Monte Carlo reposted this
AI teams gotta be careful tonight. And every other night. PS if you're planning a night in but you still wanna get a little scared, check out this roundup of 4 spectacularly public AI failures—and how they could (and should) have been avoided. (in the comments) Because the only thing scarier than pushing bad AI to production... is not having the end-to-end observability to catch it.