center Redesign After Moving to Polaris Web Components
The speaker prioritizes improving the UI before further backend polish and asks for a comparison between the current command-center dashboard and a website reference, followed by a plan before implementation. He says the move from Polaris React to Polaris Web Components temporarily set the interface back, while relating it to efforts to better satisfy Shopify app-submission requirements and policies.
- 3:45:07
Building Shopify Interfaces with Polaris Components
The speaker hopes Polaris Web Components will make dashboard construction as straightforward as the earlier React experience while still using Shopify’s required components.
- 3:48:14
Proactive Follow-up After Implementation Checks
The speaker argues that checking a task should not end with calling it incomplete or partially complete; the next steps or an implementation plan should follow, with straightforward work handled directly when appropriate.
- 3:50:07
AI Commerce Index Refresh CTA and Weekly Limit
A command-center CTA is reconsidered because it does not truly generate a report; it refreshes the figures, trend, and AI Commerce Index used to populate the interface. It should be used during onboarding and then made available once per week, while the backend mechanism is deferred to a separate task and documented for later.
- 3:54:03
A Perceived Regression in Codex Performance
The speaker described a perceived decline in Codex completing some coding tasks, based on an informal comparison with Claude using the same repository and commands. This was a personal observation, not a controlled test.
- 3:57:06
Reading Coding-Model Benchmarks Carefully
The speaker discussed reported differences in coding-model benchmarks, warning that vendor pages can be selective and may give less prominence to coding-task results. He suggested Artificial Analysis as a relatively independent source while noting that benchmark methodology takes care and learning to interpret.
- 4:00:15
Qualified Interest in Trying Hermes Agent
The speaker said he had not yet tried Hermes Agent but intended to evaluate it later for work such as customer service and marketing. He relayed, rather than personally verified, a favorable comparison with OpenClaw and noted that avoiding the need for a second computer or VPS could simplify adoption.
- 4:04:52
A Kanban Board for Coordinating Multiple Agents
Explains the idea of using a Kanban-style task board to distribute work across a team of agents instead of keeping it in one conversation, with a request to explore how such a system would look and operate.
- 4:05:59
Reviewing the Merchant Dashboard During a Polaris Web Components Migration
Reviews a merchant-facing dashboard against a wider reference design, criticizing its muted colors and compressed layout. The discussion asks whether page width is constrained by Polaris Web Components while stressing that the aim is to work within them and continue the migration from Polaris React rather than abandon it.
- 4:12:47
Should an Agent Board Be a Mission Control or a Command Center?
The discussion tentatively distinguishes a “mission control” from a “command center,” suggesting that a multi-agent Kanban board may be closer to mission control and asking what capabilities it would include beyond the task board itself.
- 4:16:47
Exploring a Kanban Board and Workspace Integration
The discussion covered a Kanban board said to support Git worktrees, workspaces, and session continuity after automatic resume, with a suggestion to find a video demonstrating its setup and appearance. The tool identity and some wording are unclear in the captions.
- 4:18:10
Polishing the Command Center Layout and Verifying a Font Change
The interface width was considered acceptable if it was the available maximum, while polish and alignment still needed attention. Adding AI-engine logos beside engine names was also proposed, before prioritizing the font fix. After a hard refresh and comparison with an earlier screenshot, no visible font change was found.
- 4:31:19
Visual Priorities for the Command Center
The discussion shifted to the command center’s visual design, treating the inclusion of the dashboard component as the key requirement even if the final form would not exactly match the current view.
- 4:32:21
When to Use Hermes and How to Divide Work Between Agents
The discussion notes that continuously running Hermes may consume many tokens, so routine work could go to a lower-cost default agent while agents such as Codex serve as specialists. It stresses that the first question is why Hermes is needed at all: the speaker does not want it to fully automate building because they want to remain closely involved, while marketing and customer service appear to be possible uses that still require more research.
- 4:34:19
Checking the Merchant UI Layout with Real Data
A discussion about reviewing the layout against a configured development store so it shows real data rather than error banners, while stressing the need to observe current behavior instead of assuming production data will populate the interface correctly.
- 4:36:48
Temporary Access-Token Expiration
A practical question about why an access token expires after one day and whether that would prevent access or a DNS change several days later. The captions do not establish a clear final answer or decision.
- 4:39:54
Hermes Kanban Capabilities for Agents
Hermes Kanban is described as supporting persistent multi-agent work, database-backed storage, multiple boards and workspaces, and agent collaboration. The speaker considers Hermes stronger, while other comparison details are fragmented in the captions.
- 4:42:07
Reviewing Visible Interface Statuses
A quick review of a result showing states such as ‘created’ and ‘pending,’ ending with the preliminary judgment that it was workable. Some other labels remain unclear because the captions are fragmented.
- 4:44:03
Refining AI Engine Cards in the Hero Section
A hero-section review identified a duplicated “AI Commerce Index” label, with proposals to rename it and place engine logos in the cards instead of color dots for clearer visual identification.
- 4:50:17
When Programmatic SEO May Help New Domains
The speaker argued that programmatic SEO may become more useful for organic traffic after a new domain passes a perceived Google “sandbox” period, estimated—without certainty—at roughly three to nine months.
Got a question about something in this stream?
Ask in the group