David Nolan

That's impossible.

Impossible.

Impossible.

So...

So I want to know, these are the kinds of things I want.

People say you can do it, but you don't.

So as a CISO, is it hard to defend focusing on basics without looking like you're settling for mediocrity?

David Nolan.

Development has been the use case that's paying dividends for LLMs.

But Keith Townsend, the CTO advisor, is skeptical if we're about to replace developers saying, quote, AI does not own outcomes.

It does not bear responsibility when an assumption turns out to be wrong.

It does not understand the difference between confidence and correctness, only how to simulate both convincingly.

AI can argue its case fluently, citing plausible metrics and familiar frameworks, but the moment you ask it basic questions, the confidence outpaced the evidence, not because it was lying, but because it was presenting claims with no accountable owner.

For him, until AI closes that gap, it's an accelerator, not a replacement.

So, Andy...

How should security leaders manage AI generated code?

How do you build a governance model for code that arise with confidence, but no owner?

Well, great advice, both of you.

Excellent job during the what's worse.

And kudos to all these sort of unwitting contributions for our show.