What this covers

Patrick talks with Ken Stanley, a computer science professor and neuroevolution researcher who co-wrote Why Greatness Cannot Be Planned. The conversation starts with what the critique covers and what it leaves alone. Modest goals such as shaving a lap time, upgrading software, or losing weight have known paths and suit metrics. Stanley's concern is the objective paradox, which applies to goals like curing cancer, reaching AGI, or lowering inflation, where the intermediate steps are unknown. He grounds it in history: researchers studying vacuum tubes from the 1700s had no thought of computation, and pulling them toward building computers would have erased the very prerequisite.

The middle of the episode follows Picbreeder, an image-evolution system. Stanley reports that in over 99 percent of cases, an interesting image's ancestors look nothing like it. An automated search aimed at a skull failed in every run, while human-guided exploration reached one in about 74 generations. He ties this to the like-causes-like fallacy, which he attributes to John Stuart Mill.

The institutional turns are the more contentious ones. Stanley says grant processes such as the NSF's reward predicted outcomes, and he contrasts Picbreeder with a crowd-voted sibling, Living Image, whose votes produced bland blobs through convergent consensus. He also argues that evolution works because survival and reproduction act as constraints rather than objectives, and that open-ended search needs such constraints. On the investing side, he calls AGI too many steps away to be a principled target, criticizes 2016 self-driving forecasts, and proposes costly endorsement as a signal worth trusting.

Sharpest takeaway

Stanley argues that setting ambitious objectives hurts progress toward discovery, because the stepping stones to a distant goal rarely resemble the goal. Innovation therefore depends on collecting interesting stepping stones and on avoiding the consensus and committee filters that prune them.

  • In Picbreeder, the ancestors of nearly every interesting image looked nothing like the final image, and an automated search aimed at a skull failed while human-guided search found it in about 74 steps
  • Committees and consensus votes converge on the status quo and prune the unusual stepping stones that enable breakthroughs
  • Interestingness, not an objective metric, should guide exploration, and constraints keep that search from degenerating into randomness

The claims · ranked63 claims · weighted by value

This asset isn't compiled yet

You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.

0.80

Modest objectives such as improving a lap time or upgrading software are reasonable and can be pursued with metrics, because their stepping stones are known; the critique applies only to objectives where the path is unknown.

normativehigh valueestablishednovelty 2/4durability 4/4· Ken

“So if you have what I call modest objectives in your life or if institutions have those, I think those can be reasonably pursued as they generally are.”

0.80

Vacuum tubes were studied for their electrical properties from the 1700s onward by researchers who had no thought of computation, yet this work became an essential prerequisite for the first working computers in the 1940s.

factualhigh valueestablishednovelty 2/4durability 4/4· Ken

“they weren't thinking about using it for computers at all.”

0.80

Major inventions happen only when the prerequisite technology becomes available, so launching computers in the 1840s or a space program in 1900 would have been insane, because the stepping stones were not yet in place.

causalhigh valueestablishednovelty 2/4durability 4/4· Ken

“It'd be insane to launch a project like that at that time or could the space program have been in 1900?”

0.80

Curiosity and exploration are principled for discovery, innovation, and self-fulfillment, but not for narrow goals with known paths, such as losing weight or driving safely, where exploration is dangerous.

normativehigh valueestablishednovelty 2/4durability 4/4· Ken

“The last thing I would want is a driver who's exploring methods of driving.”

0.78

Evolution is an open-ended stepping-stone search that produces birds, humans, and photosynthesis without any of them being written into its rules, so survival and reproduction serve as constraints rather than objectives.

causalhigh valuecontestednovelty 3/4durability 4/4· Ken

“None of them are the objective. None of them are written into the rules. If you just look at try to survive, you cannot predict bird.”

0.69

Artificial general intelligence is too many unknown stepping stones away to be a principled investment objective, but AI-adjacent applications such as computers doing human-like computation are the nearer stepping stones worth pursuing.

normativehigh valuecontestednovelty 3/4durability 2/4· Ken

“When I think about that, I think that is too many stepping stones away to realistically invest in.”

0.68

Science grant processes like those at the NSF send proposals to committees that require researchers to justify projects by their predicted outcomes, preventing them from pursuing unexplored ideas on their own.

factualhigh valuecontestednovelty 2/4durability 3/4· Ken

“I am not allowed to just go off on my own and work on that butterfly now.”

0.66

Moving quickly toward consensus in group settings cuts off unusual ideas at the start, producing convergent consensus that favors the status quo and eliminates the radical stepping stones innovation needs.

causalhigh valuecontestednovelty 3/4durability 3/4· Ken

“all the crazy things you would've done are basically cut off at the start by this surge towards consensus”

0.65

People intuitively assume that the things that lead to what they want will resemble what they want, a belief Ken calls the like-causes-like fallacy, which he attributes to Mill and considers a deep, possibly biologically built-in assumption.

definitionhigh valuefringenovelty 3/4durability 3/4· Ken

“It's called the like causes like fallacy. I think it's from Mills.”

0.63

Ambitious predictions made without the necessary stepping stones in place, such as 2016 claims about imminent self-driving cars, produce false visionary bets that lead to serious losses, including divisions being created and later sold off.

causalhigh valuecontestednovelty 2/4durability 2/4· Ken

“The stepping stones weren't there in 2016. If you invest a huge amount of money and so forth at that point in time, you are looking at some serious risk and loss”

0.61

The only way the Picbreeder discoveries happened was that their discoverers were not trying to reach them; had they targeted the final product, they would not have selected the intermediate stepping stones.

causalhigh valuespeaker onlynovelty 4/4durability 3/4· Ken

“the only way to discover any of these things was to not be trying to discover them.”

0.61

When the skull-matching search was automated and explicitly aimed at a skull, it failed every time across dozens of runs over 30,000 steps, although human-guided exploration found a skull in about 74 generations.

factualhigh valuespeaker onlynovelty 4/4durability 3/4· Ken

“We gave it 30,000 steps. This takes about 74 steps, let's say, for the first discovery by a human.”

0.61

The Living Image project, which used the same underlying code as Picbreeder but selected parents by crowd vote, produced washed-out, aesthetically bland blobs after roughly 25,000 votes, because votes diluted divergent preferences into the most pleasing average.

factualhigh valuespeaker onlynovelty 4/4durability 3/4· Ken

“after thousands and thousands of, I think it was 25,000 votes, you can look at the top ranking, all you have are amorphous rainbowy blobs”

0.61

A reliable signal that an idea deserves investment is when the person endorsing it gives up something of their own, such as time, resources, or a claim to their own ideas, because costly endorsement shows the idea matters to them.

normativehigh valuespeaker onlynovelty 4/4durability 3/4· Ken

“What should make the confirmation of something meaningful and really worth investment is if the person who's confirming it is giving something away.”

0.60

The reason discoveries follow indirect paths is that the world is deceptive, meaning the things that lead to a desired outcome typically do not look like that outcome.

causalhigh valuespeaker onlynovelty 3/4durability 4/4· Ken

“the things that lead to skulls don't look like skulls.”

0.60

If researchers working on vacuum tubes had been pulled into building computers, the stepping stone would have disappeared, so neither vacuum tubes nor computers would have resulted.

causalhigh valuespeaker onlynovelty 3/4durability 4/4· Ken

“You pull the people off the vacuum tubes and you make them work on computers. Well, you just took out the stepping stone that leads to computers.”

0.60

Collecting stepping stones is not random, because the stepping stones that get archived are chosen for interestingness, and interesting things carry potential to lead to other interesting things.

definitionhigh valuespeaker onlynovelty 3/4durability 4/4· Ken

“stepping stone collecting is anything but random”

0.60

Constraints are necessary in open-ended search because without them the archive fills with uninteresting, inert outputs; constraints select which stepping stones persist, making the search less random and keeping it interesting.

causalhigh valuespeaker onlynovelty 3/4durability 4/4· Ken

“Constraints are what make the collection of interesting stepping stones less random. It means that there's some principle that's still selected.”

0.59

The objective paradox applies when the stepping stones are mysterious, especially when several of them must be crossed on the way to the goal, so objectives such as curing cancer, reaching AGI, or lowering inflation fall into the problematic category.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“As soon as that's true where the stepping stones are mysterious, especially if there's several of them that we have to go through on the road to where we want to go”

0.59

Setting an objective actually hurts one's ability to reach ambitious, hard-to-reach goals when the stepping stones along the path are unknown, because the objective-driven method does not work for discovery.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“What the book is about is that it actually doesn't work.”

0.59

Two principles follow: first, collect stepping stones and honor interestingness as a system that individuals and institutions can deliberately build; second, recognize stepping stones when they first snap into view and become possible.

normativehigh valuespeaker onlynovelty 2/4durability 4/4· Ken

“One is collect stepping stones and honor interestingness.”

0.59

In over 99 percent of Picbreeder cases, tracing an interesting image back through its lineage reveals that at some point the stepping stones look nothing like the final product.

factualhigh valuespeaker onlynovelty 4/4durability 3/4· Ken

“if you go back far enough, you will find a stepping stone that looks absolutely nothing like it in 99.9% of cases.”

0.58

The hard constraint that organisms must make functioning copies of themselves keeps evolution from degenerating into inert blobs, because if everything reproduced regardless of complexity the space would fill with worthless forms.

causalhigh valuespeaker onlynovelty 3/4durability 4/4· Ken

“Well, the world would devolve into all kinds of inert blobs covering the earth. The constraint is basically a way of keeping things interesting.”

0.57

In Picbreeder, seeing and building on others' discoveries is implicit collaboration, but the moment a user chooses to branch from a discovery, they act alone, a structure rarely used in culture's group-based collaboration.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“the moment you choose to continue or what we call branch from the butterfly to breed it further, you are on your own.”

0.57

Steve Jobs's iPhone is better understood as recognizing that the prerequisite stepping stones had suddenly been laid, not as a visionary foreseeing the final product, and the skill that matters is recognizing when stepping stones snap into place.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“he was the first person to recognize that all the stepping stones had suddenly been laid”

0.57

Questions such as finding satisfaction, love, or one's purpose are blue-sky discovery problems with unclear stepping stones, so the same objective paradox applies to them and they cannot be solved by planning toward a fixed goal.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“In this self-discovery or 'What am I going to do with my life?' type of stuff, I think that's analogous to blue sky discovery at a society level.”

0.57

Innovation is limited by narrow organizational filters that prevent ideas from reaching the few people who could act on them, so the key failure is insufficient exposure to stepping stones rather than a lack of ideas.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“we have these filters, which are extremely narrow, which decide what comes through, and they end up pruning out things.”

0.57

A committee of experts tends to represent the status quo, whereas one person on the periphery may be the only one to see a idea's potential, so connections depend on individuals rather than group assessment.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“A committee of five people is crazy in that case. Even if they're experts in the fields, whatever, they just basically represent the status quo.”

0.57

Sharing a discovery publicly, so anyone can judge it, is the point of collaboration that matters, because it exposes stepping stones across society without a committee filter.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“It's just airing it to the world. Everybody can look and decide for themselves this is a stepping stone”

0.57

During wartime or crisis, society recalibrates what is interesting, making related problems more salient and drawing focus toward them, which is a valid shift in interestingness rather than the adoption of an objective.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“we recalibrate the things that we find interesting, depending on what's going on in the world around us.”

0.57

What counts as interesting changes constantly, and novelty is an intrinsic part of interestingness, so heuristics that identified novelty 20 years ago become obsolete and must be continually refined.

factualhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“What is interesting is changing constantly.”

0.57

There is no good strategy for reliably achieving a specific thing one does not know how to do; the most that can be done is to maximize the chance of achieving something useful.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“There's just no principled way to do that. What I can give you advice on is how to maximize the chance of achieving something useful, but not a specific thing.”

0.57

Our culture designs nearly everything around setting a goal, adhering to it, and measuring progress toward it, but this objective-driven approach does not work for innovation and discovery, and the flaw is largely unrecognized because no one debates it.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“What the book is about is that it actually doesn't work.”

0.57

A practical self-discipline is to ask where people would predict you would go next and then deliberately avoid that direction, to keep moving into novel territory rather than exploiting a comfortable corner.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“Where would people predict if they were me they would go and then not go there?”

0.57

Gatekeepers who control innovation funding should greatly broaden the filters between ideas and the people who could run with them, and should base funding decisions on whether one or two enthusiastic people see potential rather than on consensus or votes.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“What we should do is greatly broaden the filters that go from idea to exposure to the people who could run with the ideas”

0.57

Interestingness is intentionally hard to define because it is meant to encompass everything that matters in life, and it differs between people, but it remains a constrained concept that field experience can refine.

definitionhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“It's a hard to define thing, which is intentional here because it's supposed to encompass everything that matters in life.”

0.56

The standard narrative of achievement, in which a visionary sets an objective and a team executes it, is misleading because it omits the earlier stepping stones, such as vacuum tubes, that made the achievement possible.

normativehigh valuespeaker onlynovelty 2/4durability 4/4· Ken

“The story doesn't start where it starts, and that makes the story extremely misleading”

0.56

The archive of accumulated stepping stones grows more powerful the longer it runs, which is why people in 2022 can reach more places than people in 1922.

causalhigh valuespeaker onlynovelty 2/4durability 4/4· Ken

“The archive gets more powerful the longer it runs, the more stepping stones we have. That's why from today, the year 2022, we can get to more places than we could in 1922”

0.56

Having modest aims is a legitimate, less risky life approach, but pursuing discovery and remarkable outcomes requires accepting risk, because the stepping-stone process inherently involves failures that must be treated as a portfolio.

normativehigh valuespeaker onlynovelty 2/4durability 4/4· Ken

“The whole idea of our interestingness, repertoire, stepping stones, it's all about risk. So you have to take risks in order to have amazing ends.”

0.53

In Picbreeder, a car's wheels derived from the eyes of an alien face, showing that an intermediate image with no visible relation to the final product can be the necessary step to it.

factualhigh valuespeaker onlynovelty 2/4durability 4/4· Ken

“the wheels of the car, which was depicted from the side, actually derived from the eyes of the alien face.”

0.53

The lines around the block for the first iPads reflected that people saw the device as a stepping stone to unknown future uses, not merely for its current capabilities.

causalhigh valuespeaker onlynovelty 3/4durability 2/4· Ken

“What was interesting about the iPad wasn't what you could do with it, it's that we don't know all the things that you might do with it”

0.52

Necessity drives invention when stepping stones already exist, because crises make the relevant stepping stones salient, but genius or effort cannot produce solutions when the stepping stones are absent.

causalhigh valuespeaker onlynovelty 2/4durability 3/4· Ken

“It doesn't matter how many geniuses were in the room.”

0.52

Organizations seeking innovation should maximize stepping stones and exposure to them rather than objective performance, which means broadening the filters that route ideas to people who could run with them.

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Ken

“What we should do is greatly broaden the filters that go from idea to exposure to the people who could run with the ideas”

0.52

Beyond curiosity and exploration, the key superpower in discovery is recognizing deception, meaning that progress toward a goal often looks good while not actually leading to it, and this pathology is seriously underrecognized.

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Ken

“I would maybe add deception. It's curiosity, exploration, the idea of deception, which is when you're moving towards something and things seem to be going well, often they're not, actually.”

0.52

Someone who has spent years in a field is qualified to recognize what is interesting there, and this recognition is a constraint on search that is distinct from having a fixed objective.

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Ken

“If it's a field that's your field, if you spend years and years in a field, you're qualified to understand what's interesting”

0.52

People who find a low-hanging, unoccupied corner of the search space and become famous for it often exploit it until it is exhausted and then become unable to leave, because their reputation discourages moving on.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Ken

“You'll be famous after that, and then that'll look really good and you can live out your whole life based on that”

0.50

Trying to control innovation toward a chosen destination destroys it, and the vacuum tube and computer stepping stones happened by serendipitous luck rather than institutional planning.

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Ken

“You absolutely destroy innovation if you try to control it in the way that we do try to control it.”

0.50

For individuals, the practical approach is to pursue interesting activities as personal stepping stones, giving oneself multiple opportunities and using them as jumping-off points toward satisfaction, rather than planning a fixed objective and fearing randomness.

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Ken

“You give yourself multiple opportunities and you use those as jumping off points to find what makes you satisfied, interested, rich”

0.47

Culture designs nearly everything around a single archetype: set a goal, stay on objective, define a metric to measure progress, and concentrate all effort on moving directly toward the goal, on the belief that this formula delivers everything needed, including staying ahead of the curve and avoiding disruption.

factualhigh valuespeaker onlynovelty 2/4durability 3/4· Ken

“basically design everything that we do around exactly that archetype”

0.44

Picbreeder was a crowdsourced AI experiment in which internet users bred images by selecting favorites from each generation, with the full lineage recorded, creating a complete history of discovery.

definitionestablishednovelty 1/4durability 4/4· Ken

“Picbreeder was an experiment that I was running with my lab.”

0.44

Objective-driven culture produces stifling psychological pressure that begins in early childhood, creating a constant objective vice that is toxic and generates anxiety, which Ken suggests people intuitively sense is unnatural.

causalhigh valuespeaker onlynovelty 2/4durability 2/4· Ken

“It's just stifling over time. It seems to grip you from a really early age, from six or seven years old or something.”

0.34

Discovery depends heavily on the cost of experimentation, since tens of thousands of generations are cheap for a computer but prohibitively expensive for a human, so policies should promote low-cost, fast experimentation.

normativespeaker onlynovelty 2/4durability 3/4· Patrick

“30,000 generations is cheap for a computer. It's freaking expensive for a human. What you would want is to promote low cost and fast experimentation”

0.34

The shift from whale oil to rock oil for street lamps, which led to kerosene and later products such as the microchip, illustrates how stepping stones proliferate into outcomes that cannot be foreseen.

factualspeaker onlynovelty 2/4durability 3/4· Patrick

“whale oil for street lamps going to rock oil for kerosene, again, just for street lamps, and then all the things that it led to or the microchip”

0.31

The flaw in objective-driven achievement is a hidden problem that lacks public controversy or sides, unlike typical political debates, even though it has serious social consequences.

factualspeaker onlynovelty 2/4durability 2/4· Ken

“Nobody's talking about this. There are no sides.”

0.29

If the like-causes-like assumption were true, it would be obvious what to do, and cancer would already be cured, because the path to any goal would simply be the step that moved closest to it.

causalspeaker onlynovelty 2/4durability 3/4· Ken

“If that was true, then there would be no such thing as complexity.”

0.26

Exploration is often seen as wasting time, but in discovery-oriented work pursuing what one wants, such as 20 percent time, is the principled use of time rather than a waste.

normativespeaker onlynovelty 1/4durability 2/4· Ken

“It's my 20% time.”

0.26

Ken credits a friend named Ed, who funded his early research from his personal bank account and encouraged him, as the kindest thing anyone has done for him.

factualspeaker onlynovelty 0/4durability 4/4· Ken

“His name was Ed, who gave me money to do research out of his personal bank account.”

0.26

A friend named Ed gave Stanley research money from his personal bank account, unusual for academia, encouraged him as a father figure, and profoundly shaped his career, leading Stanley to name his son Edward after him.

factualspeaker onlynovelty 0/4durability 4/4· Ken

“His name was Ed, who gave me money to do research out of his personal bank account.”

0.26

Individuals, corporations, and government all rely on goal-setting as the one heuristic for reaching high-level achievement, starting from childhood goals encouraged by parents and schools.

factualspeaker onlynovelty 2/4durability 2/4· Ken

“All of these institutions and individuals are using goal-setting as the one heuristic for getting to high-level achievement.”

0.26

The failed objective-driven runs produced images that resembled the skull's silhouette or shadow but never got close, illustrating that partial progress toward a target can be deceptive.

factualspeaker onlynovelty 2/4durability 2/4· Ken

“It's like getting the silhouette shadow of what it wants, but it can't get even close.”

0.26

Truly incredible ideas are typically recognized by only about one in a hundred people, so organizations must find ways to honor and protect those rare recognitions rather than rely on broad consensus.

forecastspeaker onlynovelty 2/4durability 2/4· Ken

“When there's something incredible, it's not going to be tons of people see, it's going to be one out of a hundred see it”

0.21

Recognizing stepping stones is one of humanity's special skills, so the principle should not be frightening, because people are naturally able to see when an interesting thing opens new possibilities.

factualspeaker onlynovelty 1/4durability 2/4· Ken

“Human beings understand stepping stones. That's one of the reasons we shouldn't be so scared at the message that I'm giving here.”

0.21

Leaders who want to depart from objective-driven evaluation must have the courage to fight their organizations, because superiors and funders will not understand abandoning standard metrics and the change will not happen otherwise.

normativespeaker onlynovelty 1/4durability 2/4· Ken

“So this takes some courage, I think.”