Sign up to save your podcasts
Or
Share Alien Intelligence (Audio & video edition + extras)
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Alien Intelligence (Audio & video edition + extras)
Hey, it’s Marek.
A COO told me about their GenAI pilots. Faster drafts, fewer errors, great metrics. I asked what surprising behaviours they were seeing. Surprises they could learn from. Blank look. “We’re not measuring that.”
Nobody asks: “What is AI doing that we’d never try?”
Move 37
In 2016, AlphaGo played a move no trained Go player would make. Game 2, Move 37. Commentators went silent. “It’s not a human move.” AlphaGo won.
The untold part: a few years later, researchers analysed 66 years of professional Go. Quality was flat until 2016. The world’s best players hit an imaginary ceiling. They’d optimised within invisible boundaries.
After AlphaGo won, almost immediately, human players started making better decisions. They realised that there were other approaches to the game, ones they could now learn from the machine.
The alien intelligence revealed what 3,000 years of expertise had ruled out.
What would AI do in your business that you’d never try? And why not?
P.S. Right after I published the newsletter version of this post, Prof. Toby Walsh, a friend from the University of New South Wales, pointed me to his book “Faking It”, in which he had a section called… Move 37. There, he described a chess game, Kasparov vs Deep Blue, 1997. Kasparov lost. The pivotal moment? Game 2, Move 37. Nineteen years before AlphaGo’s Game 2, Move 37. Coincidence?
Stay curious!
Listen to the full episode for the research that proves we’ve been asking AI the wrong question, and what Game 2, Move 37 reveals about the moves we’ve ruled out without realising it.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com
View all episodes
By Marek KowalkiewiczHey, it’s Marek.
A COO told me about their GenAI pilots. Faster drafts, fewer errors, great metrics. I asked what surprising behaviours they were seeing. Surprises they could learn from. Blank look. “We’re not measuring that.”
Nobody asks: “What is AI doing that we’d never try?”
Move 37
In 2016, AlphaGo played a move no trained Go player would make. Game 2, Move 37. Commentators went silent. “It’s not a human move.” AlphaGo won.
The untold part: a few years later, researchers analysed 66 years of professional Go. Quality was flat until 2016. The world’s best players hit an imaginary ceiling. They’d optimised within invisible boundaries.
After AlphaGo won, almost immediately, human players started making better decisions. They realised that there were other approaches to the game, ones they could now learn from the machine.
The alien intelligence revealed what 3,000 years of expertise had ruled out.
What would AI do in your business that you’d never try? And why not?
P.S. Right after I published the newsletter version of this post, Prof. Toby Walsh, a friend from the University of New South Wales, pointed me to his book “Faking It”, in which he had a section called… Move 37. There, he described a chess game, Kasparov vs Deep Blue, 1997. Kasparov lost. The pivotal moment? Game 2, Move 37. Nineteen years before AlphaGo’s Game 2, Move 37. Coincidence?
Stay curious!
Listen to the full episode for the research that proves we’ve been asking AI the wrong question, and what Game 2, Move 37 reveals about the moves we’ve ruled out without realising it.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marekkowal.substack.com