00:00I went through like a coding weekend 10 years ago and locked myself in my office and I built like
00:05a
00:05little email triage system that automatically scraped and read everything that came into our
00:10company's inbox. I'm the CEO, which was the first problem. I scraped everything that came in the
00:15inbox and then triaged it and automatically sent emails out and it downloaded everything to a
00:20spreadsheet. I thought I was invincible until I like let it run in a live environment and I spammed
00:25a thousand people because I had created a loop and then my co-workers were like, who approved
00:29this? What's happening? And you know, I was trying to shut it off. So here's the issue. That was not
00:35AI
00:35waking up and having godlike powers. That was janky code that I wrote and then I put it in a
00:41environment that it wasn't prepared to be in. So we talk about agents going rogue. The first thing with
00:47every executive leadership team is we have to stop anthropomorphizing. AIs are not making a decision.
00:53They don't have consciousness, right? It's something more complicated, but it's the same kind of
00:58mistakes humans would have made.
00:59humans would have made.
