Connect your Claude, Codex or OpenClaw for full transcripts on millions of podcasts. Two week free trial, cancel anytime.

Start free trial →
DM Dan Moore Dan Moore On The Stack Overflow Podcast

“We have a couple of principles around AI agents. One is the agent needs its own identity, because if it's riding along with your identity, it's really hard to audit and really hard to differentiate. The people who let agents drive their browsers terrify me. And another is that the bounds around the agent need to be deterministic software. It's not enough to say, hey, Claude, don't break anything or be good. Because they will try to do whatever they can to accomplish their goal. That's what we've seen.”

The Stack Overflow Podcast · Hands-On Engineering Podcasts · November 2026

“We have a couple of principles around AI agents. One is the agent needs its own identity, because if it's riding along with your identity, it's really hard to audit and really hard to differentiate. The people who let agents drive their browsers terrify me. And another is that the bounds around the agent need to be deterministic software. It's not enough to say, hey, Claude, don't break anything or be good. Because they will try to do whatever they can to accomplish their goal. That's what we've seen.” — Dan Moore, The Stack Overflow Podcast

Context

Asked whether agents should log in directly and whether their access can be limited, Dan Moore of FusionAuth gives his two rules for agent security.

Listen to the episode → More from Dan Moore → More Hands-On Engineering Podcasts quotes →

Transcript

The Stack Overflow Podcast Around 04:56 into the episode
Dan Moore

Anything it's allowed to, yes. Yeah, yes.

Ryan Donovan

Yeah, I mean, my next question was: should you let it log in directly? But it sounds like there are ways to really bound what the agent is allowed to do with that account.

Dan Moore

Yeah. You know, we kind of have a couple of different principles around AI agents at FusionOff. And one is the agent needs its own identity. And because if it's riding along with your identity, like, then it's really hard to audit. It's really hard to differentiate. The people who let agents drive their browsers terrify me. And another is that the bounds around the agent need to be deterministic software. It's not like enough to say, hey, Claude, don't break anything or be good. Because they will try to do whatever they can to accomplish their goal. That's what we've seen. And so you need to have bounds around that.

Ryan Donovan

I remember when OpenCloud was first released, people were buying new computers for it. I'm sure for that very reason because it's just this YOLO software. It's just like running wild. Totally.

Dan Moore

I think that's fine for experimentation. There's actually one of the people that I follow in the identity space has an agent that's running OpenCloud, I believe, called Claudery Hepburn. And she's given it all its own accounts. And if it gets a credit card, it has limited spend. And that's the kind…

Ryan Donovan

They're limiting. Yeah. I've seen some solutions where it's very, very tightly controlled. The usage like time boxed or single use or something. You think about things like that?

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

More from The Stack Overflow Podcast

“Why do we accept a 400-line or 4,000-line PR from AI when we wouldn't from our coworkers? Why wouldn't we go back to the AI and say, you need to break this down into ten smaller things that I can actually review and understand? Part of us learning to adjust to this new tool is building cultural norms that say, just because the AI wrote it doesn't mean you have to accept it as is.” — Dan Moore, The Stack Overflow Podcast “Intelligence is overweighted on. It has been dramatically overweighted on for the past few decades at least, if not longer. There are so many things about being a person are not about how smart you are. And the more obsessed we get with our intelligence and with being the best, the more we really lose out on the journey and the joy of the … there's so much more to life than smarts.” — Charity Majors, The Stack Overflow Podcast