What if the surest way to fail at something ambitious is to have a clear plan to achieve it? What if a robot that doesn’t know it’s trying to walk learns faster than one explicitly trained to walk?
Kenneth Stanley is one of the most provocative minds in artificial intelligence. Former professor and team leader at OpenAI, he co-authored the cult classic Why Greatness Cannot Be Planned—a manifesto that challenges everything we think we know about achievement. His research discovered something shocking: in AI experiments, robots seeking novelty learned to walk better than robots explicitly trained to walk. Over the last two years, Ken has taken these theories from the lab into the real world. As Senior Vice President of Open-Endedness at Lila Sciences, he’s building scientific super intelligence—AI that autonomously discovers new biology and chemistry. He’s also the founder of Maven, a social network designed to replace popularity contests with genuine serendipity. His work reveals a profound truth: the path to greatness is paved with stepping stones you can’t predict, and the greatest discoveries come from following what’s interesting, not what’s planned.
What You’ll Discover:
🎨 The Picbreeder Paradox
- How people breeding random images discovered cars and butterflies—but only when not looking for them
- How an alien face became a car (and no one could have predicted it)
- The principle: in complex spaces, stepping stones don’t resemble final destinations
🤖 The Walking Robot Embarrassment
- How a robot seeking novelty learned to walk better than one trained to walk
- The critique isn’t about novelty’s magic—it’s about objectives’ fundamental flaw
🧭 Creativity as Intelligent Decision-Making Without Knowing Where You’re Going
- Why following interestingness isn’t random—it’s your sharpest compass
- How symmetry, oscillation, and other principles get locked in as stepping stones
- The collection metaphor: progress as gathering means, not optimizing toward ends
💼 The Founder’s Dilemma: Metrics vs. Meaning
- How to allocate resources between safety and serendipity
- Why there’s always an excuse to be conservative (investors, boards, grants)
- The regret of never following your true passion because you thought it wouldn’t get funding
🔬 Scientific Super Intelligence and the Future of Creativity
- Whether humans become curators of AI-generated ideas or retain creative edge
- Why 100 trillion parameter minds (human brains) might have infinite unique niches
- The gatekeeper argument: humans must decide which experiments get green-lit, even if we don’t fully understand
Key Insights:
“When we applied novelty search to teaching a robot to walk, the robot that was seeking novelty learned better how to walk than the robot that was being rewarded for walking better, which sounds completely paradoxical. The robot that doesn’t even know what it’s trying to do gets better than the one that is being trained to do that thing.”
“Creativity is the ability to make an intelligent decision without knowing where you’re going. You have multiple choices. You don’t know where you’re going, but you still choose a good direction to go. You’re choosing the direction that you think is most interesting. Things that have intrinsic potential are interesting and we have a very sharp nose for what’s interesting as human beings.”
“The thing you really want to do probably is the most impactful thing you could do. Because it’s what you’re truly genuinely interested in. It’s like your whole force and your whole life force is oriented towards this thing. That’s why you keep thinking about it. But you think it won’t get you the money. So you’re too afraid and you’ll never do it. Man, that would be regretful in the end.”
Resources:
Picbreeder (website)
The Picbreeder Experiment (pdf of workshop slides)
The Evolution of a Spaceship (website)
About Kenneth Stanley:
Kenneth Stanley is Senior Vice President of Open-Endedness at Lila Sciences, where he’s building scientific super intelligence capable of autonomous discovery in biology and chemistry. Former professor and team leader at OpenAI, he invented novelty search and neuroevolution techniques that revolutionized how we think about AI and achievement. Co-author of Why Greatness Cannot Be Planned, his research revealed that ambitious objectives are often counterproductive—robots seeking novelty outperform those trained for specific goals. He’s also the founder of Maven, a social network built on serendipity rather than popularity. His work bridges AI research with profound insights about human creativity, discovery, and what it means to build something genuinely new.
🎯 Perfect for: AI founders questioning their roadmap obsession, technical leaders interested in open-ended discovery systems, anyone feeling the hollowness of hitting metrics without fulfillment, or builders exploring how nature’s most creative force (evolution) has no plan.
⏰ Timestamps:
00:00 – Introduction: When Objectives Become the Enemy
03:01 – How AI Experiments Led to Social Critique
08:00 – The Picbreeder Experiment: Breeding Pictures Online
15:00 – A Car from an Alien Face (Deceptive Stepping Stones)
21:50 – It’s Not Random: The Sharp Compass of Interestingness
25:36 – From AI to Life Lessons: Permission to Follow Curiosity
32:55 – The Spaceship That Looked Like a Mistake
42:19 – Getting Lost in St. Petersburg: The Collection Metaphor
44:59 – Stepping Stones That Looked Like Mistakes
49:48 – Advice for the Founder Feeling Hollow Despite Hitting Metrics
53:29 – Balancing Investors, Boards, and Open-Ended Exploration
1:03:12 – Will Humans Become Curators of AI Creativity?
Subscribe for more conversations that matter, and let us know in the comments: What stepping stone in your career looked like a mistake at the time but turned out to be essential?
#KennethStanley #NoveltySearch #OpenEndedness #AI #Creativity #Innovation #Objectives #Picbreeder #Maven #LilaSciences