If your agents ship overnight, what happens if something goes wrong?

Viktor Farcic's last task before bed is kicking off the next round of work for his AI agents. By morning they've built, shipped, and recorded their own tests as video clips, and he signs off by watching the reel, without reading a line of code. If that sounds reckless, he's been doing it in public for months on DevOps & AI Toolkit, and mostly, it works.

This session picks up where those videos leave off: the morning after, when some of the changes occasionally go wrong.

Join us to hear:

• Viktor’s nightly setup: what he queues before bed, how much freedom the agents get, and what he checks in the morning before approving deployment
• Scenarios where it went wrong: how long the bad change ran before anyone noticed and what it took to recover
• Where he lands on the prevent-or-prepare debate: which guardrails earn their keep, and what to have rehearsed before the first overnight run