Pinned
🚀
4d •
11/10 - 11/12 ✄ Live Show Cut Up by Topic (the Daily Pulse 🔥)
Hey Architects! Lives this week start today! I'll be covering topics from my new KEDB product launch. - PAID members can access all live segments anytime (✄ cut and organized by topic) - FREE members can access the last unedited video for 24 hours only (see below👇) 📆 Upcoming: Tues-Thur (calendar) 🖥️ Watch now 11/10 - Content Update & Roadmap - AI Subscription Strategy 11/12 - Home Networking Update - Building A Lead Magnet 💬 Video request? - Make a video request and I'll cover it (click here) 💪
4
0
Pinned
6d •
GTA6-style Diwali film with AI.
Hey everyone, happy Diwali! 🪔 I just finished a GTA-style short film where the twist is that the mystery bag is full of crackers. It was a big prompt-engineering lesson, so I’m sharing the whole process. What went wrong at first (30-second single generations) The zip opening was a mess The main character stood on a car The last 10 seconds broke completely What fixed it 12 scenes of 10 seconds instead of one long clip Each scene follows a fixed structure: Scene Context, Characters, Location, Format, Camera, Action, Lighting, Texture, Audio, Positive Locks A Physics section in every prompt (weight, braking, door hinges, recoil) Positive locks: “he stands on the roof floor” instead of “don’t stand on the car” A flyer lock so the flyers stay on the door when it opens Reference sheets first, then tag them in every scene Tools: Seedance 2.5 on Higgsfield. Questions for you: What’s your biggest consistency problem in AI video right now? Do you prefer 10-second scenes or longer shots?
5d •
Automate technical web app quality control
Hi all, I'm new here and I would love to hear from experienced web app developers who may have developed a technical web application, for example, to perform or automate data analysis or create or design things based on certain known mathematical or physical constraints. My goal for a technical web app I am developing is to somehow use an agent (or subagents) to perform checks of the underlying behavior or calculations the vibe-coded (Claude) code is performing. As the functionality of my web app increases, I am consistently challenged with the ability to perform quality checks on the new outputs.
1d •
I map the process on paper before I open the tool
I used to open the tool first. Connect a few nodes, figure out the flow as I go. Now I do the opposite. 15 minutes with pen and paper. Every step, every branch, every spot where a human makes a call. Sounds slow. It isn't. Last flow I built this way shipped in one afternoon. The one before that, built tool-first, took two extra days because I hit an edge case I should have caught on paper. Paper is the cheapest debugger you'll ever own.
0
0
2d •
One n8n flow took a weekly report from 4 hours to 12 minutes
I build automations for small businesses. One client spent 4 hours every Monday pulling numbers from 3 dashboards into one report. We moved it to one n8n flow. It runs every Monday at 7am, pulls the same 3 sources, writes the report to a sheet, and emails it. First version took me a weekend. The trick that made it stable: make it fail loud. If a source is down, the flow sends an email that says 'this step broke' instead of sending a half-empty report. Silence was the old problem, not speed. Now it's 12 minutes of checking instead of 4 hours of copy-paste. One weekend to build, about an hour a month to maintain.
0
0
1-30 of 6,747
skool.com/ai-architects
Learn to build real apps and AI agents in 30 days or less. Get your first customer in 90 days. No experience required.
Leaderboard (30-day)
1
+9
2
+8
3
+7
4
+6
5
+5
Powered by