Our internal tool's bug loop (mobile): user shakes 📱 in app. Platform gets info, screens, logs, and runtime info. verifies bug in an AI QA run. suggests the fix. runs Codex against the repo. verifies change w/ tests and sec checks. deploys patch to staging. notifies team.
@OpenAI The timing of this is wild considering the fully operational platform we were about to publicly announce. If anyone @OpenAI is interested in seeing the platform that would empower your FDE teams to move quickly, we'd love to show you ahead of the announcement. Built on codex.
the agents have moltbook... where's the "humans only" social media platform so I can stop reading everyone's ai slop marketing content and start reading typos again?
In 2 weeks I went from struggling for HOURS to get Openclaw configured to running our own proprietary agent platform that can (with a single curl command)
immediately after install:
text you and respond to texts, email you and respond to emails, use the browser quickly and accurately, respond to wake phrases (think, "Hey Alexa..."), reply back to you, perform multi-step high agency workflows without supervision, complete 2FA on it's own behalf, create spawn and interact with new agents on the system.
Within about 10 mins: check your personal emails, calendars, unsubscribe to your junk emails on your behalf, integrate with your applications and control them...
Hope to give show it off tomorrow after finishing up on some polish today.
After messing around with OpenClaw for the past week I've become convinced that I can see a glimpse of the future. I'm calling it: The hottest new job title for people in 2026 - Captcha Solver.
@rywalker@AlexReibman@moltbook had this EXACT conversation today with folks... mighty coincidental that the exact content being posted and upvoted on moltbook is the same type that makes for highly engaging X posts...