Meen el hodhod dah?
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
A practical stream centered on an unfinished “Mission Control” tool for visually supervising coding-agent swarms, exposing their status and consumption, and detecting stuck loops, alongside live reviews of the interface, bird characters, and animation. The discussion broadened to goal-based roadmaps, separating software-building tools from company-operations systems, and balancing custom agent harnesses against added complexity and waste. Luqman answered audience questions about Codex, Claude, AI tools, learning, fundamentals, and hosting, repeatedly framing many comparisons as personal experience rather than conclusive tests. He also argued that distribution and demand validation can be harder than technical construction, and discussed a possible open-source GitHub release with community contributions or a competition whose prize and judging details remained undecided. The interface, motion, roles, naming, and release plan were still incomplete, and neither real token savings nor practical success had yet been established.
141 moments from the stream
- 5:10:13
Choosing a Useful and Creative Contest Prize
The speaker discusses the trade-off in choosing a contest prize: cash is flexible and useful to any winner, but may feel less creative, so he asks for suitable alternatives.
- 5:14:08
An Open Contest with AI-Based Code Review
The speaker says the proposed contest would be open both to developers and to others building with AI-assisted coding. The planned judging would use AI to inspect pull requests, code quality, and additional criteria.
- 5:15:21
A Bird-Themed Multi-Agent Concept
A creative multi-agent concept in which each bird has a distinct persona and specialized responsibility, such as review, compliance, security, or orchestration, with the hoopoe acting as orchestrator. It was framed as related to Mission Control rather than an entry into the broader automation-bot category.
- 5:17:24
Choosing a Prize That Appeals to the AI Community
A discussion of cash, subscriptions, and gift cards as possible prizes for a broadly targeted competition, with interest in an AI-related reward and promotion through TikTok and Instagram to attract attention.
- 5:19:42
What Should Be Settled Before Building a Startup?
Practical advice to explore early funding while recognizing the difficulty of pre-seed investment, then establish a monetization strategy, a dependable distribution channel, and sound unit economics before building. It also stresses studying competitors: their absence can require market education or signal a problem, while a “purple ocean” between crowded and entirely new markets may offer a balance. A high gross-margin target was presented as the speaker’s rule of thumb, not a guaranteed standard.
- 5:22:21
Opening Mission Control and Building a Community Around It
The speaker acknowledges that visual polish can be distracting, but intends to use Mission Control immediately, release it publicly on GitHub, and organize a developer competition around it to strengthen both the tool and its community. A previous demonstration of its agent-swarm integration was said to have failed because of a UI issue attributed to a change made by Codex.
- 5:26:39
Balancing Speed and Usage Limits in AI Coding Tools
The speaker discusses the appeal of an ultra-fast mode for sharply reducing the time of some tasks, while noting that it can consume usage limits quickly. He has not decided yet and may reserve it for selected tasks after weighing the tradeoff.
- 5:29:33
Designing a Multi-Agent Orchestrator
The speaker sketches an early concept for a system with an orchestrator and multiple agents, emphasizing that its structure and roles need to be designed carefully.
- 5:31:13
Assigning Agent Roles Through Bird Personas
The speaker and audience discuss which bird persona best fits the leadership role in an agent system, considering the owl and hoopoe while keeping the design open to change.
- 5:32:04
Technical Challenges Versus Marketing and Traffic
Answering an audience question, the speaker said that, at the time of the recording, AI tools could handle many coding problems, while an experienced developer was still better at establishing sound project foundations. In his view, the larger practical obstacles had shifted toward marketing and traffic, though difficult bugs still occurred.
- 5:38:29
Choosing an Affordable Image Generation Model
Answering an audience question, he said identifying the absolute cheapest option requires research and comparison. He described GPT as offering good quality, mentioned Gemini as another option, and did not recommend Claude for this task.
- 5:40:28
A Visual Cockpit for Coding Agents
He demonstrated an experimental interface representing four coding agents as birds, with a central bird intended to become an owl, and showing whether they are reading GitHub pull requests, opening a pull request, coding, testing, or needing to ask a question.
- 5:41:53
Bug Severity and the Need for Independent Testing
He explained that bugs differ in severity: some affect user experience, while others can break the whole application. He questioned relying only on the same model that built the product to judge its work, arguing for external review and testing, especially before traffic exposes scaling problems.
- 5:43:43
AI-Assisted Coding Does Not Eliminate the Fundamentals
He discussed real applications built by people with limited software-engineering backgrounds using AI-assisted coding, while qualifying that correctly learning the fundamentals still matters even when AI can perform much of the implementation.
- 5:45:03
Detecting Stuck Agents and Preventing Wasted Work
He explained that the monitoring tool was motivated by an experience in which agents remained trapped in repeated loops instead of stopping and requesting intervention. Its purpose is to expose each agent’s activity so stalls can be detected earlier, while he acknowledged that the interface was still unfinished.
- 5:47:13
Scalability Versus Reaching Users
He argued that scaling problems arising after real demand can be a solvable kind of problem, whereas spending a long time building without being able to reach users is a harder risk. The comparison balances technical and distribution risks without dismissing the need to fix failures.
- 5:51:18
Rethinking Alerts for Agents Needing Intervention
He suggested revising the interface’s attention panel and reconceptualizing notifications for “needs you” and “stuck” states so cases requiring intervention are clearer.
- 6:00:39
Builder Platforms in a Crowded Market
An opinion praising the idea of a platform for the growing builder community, while warning that the market is competitive, execution is difficult, and investment support may be worth considering.
- 6:01:28
What Does It Cost to Build an App or SaaS with AI Coding Tools?
An answer to a question about building an app or SaaS with tools such as Cursor and Codex: there is no fixed cost because it depends on application size and other factors, though a low-cost monthly subscription and patience may be enough for a small project.
- 6:10:55
A Side Tool for Building Agent-Based Projects
The speaker explains that the “birds” are not the main project but a still-in-development set of supporting tools intended to make project building clearer, especially when using agents, with a control-center-like role envisioned for it.
Got a question about something in this stream?
Ask in the group