Inbound vs Outbound Traffic
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 a widening gap between building products with AI—which is becoming easier—and the harder work of distribution, customer acquisition, and sales. Luqman reassesses a heavy reliance on inbound content and considers a broader mix of outbound outreach, cold email, advertising, partnerships, and SEO, with the practical principle of testing channels and concentrating on those that show real conversion signals. Practical work includes reconsidering Burnish’s pricing by replacing a time-limited trial with a free plan, investigating a reported HTTP 500 error and an observability gap, and simplifying AI-agent workflow diagrams; the records do not establish that these efforts were completed or successful. The discussion also explores multi-agent systems organized around shared goals, context, memory, and explicit follow-up, with a preference for starting small instead of introducing early complexity. In comparisons of Codex, Claude, and AI providers more broadly, Luqman emphasizes balancing quality, speed, price, usage limits, and completed workload, while repeatedly qualifying his views as personal and treating mixed reviews and benchmarks as insufficient for declaring an absolute winner. Audience questions broaden the stream to SaaS development, AI-engineering portfolios, content automation and scheduling, visibility in AI-generated search answers, team size, personal branding, transparent affiliate marketing, and AI support for study or business management. Major unresolved questions include the best acquisition channel, the commercial value of content, the right agent architecture, and the changing choice between tools and subscription plans.
133 moments from the stream
- 4:20:57
How Can a Pharmacy Student Use AI?
In a tentative response, the speaker suggests that AI may help a pharmacy student study and understand scientific and theoretical material. He offers no clear suggestion for earning income through the field, and the answer ends before additional practical uses are developed.
- 4:23:35
AI as a Study Aid in Specialized Fields
Responding to a question about AI in a specialized field, the speaker said he did not know a specific professional use case, but suggested using it to support studying and seeking input from people in the same field.
- 4:25:10
Context Limits for an Agent Managing Other Agents
A conversational test showed that a top-level agent could not answer questions about the structure beneath it unless that data was available in the current context, even though the text interface appeared to have broader visibility. The speaker regarded connecting that context to voice as a solvable limitation.
- 4:27:56
Diversifying Content Formats Across Social Platforms
The speaker considered expanding the plan beyond videos and short-form clips, asking which formats suit each platform and what could be published on days without video posts.
- 4:30:53
Checking the Status of a Previous Workflow
Before proceeding, the speaker asked whether a previous workflow had been paused, deleted, or was still running, highlighting the need to verify legacy automation state before building on it.
- 4:31:52
Reassessing the Inbound Traffic Direction
Reviewing the recent direction, the speaker concluded that the problem was not the tools, technical work, or automation, but the approach used to generate inbound traffic, which he described as the wrong direction in its current form.
- 4:33:17
Diagnosing a 500 Error While Connecting a Store to Burnish
A user reported a 500 error while connecting a store to Burnish during a major update. Luqman explained that the status usually indicates a server-side problem, requested a screenshot and the exact failing step so it could be reproduced, and asked for verification of what had reached production.
- 4:42:16
Understanding AI Coding Tool Costs and Usage Limits
Luqman discussed a positive experience with an AI model’s speed and usage limits, while remaining puzzled by published cost tables—particularly cache pricing—and uncertain whether API-based comparisons map directly to usage inside coding tools.
- 4:46:15
Combining Codex and Claude in a Coding Workflow
Luqman suggested diversifying a coding workflow across Codex and Claude when both are available, presenting the combination as a practical alternative to relying on only one tool.
- 4:48:07
Treating AI Model Benchmarks with Caution
Luqman said AI benchmark results are not easy for him to interpret and that understanding them deeply takes time, so he remained uncertain about how much confidence to place in the published comparisons without further analysis.
- 4:48:49
An Observability Gap in a Burnish Installation Error
Luqman returned to an error encountered by a store after installing Burnish, noting that its source was unclear because Sentry had not reported it, highlighting the diagnostic difficulty when monitoring fails to capture an incident.
- 4:54:28
Building a Portfolio for an AI Engineering Career
In response to a graduate asking how to enter AI engineering, the advice was to start with small software projects built with AI and present them in a portfolio that demonstrates practical ability, while noting that the graduate's academic background was unknown.
- 4:57:06
Contributing Through Beta Testing
Beta testing was suggested as a practical way to contribute: keep using the application for a while and report any errors or problems that emerge.
- 4:58:52
Work Progress and a Free-Plan Pricing Hypothesis
The work recap mentioned giving outbound traffic more attention, while the tool choice remained unresolved. The speaker also presented replacing a free trial with a free plan as a personal pricing hypothesis that might accelerate revenue.
- 5:00:45
Testing Multiple Acquisition Channels, Then Doubling Down
The speaker proposed testing a broad mix of acquisition channels, including inbound traffic, outreach, influencers, programmatic SEO, and advertising. Once a channel shows the first signs of acquiring customers, the suggested next step is to focus on it more deeply.
- 5:04:40
AI-Assisted Web Design from Reference Sites
The speaker describes asking AI to capture much of the codebase of reference websites, placing it in his repository, and then modifying the site, content, and presentation. He distinguishes this from simple imitation and says the approach is faster and cleaner for him than generating a design from scratch, while acknowledging that AI tools may initially refuse on copyright grounds. This is his personal workflow and opinion, not a controlled test.
- 5:09:16
A Software Agent Designed for Broader Use
After a brief trial and a question-and-answer call, the speaker describes an unnamed tool as an always-available agent that can follow tasks rather than merely serving as a direct substitute for other tools. He considers it easier to use and potentially suitable for a broader, nontechnical audience as well as developers.
- 5:15:51
How Task Acceleration Changes the Productivity Gap
The speaker explained that reducing a long task to a small fraction of its usual duration could radically change the pace of work, and argued that people able to afford such tools may advance faster than others. The cited speed multiplier was approximate, not a documented test result.
- 5:18:23
Choosing Between Codex and Claude—and Whether to Consolidate
The speaker discussed the difficulty of settling on Codex or Claude as each tool’s capabilities and offers change, saying he generally leaned toward Codex while remaining uncertain. Using both was considered, though he found consolidating work in one tool more comfortable and treated price and efficiency as key factors.
- 5:27:17
Choosing an AI Tool by Use Case
The speaker says tool choice is subjective and depends on the use case. He personally leans toward Codex because he prioritizes efficiency and believes the quality gap between Codex and Claude is small; this is an opinion, not a documented benchmark.
Got a question about something in this stream?
Ask in the group