One AI Department Done. 3 to Go.
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 work-in-progress automation system that organizes AI agents like connected departments for research, coordination, content production, publishing, and analytics. Luqman explains a possible progression from scheduled tasks and role-specific instructions toward shared goals and indicators, with research handed to agents that adapt posts, articles, and short videos for each platform. Practical work includes reviewing the post-stream pipeline, trimming silence and adjusting clips, and preparing operations across LinkedIn, X, Facebook, and Instagram while retaining human review initially. Speed, quality, resource use, queues, measurement, and agent autonomy still require testing and refinement; the stream does not establish a successful end-to-end system. It also discusses distribution and building in public, comparing direct outreach with SEO, social content, and advertising, alongside the speaker’s view that understanding an audience and solving a real problem matter more than selling a generic chatbot. Audience questions broaden the discussion to changing developer roles, learning and working with AI, game development, video-editing tools, background or local agent execution, community platforms, tools, and subscriptions. Other cautious, unresolved topics include AI jobs and pricing, proposed Egyptian data centers, model and hardware performance, YouTube Shorts growth, and ideas for competitions and products; these are generally presented as opinions, proposals, or open questions rather than verified findings.
171 moments from the stream
- 1:40:37
Building an Agent-Activity Graph from Execution Records
A proposed architecture records each agent action as evidence of activity and resource use, then uses those records to generate nodes and links in a graph. This was an architectural explanation, not a documented test result.
- 1:45:59
An Agent That Reviews Page Visuals and HTML for SEO
The discussion considered an agent that reads a page section by section and combines its visual appearance with HTML, elements, and tags to suggest practical SEO improvements when major design changes are unavailable. A recurring need might suit a persistent agent, while a one-off audit might require only a single run, but the task requirements were acknowledged as too unclear for a definitive answer.
- 1:49:14
Speed and Quality Trade-offs in Local Model Use
A brief, inconclusive discussion covered running models locally, including the speaker's impression that some options may be slow and deliver middling quality, with suitability depending on the task. No benchmarks or verified tests were presented.
- 1:49:43
What Happens When Clip Reviews Accumulate?
A workflow question asked what happens when a new stream is picked up before clips from the previous stream have been reviewed: does the work accumulate in a queue? The page contains no answer, so the system behavior remains unspecified.
- 1:50:28
Question About the Best Channel for B2B Client Acquisition
A question asked which marketing approach works best for acquiring B2B clients and whether direct outreach is the appropriate channel. The excerpt ends before any answer or recommendation is given.
- 1:50:39
Direct Outreach vs. Traffic Generation for B2B
The speaker considers direct outreach the fastest route to B2B customers, while noting that it can be costlier, more cumbersome, and dependent on sales skill. He contrasts it with more automated traffic generation through SEO, social media marketing, and ads, and mentions Apollo, email collection, and phone calls as direct-outreach methods.
- 1:56:23
Separating Conversations from the Processing Queue
After observing that the system appeared stuck while it was processing work without reporting its status, the speaker proposes separating conversations from the processing queue so interaction with AI workers can continue during background work. This is presented as a design requirement, not a completed test or verified success.
- 1:57:29
Directing a Time-Driven Musical Introduction
The direction asks for the track’s introduction to feel majestic and magical, with a ticking-clock sound synchronized seamlessly to the music, followed by an immediate transition whose exact description is unclear in the captions.
- 1:59:22
Growing Interest in Entering the Gaming Industry
The speaker observes that several people around him have recently been discussing entering the gaming industry, without assessing its viability or outlining practical entry steps.
- 2:00:39
Preparing LinkedIn and Social Operations with an AI Agent
The segment shows an agent being tasked with reviewing the social-platform workflow, checking Burnish’s LinkedIn company page for potential improvements, and locating and preparing the article-publishing area before posting begins. A similar approach is proposed for X, Instagram, and Facebook, without exposing private account details.
- 2:01:48
A Limited Impression of NVIDIA DGX Spark
Asked about NVIDIA DGX Spark, the speaker says he is not deeply knowledgeable about hardware and that his positive impression is based only on what he has heard, not on personal testing.
- 2:05:49
A View on Crowding in Gaming and SaaS Markets
In an opinion-based discussion, both gaming and SaaS are described as crowded markets. The speaker suggests that success in gaming may depend on genuine interest and the strength of the game idea, while noting similarities to SaaS product-building steps without claiming that games are literally software as a service.
- 2:11:59
Can AI Agents Be Presented as Employees?
An exploratory idea: treat long-running AI agents as real employees and clearly disclose their AI nature when presenting them professionally on LinkedIn. The speaker remains unsure whether the idea is sound.
- 2:13:37
Reviewing a LinkedIn Presence, Newsletter, and Articles
An example of reviewing a professional LinkedIn presence: spotting banner text obscured by the logo, organizing a newsletter, and using the articles section for full-length writing on a chosen topic.
- 2:17:03
Audio Direction: Mysterious Clock Ticks into a Beat Drop
Practical audio direction for a sequence beginning with pure, mysterious clock ticks over a very subtle muted beat, followed immediately by a beat drop when the ticking ends.
- 2:18:34
Token Optimization Does Not Always Mean Using Less
The discussion reframes token optimization: in some situations, the goal is to get more useful work done within usage limits, not merely minimize consumption. Codex is mentioned as the tool used more recently, without a systematic comparison.
- 2:21:44
Checking In with an AI Agent While It Works
A demonstration of messaging an agent while it processes a task, asking for progress and when results will reach the review area. The speaker also criticizes the agent’s replies for opening with a rigid self-introduction despite conversational prompts.
- 2:25:02
Cautious Expectations for the Next Gemini Model
A cautious response to whether Google might release a model matching another model whose name is unclear in the captions. The speaker says the outcome is unknown, awaits the next Gemini release, and personally doubts it will reach that level.
- 2:25:19
Discord or Telegram for an Arabic-Speaking Audience?
The speaker argues that Telegram creates less adoption friction for a broad Arabic-speaking audience, while Discord may require a new download and is often associated with gaming communities. This is the speaker’s assessment, not a demonstrated test result.
- 2:26:01
An Opinion on Google AI’s Strengths Beyond Coding
The speaker gives a negative opinion of Google’s position in the AI coding race, while acknowledging that it has strong models in other areas and explicitly ending on uncertainty.
Got a question about something in this stream?
Ask in the group