Chief Executive Officer
Chief of Staff
Managing Principal
Most AI headlines ask whether the robots are taking over. But the more useful question is: who’s steering? In this episode, we sit down with Joe Norris of Data Center Results to unpack the stories about “AI agents escaping containment” and what they actually tell us about the technology. Instead of treating unexpected AI behavior like a sci-fi problem, we look at the people, incentives, instructions, and systems behind it.
In this recording, we’ll review:
- Emergent behavior or a control problem: Why the difference between “AI went rogue” and “we gave it bad instructions” matters more than the headlines suggest
- Who benefits from the fear: How a viral insider story and a legitimate safety concern can look almost identical from the outside
- A mission is a mission to a computer: Why unexpected behavior often comes back to the instructions, objectives, and guardrails humans put in place
- What the nuclear comparison can teach us: How powerful technologies can become safer when the incentives to cooperate become stronger than the incentives to compete
- Building responsibly instead of trying to stop it: Why the teams that succeed with AI will be the ones asking better questions about control, accountability, and responsible development
And much more.
