Burnishing the 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 centers on refining Burnish’s landing page so it clearly explains GEO and its value to Shopify merchants, while turning the page into an acquisition funnel that offers a free scan rather than demanding an immediate subscription. Reviews cover the hero, pricing, product explanation, calls to action, and mobile experience, alongside unfinished visual experiments involving WebGL, GSAP, an interactive logo, and a light-reactive grid; several results remained short of the intended design. The speaker argues that clear messaging, a provisional V1 focus on GEO, and one ideal customer profile matter more than endless polish, while acknowledging that scope, pricing, and willingness-to-pay assumptions still require validation. The discussion also addresses Shopify review risks and the need to begin marketing and market validation before approval, since building alone does not ensure merchant acquisition. His Codex and Claude comparisons are explicitly personal: reusable skills and critique sessions may improve the starting point, but they do not replace a clear vision, repeated review, or early checks of repository, branch, and implementation plans. In audience discussions, he advises beginners to build a small personal project with Codex and also covers WebGL, GSAP, usage limits, agents and automation, and the economics of software and dropshipping, presenting tool comparisons and market forecasts as tentative opinions rather than proven findings.
71 moments from the stream
- 0:03:29
Finishing the Landing Page Before Marketing
The discussion treats Burnish’s interface as a necessary foundation for starting its go-to-market work, allowing the product to be presented and marketed before the full build is complete.
- 0:04:12
AI Skills Do Not Eliminate Iteration
A practical reflection on design and review skills used with Codex and Claude. The speaker says they can provide a useful starting point but do not replace repeated iteration, and offers his personal assessment of how the two tools performed.
- 0:05:57
Every Website Section Should Convey a Message and Feeling
He explains that landing-page sections should not be random decoration: each part should communicate something about the product and create an intentional impression for the merchant, even when the message works indirectly.
- 0:07:17
Balancing UI Polish with Product Progress
The discussion argues for limiting UI-polish time and avoiding perfectionism: begin with an initial version that can evolve, then move on to the application and its backend engineering concerns.
- 0:08:41
Narrowing V1 to GEO
He presents a provisional decision to focus V1 on GEO rather than combining it with SEO, despite overlap between the fields, to avoid expanding scope too early. He says the decision will be revisited after the landing page is finished.
- 0:09:18
Reviewing Pricing, Explanation, and CTA Sections
He reviews the landing-page structure: the pricing content is provisionally acceptable but its design needs reworking, another section should explain what the application does, and the call to action should define the intended visitor outcome.
- 0:09:46
Turning the Landing Page into a Value-Exchange Acquisition Funnel
He proposes using the landing page as an acquisition funnel that does not demand an immediate subscription. Instead, it offers merchants a free store scan and information about their visibility across AI answer channels in exchange for appropriate lead details, followed by progression through the marketing funnel.
- 0:14:48
Explaining GEO as a New Category on a Landing Page
The speaker argues that every line on a landing page should serve a purpose in persuading the merchant. In the speaker’s comparison of U.S. Google Trends data for the preceding five years, GEO’s interest curve was far below SEO’s and appeared from mid-2025; he therefore argued that the page should explain the category and its value to Shopify merchants clearly and without excessive copy.
- 0:17:46
A Practical Way to Start with AI Using Codex
Responding to a question about getting started with AI, the speaker suggests installing Codex and using it to build a small personal project, then learning progressively by exploring what it can enable.
- 0:18:32
WebGL, GSAP, and Video Generation for Richer Interfaces
The speaker considers a visual direction combining WebGL and GSAP with AI video generation as tools that could make a website more visually appealing. It is presented as a possibility, not as an implemented or tested result.
- 0:19:35
Splitting Website Improvements into Focused Sessions
The speaker organizes work into separate labeled sessions, with each session focused on a specific website section. The example is dedicating one session to the pricing page or section and continuing until it reaches the desired standard before opening another session.
- 0:21:12
Comparing the Grill Me Experience with Claude and Codex
The speaker gives a personal impression of using the Grill Me tool with Claude and Codex: Codex appeared to provide shorter responses and rely more on successive questions, while Claude seemed to offer longer paragraphs and begin a deeper discussion. This is the speaker’s anecdotal experience, not a systematic test.
- 0:21:56
Considering a GSAP Resource for an Animation Problem
A GSAP-related resource or “skill” was suggested as possible help with an animation problem. The speaker planned to download and inspect it after finishing the current section, but the captions do not establish its capabilities or the outcome of any trial.
- 0:25:32
Polishing LuqDex and Explaining the Cost of Ignoring GEO
The speaker reviewed the landing page structure: the hero was considered nearly complete and in need of only minor polish, while the following section required more thought. Its stated purpose was to show merchants the potential harm of not focusing on GEO today.
- 0:27:39
Discovering Wrong Branch and Repository Context and Reviewing the Agent Plan
The speaker discovered that work had been running on the wrong branch and then the wrong repository, which he described as wasted token usage. After waiting for Claude’s result, he judged the plan weak, noted that he had not reviewed it beforehand, and attributed the poor output to that before stopping the task. The discussion highlights checking repository and branch context and reviewing an agent’s plan early; it is the speaker’s assessment of one experience, not a formal comparison.
- 0:32:55
Restarting After a Repository Reset
The speaker explains that he restarted from scratch after ending and relaunching a Codex session changed the repository state, to avoid wasting time or tokens on the wrong path.
- 0:35:23
Polishing the Website Within a Phased Launch Plan
The speaker describes improving the website step by step so the landing page and interface express the product vision before marketing, while parts of the backend remain unfinished. He frames this within development, review, and later marketing phases rather than treating the current UI as final.
- 0:37:28
A Command Center Where Agents Execute a Roadmap
Recapping the work, the speaker describes a command center that contains the project roadmap and lets agents execute tasks under user supervision and tracking. He then says the current work has moved to a Shopify app related to GEO, a field with some similarity to SEO.
- 0:39:04
A Clear Vision Supports Consistency
The speaker argues that consistency and motivation depend on having a very clear vision: the clearer the intended outcome, the easier it is to sustain momentum. This is presented as his personal view, not a proven rule.
- 0:39:55
Why the WebGL and GSAP Implementation Stalled
The speaker attributes the agent's stalled work to a weak implementation plan and demanding WebGL and GSAP effects, rather than blaming the tool alone. He notes that waiting on the stalled agent also blocked his own workflow.
Got a question about something in this stream?
Ask in the group