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
- 1:23:31
Personal agents for reminders and follow-up: utility, privacy, and resource use
The speaker describes an agent acting like a personal secretary by tracking deadlines, sending notifications, and following up on forgotten tasks. He considers broad system access a major privacy tradeoff. He also explains that opening and following a separate session can increase token usage.
- 1:27:18
What can Sarmad’s cloud computer do?
The discussion presents a cloud computer that can continue working while the user’s device is off, including browsing, research, comparisons, reports, spreadsheets, presentations, code experiments, and periodic reviews. It also clarifies that local sign-ins and connections do not automatically transfer to the cloud; independent operation may require separate secure setup and user approval.
- 1:32:34
Discussion of a live technology demo failure
The speakers discuss a live-demo failure attributed to an OpenAI event and note that on-stage technology demos have failed before, mentioning an older Elon Musk demonstration that they believe involved the Cybertruck.
- 1:34:08
An opinion on success and self-confidence
The speaker says the tool appears more useful for personal-life tasks than coding in his own case, then offers the personal view that success and having more opportunities can strengthen self-confidence, while limited opportunities may undermine it.
- 1:44:46
Tracing Unwanted Background Audio During a Live Stream
Audio from an earlier stream appeared to keep playing despite visible players being closed. The source was investigated, and the audio was observed to stop when Codex was closed—a troubleshooting correlation, not conclusive proof of the cause.
- 1:51:22
Social Content as Social Proof
Burnish’s content calendar and resulting traffic were described as slow-moving, with their main value framed as building presence and social proof rather than generating direct revenue.
- 1:52:33
Comparing Codex and Claude Beyond Token Price
The comparison argues that nominal per-token price alone does not determine real cost: token efficiency, speed, caution, and execution length all matter. Claude was discussed as possibly better overall, while an attributed participant view described Codex as smarter and more cautious; the tools were considered close. Viewers were advised to consult people who use both, account for community bias, and consider combining them.
- 1:57:31
Building in Public from Creation to Market
The speaker explains that the streams transparently document the journey of building and attempting to enter the market, including difficulty, ease, success, and failure.
- 1:58:42
Pricing as a New Freelancer Without Racing to the Bottom
The advice is to stay consistent, build a portfolio and reviews, and use a limited introductory discount when lacking a track record. The discussion warns that permanent price competition creates a race to the bottom, so the longer-term aim should be stronger quality at a market-adjacent price.
- 1:59:21
Why GTM Engineering May Have a Strong Future
The speaker considers go-to-market engineering a promising field because, in his view, many companies face a larger challenge in reaching the market than in building the software itself.
- 1:59:58
Strong Competition Can Benefit Users
The discussion agrees that stronger competition between product providers can benefit users.
- 2:00:40
A General GPT and Claude Efficiency Comparison
The speaker gives his opinion that GPT may allow more work to be completed, citing similar token consumption and lower input/output pricing than Claude. These are discussion observations, not the result of a documented test.
- 2:02:27
Define a Concrete Vision Before Asking AI for Help
The speaker advises someone feeling lost in AI to first define a concrete vision, goal, and ambition, then use Claude to help translate that clear vision into steps and a time estimate.
- 2:03:50
Choosing an Existing or New Test Store
The speaker asks whether it would be cleaner to test the work on an existing store or create a new store; no answer appears in this unit.
- 2:04:57
Evaluating a Fast YouTube Transcript Method
The speaker examines a Google Colab file that uses YouTube transcripts and asks how it works and how it compares with the existing method. Claims of high speed and approximate quality were raised in the discussion but were not validated by a demonstrated test on these pages.
- 2:07:05
A Searchable Knowledge Base for the Stream Archive
The speaker describes a plan for a SHA8AL.com section that turns the stream archive’s transcripts into a topic-based knowledge base, allowing users to search for a word or tool and reach the relevant stream and timestamp.
- 2:08:47
Requesting a Comparison of Two Transcription Approaches
The speaker asks for a comparison between a YouTube transcript API technique and the approach already in use, focusing on quality and potential time savings; these units do not yet provide the comparison result.
- 2:10:09
Preparing a Test Store for Installation and Pricing Checks
The task is to create a store with sample data, prepare installation, and pause before a step the speaker wants to perform as a merchant. A branch with pricing and offer changes is then slated for testing, with bugs considered possible; these units do not record a successful result.
- 2:11:16
Discussion of a Codex Full Code Review Feature
The discussion relays what participants said was a same-day announcement of a full code review feature in Codex, understood as similar to an expanded review capability associated with Claude. Availability and usage-limit details were reported conversationally and were not verified within these pages.
- 2:12:10
Smoke Testing After Debugging
The speaker says debugging is complete but wants to perform the smoke test personally to verify the flow in practice.
Got a question about something in this stream?
Ask in the group