New SHA8AL.com + Sarmad
Stream summary 75This stream in minutesWant 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 centered on turning SHA8AL.com into a community hub and searchable bilingual knowledge base that links topics and tools to timestamps across the broadcast archive while showcasing members’ projects. Luqman discussed an automated workflow for downloading YouTube captions, filling gaps locally, and then using models to summarize, select passages, and produce short and longer videos. Arabic transcription accuracy, selection criteria, video quality, and archive-processing cost still required evaluation, and no completed system was demonstrated. Practical work included reviewing Burnish and smoke-testing its installation and free-plan flow, with proposals to reduce onboarding friction, clarify what counts as an install, improve the analytics dashboard, and track Shopify App Store visits. A recurring conclusion was that distribution and traffic—not product construction or the funnel alone—were the main bottleneck, prompting exploration of measurable channels such as outbound outreach, agency partnerships, advertising, and communities, without claiming final results. The stream also covered Sarmad as an automation graph system, Paperclip and goal-directed agent coordination, build-versus-buy tradeoffs, and privacy and usage-limit constraints. Luqman offered qualified, non-benchmark impressions of Codex, Claude, GPT, transcription, and video tools, emphasizing hands-on testing, adequate context, and validation when model outputs conflict. Audience discussions additionally touched on building a Micro-SaaS with AI, market choice, freelance pricing, AI’s impact on jobs, and automated social replies, while several proposed workflows, comparisons, and tests remained unresolved.
177 moments from the stream
- 5:29:41
Distributing an AI Feature Through Shopify Apps
A proposal suggests exchanging a free AI-answer feature for measurable exposure inside other Shopify apps, with an agreement that places it before merchants as the first concrete deliverable. It is presented as a research output that still needs review and evaluation.
- 5:31:22
Using Research and Handoff Prompts for Acquisition Ideas
The workflow asks AI tools to review work since launch, conduct broad research, and prepare strong handoff prompts for other tools to uncover new ideas for addressing audience-reach difficulties. Some tool names are unclear in the captions.
- 5:33:39
Explaining Inbound, Outbound, and Targeted Distribution
Traffic is compared to people entering a shop: directly calling people in represents outbound traffic, while attracting them through what the shop offers represents inbound traffic. The discussion then argues that outcomes may depend on exposure, distribution partnerships, and targeting the right parties rather than click volume alone.
- 5:36:17
An Agent Writing About Trending Hashtags
An agent is proposed to target trending hashtags and write about them, with the qualification that this approach may be very slow.
- 5:36:26
Can AI Be Used to Build a SaaS?
A brief answer says it is possible to build a SaaS product with AI without necessarily hiring a programmer, but the page ends before any conditions or practical details are given.
- 5:36:34
Start With a Micro-SaaS When the Scope Is Uncertain
The speaker says a programmer may deliver a cleaner and faster implementation, but AI can be suitable for a modest SaaS product; when the idea is uncertain, starting with a small Micro-SaaS keeps the initial bet limited.
- 5:38:19
A Neutral Prompt for Judging Two Models’ Research
The speaker drafts a handoff prompt for a fresh session that receives research from two models, compares it without bias, identifies overlaps, and recommends the stronger direction or a useful combination and rethink. He notes that he has not reviewed all the material yet.
- 5:42:39
Verifying a Comparison Between AI Results and Ad Spend
One model output makes a general claim that spending on AI-generated results could outperform advertising, but the speaker pauses before endorsing it and says its basis must be checked.
- 5:43:43
Choosing AI Tools to Build an App With Payments
Answering a question about building an app and configuring payments with AI assistance, the speaker recommends Claude or Codex when possible. He says Replit is not a bad option, but it is not his first choice.
- 5:51:27
Evaluating an AI Assistant Beyond Coding
He clarified that he had not tested the assistant for software development, but found it useful for reminders and noticing selected tasks. His assessment therefore concerned assistance and organization, not coding quality.
- 5:52:44
Using AI to Modernize Older Games
A participant said he was impressed while using AI to update a game he had made around 2017. The host separately relayed a friend’s claim that AI turned a prototype from 2016 into a game within two days, and another participant discussed updating games made from 2017 to 2023. These are participant reports and a second-hand anecdote, not documented benchmarks or independently verified results.
- 5:54:19
Cross-Judging AI Research Responses
He described comparing two research responses through independent sessions, asking each system to identify the stronger answer and the overlap between them. The method is a practical way to seek a clearer direction, not proof that the resulting judgment is correct.
- 5:59:34
Comparing Claude and Codex to Form Your Own View
He recommended trying Claude and Codex when possible so users can form their own usage-based opinion, while favoring official access and avoiding unofficial channels. The passage did not provide a benchmark or a definitive ranking of the tools.
- 6:00:01
Self-Teaching Programming with YouTube and Free Courses
A participant explained that he learned programming gradually through YouTube and later used free course coupons. The discussion presents a practical, incremental self-learning path.
- 6:03:50
An Opinion on Codex’s Value and Efficiency
The speaker recommends Codex and says his choice is based on what he sees as genuine value and high efficiency, while also acknowledging that he likes Claude. This is a personal assessment, not the result of a documented comparative test.
- 6:04:33
Choosing a Main Goal While Keeping a Validation-Dependent Alternative
The speaker selects option C as the main goal for now, while suggesting that quickly building and validating option A could make option B viable. The available captions do not clarify what the options represent.
- 6:05:25
Be Cautious with Time Estimates Before Testing the Roadmap
The speaker questions time-and-effort projections made before working through the roadmap, arguing that long or precise estimates may be unreliable without trying the tasks. He also offers the experience-based opinion that Claude can struggle with timing estimates.
- 6:07:49
Two Conditions for Joining a New Venture
The speaker explains that he now prefers ventures with either early funding and clear momentum or a reasonably assured distribution and sales path, reflecting the difficulty of building without a route to market.
- 6:13:52
Building in Public and Investor Concerns
Responding to a suggestion to avoid building the project live, the speaker says he expects not to work on it during the stream, while arguing that openness is not necessarily an obstacle to investment, citing open source and emphasizing the need to consult his partner.
- 6:14:47
A Practical Comparison of Codex and Claude
At the time of the recording, the speaker argued that Codex could handle most general work Claude could do and might be more efficient when used well, while Claude might distribute tokens more efficiently. He considered them closely matched at that time and recommended testing them directly.
Got a question about something in this stream?
Ask in the group