2. Remember 3 mile island and this is a great example to weave in.
3. Love this article, thank you. The overall premise is so timely and relevant and the takeaways + reminder that being human has advantages. I've been saying to people worried about AI taking over music, movies and art - "there are thousands of directors and one Tarantino, AI isn't going to create the magic that Tarantino does, ever. Same with Gaga."
Zack, your Three Mile Island framing nails something most AI commentary misses: the coming failures won't be technical—they'll be trust collapses. Your health insurer scenario (system "working perfectly" while denying cancer patients) is the exact pattern I'm seeing with enterprise clients. The optimization worked. The alignment didn't.
What I'd add: organizations that survive your predicted reckoning will be the ones who built what I call "trust infrastructure" before they deployed the tech. The 70-95% AI implementation failure rate isn't a technology problem—it's a people problem dressed in technical clothing.
Your "Human Renaissance" section is the piece I hope gets the most attention. We're not going to outcompute the machines. But courage, compassion, and wisdom aren't features you can fine-tune.
Looking forward to continuing this at the Granada later this month. Say hi to your dad for me.
1. Like the Escherish image.
2. Remember 3 mile island and this is a great example to weave in.
3. Love this article, thank you. The overall premise is so timely and relevant and the takeaways + reminder that being human has advantages. I've been saying to people worried about AI taking over music, movies and art - "there are thousands of directors and one Tarantino, AI isn't going to create the magic that Tarantino does, ever. Same with Gaga."
Brilliant Zack 💡👍
This is so insightful. Looking forward to your book that's coming out soon -- 1/17?
Thanks! Your favorite online store will ship it to you today and your local bookstore will have it on 1/13.
Awesome! Just ordered :)
Thank you!
Zack, your Three Mile Island framing nails something most AI commentary misses: the coming failures won't be technical—they'll be trust collapses. Your health insurer scenario (system "working perfectly" while denying cancer patients) is the exact pattern I'm seeing with enterprise clients. The optimization worked. The alignment didn't.
What I'd add: organizations that survive your predicted reckoning will be the ones who built what I call "trust infrastructure" before they deployed the tech. The 70-95% AI implementation failure rate isn't a technology problem—it's a people problem dressed in technical clothing.
Your "Human Renaissance" section is the piece I hope gets the most attention. We're not going to outcompute the machines. But courage, compassion, and wisdom aren't features you can fine-tune.
Looking forward to continuing this at the Granada later this month. Say hi to your dad for me.