@georgedagg_ @georgedagg_ ·
I had a holy shit, the world is not the same anymore moment last night.
Put BigC, my @clawdbot on a hetzner server and my mind is just blown.
We all know that AI is transformative but I don't think you really feel it until you step outside the prompt/response loop and step into the prompt/take real actions on my behalf/response loop.
- I got BigC to check on the deployment of a new railway project
- Builds for each of the services in the project failed
- BigC reviewed the logs
- Then reviewed the project on the server
- Identified the root cause (incorrect build commands on each service)
- Updated the configs of each in railway
- Redeployed
- Confirmed everything worked
- I then tested it
- Identified a design issue
- Got BigC to review, fix and submit a PR
- job done
oh and all this was done over voice when walking the fucking dog...
More to come as I try and make sense of this new world.
What you need
- Connected apps
- Railway.
What prompted it
Builds for every service in a new Railway project had failed.
What OpenClaw did
- Asked BigC, the author’s OpenClaw on a Hetzner server, to check on a new Railway project’s deployment.
- BigC found the builds had failed, reviewed the logs and then the project on the server.
- It identified the root cause: incorrect build commands on each service.
- It updated each service’s config in Railway, redeployed and confirmed everything worked.
- The author tested it, found a design issue, and had BigC review, fix and submit a PR.
Evidence and limits
- Original post checked on X (Oct 19, 2026)
- Outcome reported by the author
- Not reproduced by anyone else yet
