Optimizing Sarmad
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 developing Sarmad as an interconnected agent system: defining roles and delegation paths, monitoring token usage, experimenting with Telegram-based control and reporting, and reviewing the interface to improve visual separation. It also demonstrates automated content creation and cross-platform publishing; only the social workflow had been verified, while other channels still needed refinement. A major theme is that building a product is insufficient without distribution. Luqman discusses traffic acquisition, measuring content impact, competitor research, gradual ad testing, and a possible shift from Sarmad’s free trial to a free tier. He stresses that limited traffic and data prevent firm conclusions about pricing or conversion, and that cost reductions should not override privacy or healthy margins. Audience questions cover choosing tools such as Claude and Codex based on practical efficiency and usage limits, learning through real projects, AI-engineering paths, marketing, freelancing, and workstation choices. Several ideas—including open-sourcing Sarmad, proactive agent hierarchies, traffic projections, and revised pricing—remain proposals or early tests rather than completed outcomes.
116 moments from the stream
- 2:46:45
The Economic Value of Problem-Solving
The speaker explained that solving a real company problem can become paid work because the value comes from addressing a problem the company needs resolved.
- 2:47:16
Choosing an AI Tool by Project
The speaker noted that Codex is not necessary in every case and that the choice of tool or model should depend on the project.
- 2:48:16
Mac mini, Mac Studio, or Laptop?
The speaker discussed choosing hardware for development with Codex while acknowledging that hardware is not his strongest area. He leaned toward a Mac Studio for higher performance, said a Mac mini could still be sufficient, and found a desktop with a large display more comfortable for substantial projects. He also advised keeping a laptop for travel and mobility when practical.
- 2:51:34
An Unresolved Usage-Consumption Observation
The speaker observed that a model he called GPT-6 Sol did not appear to consume usage as expected and explicitly wondered whether this was a bug. No test or explanation established the cause.
- 2:55:03
When Should the Prompt Be Launched?
The speaker asks whether a prompt can be launched immediately or requires answers first, and requests a concise English description of what it will do. The prompt’s name is unclear in the captions.
- 2:58:26
Mentorship Boundaries and Willingness to Help
In response to a mentorship request, the speaker says he cannot provide personal mentorship but remains open to helping and supporting one another.
- 2:59:41
Explaining a Decision and Using Multiple Agents
The speaker requests a concise plain-English explanation before deciding whether to lock or loosen something, then asks multiple agents to examine different areas of a target whose name and nature are unclear in the captions.
- 3:04:09
Trying to Make the User Experience Psychologically Compelling
The speaker describes a creative aim to make the experience psychologically compelling to users; the brief fragment does not establish the implementation method or results.
- 3:04:33
French as a Bridge to Learning Spanish
In a casual language discussion, the speaker expresses the view that knowing French can make Spanish easier to learn.
- 3:06:22
Traffic and impressions as the bottleneck
The speaker believes one part of what is being discussed is largely solved and identifies traffic and impressions as the main challenge.
- 3:07:39
Providing runway without delaying conversion
The speaker stresses that users should not have to wait two weeks before converting, while proposing a two-week runway structured so it does not delay conversion itself.
- 3:08:16
Reaching French-speaking viewers in North Africa
The speaker explains that the French conversation reflects viewership from young people in North Africa, where French is widely spoken.
- 3:09:43
Adding a fast decision-making tool to a workflow
The speaker proposes inserting a not-yet-tested tool into selected parts of a workflow to speed up decisions and reduce hesitation, while clarifying that it is not a full system replacement and the idea remains hypothetical.
- 3:14:35
Reviewing an inbox interface implementation
The speaker says the implementation is unfinished and asks for comparison with the attached screenshot, pointing to an issue involving the inbox section and its apparent similarity to the header; the page ends before the explanation is completed.
- 3:15:25
Visually Differentiating Sarmad’s Interface Sections
Direct interface feedback notes that the sections look too similar and calls for more precise color choices to distinguish the header and sidebar, while observing that the implementation still needs cleaner refinement.
- 3:18:30
Rising Memory Usage with a Large Session Folder
A viewer described memory usage increasing on every request while working with a large session folder until the coding application crashed. No firm diagnosis was offered; Luqman later noted his limited hardware expertise and that he had not encountered the issue himself.
- 3:19:09
Planning Traffic Acquisition for a New App
Answering a question about marketing an app through the App Store or Google Play, Luqman qualified his limited experience with store traffic, suggested Product Hunt as one possible channel, and recommended using AI to help formulate a structured acquisition plan rather than relying on a single tactic.
- 3:20:14
UI Constraints Inside Shopify Apps
The discussion explains that a general design may suit an independent website, while Shopify apps follow a defined system using Polaris components and Polaris React, so an arbitrary interface cannot simply be inserted unchanged.
- 3:24:21
Limited Experience and a Positive View of Firebase
Luqman said he had used Firebase the previous year on work handled mainly by a team, so he did not claim extensive recent hands-on experience, but he expressed a positive view of the service and noted that it belongs to Google.
- 3:26:06
Using AI to Develop a Traffic Strategy
The discussion suggests having a serious conversation with AI tools about marketing and traffic acquisition, including potentially using them to help run ads. This is presented as general advice, not a guaranteed plan.
Got a question about something in this stream?
Ask in the group