Building Intelligence + Manifesto
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 hands-on review of Burnish before V1: auditing the first-time-user experience, documenting bugs in product-readiness states, publishing, links, and image alt text, and testing autopilot batches, timing, loading, and pause behavior. It also traces failures in manifesto generation and PDF downloading and a missing FAQ presentation, using a workflow of visual verification, issue logging, root-cause research, separate implementation handoffs, and incremental merging and deployment. The work remained unfinished: autopilot, the Intelligence page, and downstream manifesto integration still required testing and polish, while several bug causes were unresolved. The discussion also develops an exploratory Telegram-based virtual office for collaboration and project progress, including prototypes, character movement, and collisions, although its lasting purpose, implementation scope, and privacy model remained open questions. Audience exchanges cover launching mobile apps, choosing B2B or B2C and web or mobile, competitor research, unit economics, and marketing, alongside qualified opinions about agentic coding tools, AI models, infrastructure, and token pricing. The speaker’s recurring view is that workflow fit matters more than broad model claims and that building a product is only the start of market testing, marketing, and sustained execution.
82 moments from the stream
- 1:04:00
A Qualified Discussion of AI Data Centers and Token Pricing
An audience question led to a discussion of remarks attributed to Jensen Huang at GTC: a shift from data centers that retrieve information toward infrastructure that produces tokens, and what that might mean for business models and AI usage costs. The speaker made clear that he had not heard the keynote and did not know its specifics, expressing only a hope that the trend could lower costs for users.
- 1:10:24
A Community for Sharing AI Ideas and Knowledge
The Telegram community was described as a space where people build projects and exchange ideas and information about AI.
- 1:11:10
Aligning Product Status Names and Meanings
A review of possible confusion among queued, healthy, and partially published states, with a proposed naming change so each label matches its actual meaning in the product tab.
- 1:14:36
Fixing the Product Page FAQ Display
A discussion of a recurring problem where the FAQ section looked broken on the product page, and of an HTML-configurable block created as a better presentation option. A previously polished version had disappeared, but the cause remained uncertain, including whether it had not been committed.
- 1:17:05
Adding a Bot to a Group as an Administrator
A practical attempt to add a bot to a group and grant it administrator access. Later captions confirm that the bot was in the group as an admin, but the supplied captions do not contain enough steps for a complete procedural guide.
- 1:20:46
The Product Journey Starts After the Build
Responding to a question about selling a service, the speaker argues that building with AI is not the end of the journey. Entering the market, marketing, and persisting after encountering real-world challenges are essential parts of the process; the aim is to show the full journey, including how the product is built and how decisions are made.
- 1:23:04
Validate Margins and Unit Economics Before Building
The speaker stresses evaluating profitability, monetization, market demand, margins, and unit economics before spending months building an idea that may not work commercially. Marketing is presented as a core part of the journey, while the expected launch timing remains uncertain because features keep being added.
- 1:26:46
Building in Public and Growing an Arabic Vibe-Coding Community
The speaker describes the channel as a journey beyond vibe coding: building applications, entering the market, and marketing products while showing the thinking process transparently. The stated aim is to foster an Arabic community around vibe coding.
- 1:27:52
Checking Shopify Product Visibility and the QA Workflow
The work includes asking whether a single Shopify app deployment could resolve the issue, then changing a product to draft to check whether it disappears and observing its active state. The discussion then moves to continuing a QA walkthrough and reviewing a recent code change before merging and deploying it; the fragmented captions do not establish the issue’s final technical cause.
- 1:40:39
Testing Item Visibility After Switching to Draft
He switched an item to draft expecting it to disappear from the interface tabs, refreshed, and confirmed that it was hidden. He also noted that restoring it would make it reappear before deleting it entirely.
- 1:44:41
Reviewing Autopilot Before V1
The speaker said V1 could be ready if the autopilot worked properly, but first wanted to understand where a UI panel came from and what it was meant to communicate. Later, he said the autopilot page still contained bugs while the rest had largely been polished, with the hope of wrapping up the frontend soon.
- 1:51:39
Recalling the 2009 Egypt–Algeria Football Dispute
The football discussion turned to the major match-related dispute between Egypt and Algeria, with the year corrected to 2009 and a reflection on how quickly time has passed.
- 1:54:44
A Humorous Recollection About Bitcoin and CAPTCHAs
The speaker jokingly and tentatively suggested that Bitcoin may once have been offered for solving CAPTCHAs, reflecting on how trivial it seemed then and how holding it might later have created substantial wealth.
- 1:57:24
Blind Spots in the Vibe-Coding Era
The speaker argued that the vibe-coding era still contains overlooked opportunities and ideas, even when people believe they are already working on everything important.
- 1:59:51
Cautious Speculation About Upcoming AI Models
Asked about a possible GPT release, the speaker stressed that he was not following the subject closely, then thought aloud about whether a quick release or a longer wait for a stronger model might follow from competition with Anthropic.
- 2:01:33
Autopilot Testing Reveals Batch and Timing Issues
Testing showed that the next batch did not push products into autopilot as expected, which the speaker classified as a critical bug. He also requested a 60-second initial countdown and verification of local-time and next-batch calculations.
- 2:04:07
Show the Next Batch Loading Immediately
A review of an autopilot UI bug: the batch ran in the background, but products did not appear in recent optimizations until a refresh or pause action. The requested behavior is for the interface to enter a visible loading state immediately when the countdown ends while products are fetched and prepared.
- 2:05:24
Comparing Models in Agentic Workflows
A personal-experience discussion of model versions referred to as 4.7, 4.8, and 5.5. The speaker found one model more comfortable for agentic workflows, while saying another could produce strong work when the user knows the exact goal and supplies a detailed prompt; this was not presented as a benchmark of overall capability or intelligence.
- 2:12:40
Simplify Autopilot Pause While Keeping the Timer Running
A proposal to replace multiple pause-duration choices with one pause button. The autopilot card should show a paused state, while the next-batch countdown continues in the background so resuming reveals the true remaining time instead of automatically resetting it.
- 2:17:09
QA for Level Unlocks and Telegram Links
Discusses how to QA progression from level one to two and then three when unlocking depends on reaching an approval threshold. A proposed approach is to move progress close to the threshold and complete the final steps manually, while questioning whether that would distort the natural user journey. It also examines the expected behavior when opening a link into a Telegram chat.
Got a question about something in this stream?
Ask in the group