I'm building a SOFTWARE FACTORY. The Factory Goal: - Be able to code an app on autopilot once the direction is set - Add in new features and fix bugs by user feedback directly from WilsonOS - Track and steer progress over time PROBLEM: small request, like changing the label on a card took 16 minutes of waiting time SUCCESS: I got that time down to under 6 minutes now. Here's a link on how: https://claude.ai/artifact/HaC6TfrVYeyot2ACX4UDWM TLDR The biggest needle movers: - Run each check once, in the right place. The same test suite was running three times in a row. Now the builder checks only its own files (4 min → 17 s), GitHub checks the whole app alongside QA, and QA uses the browser. - Get heavy work off the shared box. Every failed run was memory, not CPU. CI moved to GitHub, and leftover dev servers are cleaned up after each job. - Keep QA's app warm and fast. A dev server that starts while the builder works, on a faster bundler: QA went from 7:20 to 1:45. - Kill the waiting. By the end, hand-offs were the slow part. An always-on loop every 8 s, merging in the same round QA passes, and a fix for one step that re-read every job ever run (30 s → 4 s). - Tell QA where to look. Every plan names the page and what to check, so QA goes straight there. The lesson: measure every step. Half our early guesses about where the time went were wrong. Let me know if you're interested in seeing the software factory in action or if you want a tutorial on how to build your own and I'll add that to the community.

Heck yeah!!!! Thanks for the invite! I look forward to learning with this group

You may have heard this horror story... So there Mia (my agent) was, doing the work for me. Social content on autopilot. Sales and leads never better. More free time to do what I love. Then one day, something strange happened. My money dissappeared. My socials posted embarrassing garbage. That's when I realized, Mia just gave away the keys to my entire system... Ok, this is a scary extreme story. But if it is something that keeps you up at night One open source solution I'm exploring is 'Infisical'. Infisical is the open source security infrastructure platform that teams use for: - secrets - certificates - privileged access management. https://github.com/infisical/infisical What security methods are you exploring?

Claude auto mode allows our agents to work more autonomously which many of us want. But there are risks to consider... Here's Claude's architecture to ensure the safest auto mode experience: https://www.anthropic.com/engineering/claude-code-auto-mode So, do you think auto mode is worth it?