Claude and GPT compete to design my Website
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 preparing Burnish’s website as an early part of its go-to-market strategy: defining the brand essence and customer journey, communicating the product’s value within seconds, and testing landing-page messages and lead capture before the application is finished. Luqman practically compares Claude, GPT, and Codex for brainstorming, planning, and implementation, using visual references, screen recordings, and a context-aware judge; his preferences and impressions about model quality and efficiency are presented as personal observations, not general tests. The work includes reviewing the visual identity, palette, and temporary logo, balancing bold motion against clarity, and exploring Remotion and image-to-video workflows without documenting a completed final result. Audience discussions cover pricing, subscriptions, free offers, market selection, local models and tool security, and business ideas such as an AI trading bot, with emphasis on validating demand, margins, acquisition, and retention rather than merely building. A separate exploratory discussion considers Egypt’s El Dabaa nuclear power project under construction, its reported connection to the Russian company Rosatom, and a then-discussed target for the first of four plants or units to begin generating electricity in 2028; these details were mentioned in the conversation and not independently verified. Decisions about email gating versus deposits, some sales tools, the final design, and several technical experiments remain unresolved, while stream stutter and encoding lag were investigated without a confirmed fix.
78 moments from the stream
- 0:04:29
Testing the Market Before the Product Is Finished
The speaker explains why marketing should begin early to test market interest and explore expected pricing, allowing the findings to shape a product that better fits the market.
- 0:06:18
Turning Brand Essence into a Cohesive Landing-Page Journey
The discussion covers defining a brand’s voice, vision, purpose, and problem, then translating them into an immediate first impression and a consistent visual and narrative journey from the top of a landing page to its footer, with reference to a site inspired by Greek mythology.
- 0:08:00
A Collaborative Workflow Using Claude, GPT, and Codex
The speaker outlines a planned workflow in which AI models brainstorm the website, brand decisions are condensed into a shared handoff prompt, and proposals are compared before Codex handles the coding execution.
- 0:10:21
Designing an Outbound Sales Pipeline with ROI in Mind
He explains that choosing sales tools is a strategic operator decision and outlines a general pipeline: source leads, verify email addresses, then run cold outreach. He stresses evaluating paid tools against expected return on investment while noting that some choices still require research.
- 0:14:22
Comparing Claude and GPT for Website Design and Adding Visual Context
The plan is to have Claude and GPT propose or compare website ideas, then use Codex to build the selected direction. Because code access alone may not fully convey animations and interactions, he records the site to give the AI additional visual context.
- 0:16:26
Troubleshooting YouTube Stream Lag and Encoding Delays
He investigates an unstable stream: OBS and frame delivery initially appeared normal while YouTube reported insufficient video, and skipped frames from encoding lag appeared later. He suspected and stopped the screen recording, but the issue persisted, so the cause and fix remained unresolved in these pages.
- 0:33:04
Diagnosing Stream Stutter During Screen Recording
YouTube showed an excellent stream status, and OBS showed no dropped frames, but its statistics reported skipped frames due to severe encoding lag. The speaker attributes the problem to starting an additional screen recording while OBS was already recording, disrupting frames sent to the stream.
- 0:37:26
Reducing Agent Milestones and Planning Before Website Execution
In response to an audience question, the speaker says that a large milestone count is not inherently a problem and that the agent can be asked to reduce it. He then discusses establishing a tangible website plan and prefers adding the milestones to the command center before execution begins.
- 0:40:09
Screen Recording Can Affect Frame Rate
Luqman noted that he had not realized screen recording during a stream could affect frame rate, while discussing efforts to resolve streaming issues.
- 0:40:31
Preparing a Shared Prompt to Compare Claude and GPT on Website Design
Luqman prepared a detailed handoff prompt for Claude and GPT to tackle competitively, using a reference website and a clear quality benchmark with the aim of producing a stronger design.
- 0:43:29
The Importance of a Careful Roadmap
Luqman emphasized the need for a carefully prepared roadmap because it will guide the entire workflow.
- 0:47:11
The Website as a Foundation for Go-to-Market Strategy
Luqman explained his decision to prioritize the website before expanding technical development, treating it as the foundation for marketing and as a place that communicates the project's value and converts traffic into sign-ups. He also distinguished inbound sales, where prospects arrive themselves, from outbound sales initiated through direct outreach, which he described as costlier but potentially producing greater results.
- 0:50:24
Using Remotion with Codex for Programmatic Video
The speaker explains installing the Remotion plugin in the Codex app and referencing it in a prompt to animate a logo or create a marketing video. He describes Remotion as a way to build video through code with AI agents, while noting that his own hands-on experience with it is still limited.
- 0:52:07
Dividing Work Between AI Tools
The speaker gives the tools the same inputs so they can produce independent ideas to compare and score before translating the result into milestones. As an opinion and expectation, he suggests Claude may be stronger for brainstorming while preferring Codex for execution because he perceives it as more token-efficient and able to handle more work.
- 0:55:06
Turning Brand Strategy into a Website Clear in Three Seconds
The speaker describes documenting the brand essence, voice, philosophy, problem, audience persona, mission, vision, and storytelling approach, then giving those materials and the existing website code to AI tools. The aim is to develop the logo concept, site sections, and marketing direction so a visitor can understand within roughly three seconds what the site does and whom it serves.
- 0:57:59
Having AI Study References Before Proposing a Design
The speaker says the prompt instructs the tools to study a reference website, a screen recording, the codebase, and the desired effect before making proposals. After anticipating different outputs, he moves to try a GPT-generated prompt for logo creation; this page does not show the test result.
- 1:00:12
Comparing Claude and GPT Logo Ideas for Burnish
The speaker reviewed different logo proposals prompted through Claude and GPT, explaining that the visual direction connects Burnish’s name with polishing and shine. He regarded the gleam effect as a reasonable start without declaring a final result.
- 1:02:32
Choosing an AI Tool by Workflow Stage
The speaker described his personal practice of giving Claude and GPT the same prompt and handoff context. He said Claude might perform better at reasoning and brainstorming in this session, while he would still choose GPT for execution because it fits his working habits; this was a personal preference, not a general verdict.
- 1:05:05
Moving from Brainstorming to Executable Work
The speaker chose to reduce time spent reading lengthy brainstorming outputs and use a judge for comparison, aiming to move more quickly into practical building. He noted that both proposals were based on the prompt and screen recording supplied to the models.
- 1:07:17
Using a Judge to Compare Design Philosophies
The speaker drafted a request for a judge to assess both models’ responses against the brand essence, compare their philosophies and ideas, and identify overlap. Based on his personal experience, he treated rare agreement between independent brainstorming outputs as a useful signal—not established truth—worthy of further consideration.
Got a question about something in this stream?
Ask in the group