Chill Build: Automation & 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 building a controllable automation system that organizes AI agents as a graph and turns research into content, short clips, and multi-platform posts, supported by a calendar, analytics, reports, and Telegram-based coordination. Practical work included checking the publishing pipeline, correcting topic relevance, refining tone and calls to action, and dividing reasoning tasks from deterministic local scripts. Luqman repeatedly emphasized that quality, measurement, and human review should precede scaling or unattended operation. In his view, the larger challenge is no longer software construction alone but acquiring and measuring traffic through tools such as Search Console and Google Analytics, then improving the product and pricing once enough evidence exists. Audience questions covered coding tools and plans, Codex versus Claude, learning through Micro-SaaS, agent tools, e-commerce platforms, SEO, editing, and AI video generation. Other discussions covered opinions on AI safety and the software industry’s future, beginner and freelancing advice, creator-platform monetization, payment-platform choices, the risk of exposed API keys, and the ethics of AI-altered product images. Claims about unidentified models, agent security, usage limits, and tool costs were treated cautiously rather than as established facts. The automation, publishing, and traffic outcomes remained unresolved: some tests and refinements looked promising, but several workflows stalled or were still being configured, and there was not yet enough performance data to call the system successful.
131 moments from the stream
- 0:03:58
Discussing a New AI Approach and Its Claims
An initial discussion of a newly encountered AI tool or approach whose name is unclear, asking whether anyone had tried it and suggesting a possible trial. Claims that it was faster, cheaper, and might avoid hallucinations through a different output method were reported as hearsay, without testing or verification.
- 0:04:33
Debugging a Social-Media Posting Automation System
A plan to debug an active automation system and evaluate whether its agents can continue social-media posting with remote monitoring. This was a proposed test, not a demonstrated success.
- 0:07:31
Mixed Experiences with a Coding Assistant’s Usage Limits
The speaker reported mixed feedback about a coding assistant’s usage limits: some users were satisfied while others were frustrated, leading him to wonder whether the experience varies by user. No cause was established.
- 0:10:09
Graph Engineering for Higher-Level Control
A brief update described the current focus as a graph-engineering system intended to replace an existing bot and provide a higher level of control. Component names and some implementation details remain unclear in the captions.
- 0:15:27
Building Is Easier Than Selling: The Traffic Problem
The speaker observes that many developers can now ship products and software but still struggle to sell them. He identifies traffic acquisition as the practical problem he is trying to solve, presenting this as his experience rather than a proven universal rule.
- 0:16:01
An Untested Description of a Different AI Approach
The speaker gives a tentative, unverified account of an unidentified AI service that he believes uses a different type- or TypeScript-related approach, alongside claims about speed and reduced hallucination. He explicitly says he has not tried it and knows no one who has, so both the interpretation and claims remain uncertain.
- 0:20:56
Evaluating a Hosted Multi-Agent Workspace
The speaker describes testing a hosted workspace where agents had separate roles, instructions, contexts, computing environments, inter-agent communication, and mobile access. He liked the prospect of work continuing away from his computer, but found token consumption high and concluded that he needed greater control.
- 0:27:19
A Trial Blocked by Closed Sign-Ups
The speaker tried to access an AI service whose name is unclear, but sign-ups appeared to be closed, possibly because of an influx of new accounts. He therefore postponed the trial and did not claim to have completed a test.
- 0:28:34
Adding Music to an Automated Video Pipeline
The speaker discusses adding background music to short clips produced by an automated pipeline that also includes voice-over. He says he was using Suno but, at the time of recording, believed it offered neither an API nor MCP access and therefore required computer control, prompting a search for an alternative that could run while idle or traveling. He described the shown clips as fully automated; this is his characterization, not an independently verified result.
- 0:31:01
Layered Content Production and Automated Publishing
A workflow combines video, captions, voice-over, and an added music layer before publishing content regularly across multiple platforms. The speaker acknowledges that the initial quality is limited and can be improved over time.
- 0:33:55
Designing an Integrated Traffic and Measurement System
The speaker outlines a marketing system spanning affiliate marketing, inbound and outbound traffic, and SEO, while stressing that publishing alone is insufficient: improvement, traffic sources, and content performance should be tracked. They also observe that reduced content activity may coincide with weaker search visibility, without presenting a causal test.
- 0:39:25
A Central Calendar for Tracking Shorts and Their Impact
The speaker describes a workflow that processes stream material into short clips and publishes them across multiple social platforms. A calendar inside the system supports reviewing scheduled posts and opening them on their destination platforms, with emphasis on impressions, views, and resulting traffic rather than comments alone.
- 0:41:04
Analytics for Measuring Automated Content Performance
The speaker proposes an analytics section that tracks agent usage and work progress, allowing the AI to identify content that attracts the most reach and traffic. In his view, publishing across many platforms can help growth, but quality may decline without careful monitoring.
- 0:44:20
Human Oversight for an Automated Content Pipeline
The speaker describes an automated pipeline that researches topics, writes scripts, adds voice-over, music and captions, and publishes the result. He is not yet comfortable posting to personal accounts without review, and considers video generation more in need of improvement than image or text generation.
- 0:48:58
Improving Traffic Before Pricing and Product Polish
After a relative decline in visibility, the speaker says he is postponing product polish, pricing review and introductory offers until traffic improves. His current approach is to increase publishing while protecting quality, because he believes product or pricing changes will have limited effect without stronger market visibility.
- 0:50:28
Offering Product Trials to Interested Users
Responding to an audience question, the speaker clarifies that an existing product trial is available to people who want to test it, after the original request was unclear about whether it concerned account links or trial access.
- 0:53:17
Trying an Unidentified Model Through OpenRouter and OpenCode
Luqman tried to access an AI model whose name is unclear, mainly to experience its response speed and see whether it appeared less prone to hallucination, rather than to use it for coding. A prior website trial was mentioned, but the captions do not document a completed test or confirmed result.
- 0:53:34
Audience Question About an Unidentified Tool
An audience member asked about a tool whose name is unclear, and Luqman directed them to open its website on a desktop and review its short introduction. This page does not itself explain what the tool or service does.
- 0:58:50
An Uncertain Description of a Model’s Design
Luqman relayed—while explicitly saying he did not fully understand it—an unclear explanation claiming that the model’s design differed from that of typical large language models. The details and model identity remain uncertain because the captions are unclear.
- 1:04:37
Testing Tone Classification and Resistance to Misleading Instructions
The speaker describes testing a classifier that changed its tone label from “frustrated” to “calm” after the scenario wording was adjusted. In another test, the selection grounded in structured state remained stable despite misleading text embedded in a system-update field. He viewed the result as promising while noting that broader uses still needed exploration and a concrete use case.
Got a question about something in this stream?
Ask in the group