In fact, we believe that human achievement has no limits. It’s just that we’re going to highlight a different path to achievement, without the need for objectives. (Location 142)
They’ve stolen our freedom to explore creatively and blocked us from serendipitous discovery. (Location 143)
if only we can let go of the security blanket of the objective. (Location 145)
For example, scientists often demand hypotheses from each other because they don’t want to fund research only because it sounds “interesting.” (Location 154)
Usually progress towards the objective is somehow measured (Location 165)
The process of setting an objective, attempting to achieve it, and measuring progress along the way has become the primary route to achievement in our culture. (Location 176)
objectives become more ambitious, reaching them becomes less promising—and that’s where the argument becomes most interesting. (Location 182)
think about all the things that are possible (Location 195)
Of course, the vast majority of possible images are of no interest whatsoever—like television screens tuned to the wrong channel, just random static. (Location 199)
do not yet exist because no one has painted or drawn them yet. Put another way, these future masterpieces are not yet discovered (Location 201)
So we can think of creativity as a kind of search. (Location 209)
The nice thing about thinking of discovery in terms of this big room is that we can think of the process of creation as a process of searching through the space of the room. (Location 217)
sense, civilization has been exploring this room since the dawn of time. (Location 219)
In this way, artists are searching the great room of all possible images for something special or something beautiful when they create art. The more they explore the room, the more possibilities open up. (Location 221)
In a sense, the places we’ve visited, whether in our lives or just in our minds, are stepping stones to new ideas. (Location 226)
Before anyone had explored the part of the great room filled with computers, no one could know what was even possible there. (Location 240)
Stepping stones are portals to the next level of possibility. Before we get there, we have to find the stepping stones. (Location 243)
The idea of a large imaginary “room” whose space is being searched is a metaphor that helps to see why this principle holds. (Location 244)
search space, that refers to this very concept—it’s the general idea that creation and discovery happen within a space of possibilities that contains stepping stones leading from one discovery to another. (Location 246)
want to be in the future—the path to the objective. Sometimes, figuring out the path from here to the objective isn’t exactly a major challenge. (Location 248)
In other words, you’ll have to search for the right stepping stones, and if you’re lucky and clever enough, you might discover the ones that lead to the objective. (Location 252)
Objectives are well and good when they are sufficiently modest, but things get a lot more complicated when they’re more ambitious. (Location 257)
like those involving discovery, creativity, invention, or innovation—or even achieving true happiness. (Location 258)
Not only that, but this paradox leads to a very strange conclusion—if the paradox is really true then the best way to achieve greatness, the truest path to “blue sky” discovery or to fulfill boundless ambition, is to have no objective at all (Location 260)
Put another way, if you were walking through one of these great rooms of all possible things, how the items in the room are arranged would be unpredictable and confusing. (Location 263)
The problem is that the stepping stone does not resemble the final product. (Location 272)
once you’ve got vacuum tubes you’re very close to having computers, if only you could see the connection. (Location 274)
The problem is, who would think of that in advance? The arrangement, or structure, of this search space is completely unpredictable. (Location 275)
The first engine was not invented with airplanes in mind, but of course the Wright brothers needed an engine to build a flying machine. (Location 277)
the rest of your life to refining vacuum tube technology. But in all these cases, what you would never do is exactly what you should have done. (Location 284)
The paradox is that the key stepping stones were perfected only by people without the ultimate objective of building microwaves, airplanes, or computers. (Location 285)
The structure of the search space—the great room of all possible things—is just plain weird. (Location 286)
It’s so bad that the objective can actually distract you from its own stepping stones! (Location 287)
The problem is that ambitious objectives are often deceptive (Location 288)
The problem is that ambitious objectives are often deceptive (Location 288)
But strangely in the end we often must give them up ever to have the chance of reaching them. (Location 289)
Chances are that if we plan a path based on our objective, then it will miss the stepping stones. (Location 291)
It often turns out that the measure of success—which tells us whether we are moving in the right direction—is deceptive because it’s blind to the true stepping stones that must be crossed. (Location 295)
At a deeper level, we might ask why we think ambitious pursuits should be driven by objectives at all (Location 297)
“If you’re going to come up with a new direction or a really new way to do something, you’ll do it by just playing your stuff and letting it ride. (Location 304)
We just happened to be traveling the right path through the search space of musical genres to bump into it in the late 1940s [6]. (Location 307)
Who would have thought that “acting the fool,” and not some ambitious drive to overhaul popular music, would transform the world of rock and roll? (Location 313)
but that having no objective can lead to the greatest discoveries of all. (Location 314)
and the promise of success if only we try hard enough. No doubt, the thought of aimlessly wandering through the space of possibilities with no clear purpose is not going to inspire many can-do achievers. (Location 317)
We want to show you that it’s possible to explore a search space intelligently even without an objective. (Location 319)
We can align ourselves towards discovery and away from the trap of preconceived results. (Location 321)
Sometimes the best way to achieve something great is to stop trying to achieve a particular great thing. (Location 322)
Even though serendipity is often portrayed as a happy accident, maybe it’s not always so accidental after all. In fact, as we’ll show, there’s a lot we can do to attract serendipity, aside from simply betting on random luck. (Location 326)
We’ll show you why not only can you trust your gut instinct when it tells you something important is around the corner, but you should trust it, even if you can’t explain what that something is. (Location 329)
Do you ever tell yourself that you can’t do something because it’s not justified by a clear purpose? (Location 334)
It’s about how you choose your major, find your spouse, and choose your profession. (Location 352)
Sometimes things work out the way we intended. Sometimes you study accounting and become an accountant. (Location 370)
But the great successes, the ones that come crashing out of nowhere and shake up the system, they don’t usually follow this kind of script. (Location 371)
Being open and flexible to opportunity is sometimes more important than knowing what you’re trying to do. (Location 385)
It turns out that those who focused on the goal of counting the photographs took significantly longer to complete the task than those who were less focused on the objective. (Location 389)
planned. It was simply that Depp wanted to make music and was open to opportunities when they came. (Location 395)
The common thread in all these tales is that the successful wander from their original paths, (Location 429)
Note the emphasis on unplanned experience. This isn’t the usual attempt to figure out the best job and pursue it as an objective. (Location 439)
Because the stepping stones that lead to the greatest outcomes are unknown, not trying to find something can often lead to the most exciting discoveries (or self-discoveries). (Location 444)
The key is to be open to change, to a shifting landscape where appearances can be deceiving yet liberating at the same time. (Location 447)
What is important in these scenarios is to avoid locking into rigid commitment to the original ambitious objective, and instead remaining mindful and open to where the present stepping stone might lead. (Location 449)
But there’s a deeper value to such endeavors than just the endeavors themselves: They reflect that we don’t know which stepping stones might lead to something interesting. (Location 484)
But what if Joseph Herscher one day wakes and realizes that one of his Rube Goldberg machines solves an unexpected problem? (Location 488)
Interestingly, this kind of pivot is fairly common behind the scenes of many success stories, such as in today’s Internet businesses. (Location 496)
If you take one thing from this chapter, perhaps it should be that you have the right to follow your passions. Even if they deviate from your original plans or conflict with your initial objective, the courage to change course is sometimes rewarded handsomely. (Location 509)
Of course life is full of risk and some choices indeed won’t work out, but few achieve their dreams by ignoring that feeling of serendipity when it comes. (Location 516)
there is ample evidence that sticking to objectives just isn’t part of the story in many of the biggest successes. (Location 519)
In fact, it turns out that we stumbled upon this principle originally through a scientific experiment, as the next chapter reveals, involving hundreds of users on the internet breeding pictures. (Location 521)
In other words, Picbreeder seemed to work best when visitors were open minded about what they hoped to find. To see how we can be sure of that, and why it ultimately relates to the impact of objectives in many facets of life, it helps to understand a few details on how the site is set up. (Location 606)
Even with the new artificial DNA, we hadn’t realized how lifelike and meaningful the images would become. (Location 638)
It turns out that it’s a bad idea to set out with the goal of evolving a specific image. In fact, once you find an image on Picbreeder, it’s often not even possible to evolve the same image again from scratch—even though we know it can be discovered! (Location 647)
The only time these images are being discovered is when they are not the objective. (Location 653)
If you want to find a meaningful image on Picbreeder, you’re better off if it is not your objective. (Location 670)
In other words, the most successful users have no objective. (Location 679)
other words, no matter how tempting it is to believe in it, the distant objective cannot guide you to itself—it is the ultimate false compass (Location 692)
In fact, the term “stepping stone” is meant to remind us of this idea of crossing a lake on small stones protruding through the water. (Location 719)
Because of such uncertainty, the lake we’re crossing is often shrouded in mist. (Location 722)
The problem of predicting stepping stones becomes especially hard for ambitious objectives. (Location 726)
Another way to look at the challenge of ambitious problems is to say that their solutions are more than one stepping stone away (Location 728)
It’s literally a method to measure progress towards the objective. (Location 736)
The deception of the Chinese finger trap is that the path to freedom is to push inward, away from freedom. (Location 755)
If your objective is freedom from the Chinese finger trap, then measuring progress by how close you are to freedom is exactly the wrong approach. (Location 757)
In fact, the search space of any complex problem is sure to be littered with deceptive stepping stones. (Location 761)
To arrive somewhere remarkable we must be willing to hold many paths open without knowing where they might lead. (Location 776)
Natural evolution is a process of discovery and creation that produced an astronomical diversity of exceedingly complex organisms. (Location 786)
include 100 trillion neural connections, far exceeding the complexity of anything designed by engineers. (Location 788)
The problem is that the stepping stones to intelligence do not resemble intelligence at all. (Location 807)
Is it any less ridiculous to try to achieve any far-off, ambitious objective by measuring how close it is to our best candidate so far? (Location 819)
Earth was not trying to evolve human-level intelligence. And that is the only reason it did (Location 825)
is by not setting it as the objective. (Location 827)
but because they are well-adapted in their own right. (Location 830)
because it was successful in its own niche at its own time. (Location 831)
are spun around open-ended constraints like “survive and reproduce,” because meeting a constraint is much different from what is usually meant by objective-driven achievement. (Location 845)
Perhaps survive and reproduce can be more naturally seen as a constraint on evolution. (Location 857)
But it describes nothing about the products that might be created, nothing about the difference between where we are today and where we might be tomorrow, and nothing about the potential for greatness lurking behind (Location 859)
Constraints like survive and reproduce are something else—they are part of the alternative, in which we let go of defined objectives and instead explore the stepping stones. (Location 863)
it allows evolution to identify successful organisms that may lead to other successful organisms, nothing less and nothing more. (Location 865)
The fundamental problem is that 5,000 years ago the stepping stones to efficient computers were not yet uncovered. (Location 875)
In fact, it’s much more likely that you will think of vacuum tubes if you’re interested in electrical experiments, not computation. (Location 881)
The best way to get computation is not to force great minds to waste their lives pondering a distant dream, but to let the great minds pursue their own interests in their present reality. (Location 882)
Because most interesting inventions are simply the most recent fruits of chains of ideas spanning centuries, they too will necessarily depend upon a prior invention created for an entirely different purpose. (Location 886)
Almost no prerequisite to any major invention was invented with that invention in mind. (Location 889)
We simply lack the foresight to comprehend what one discovery will later make possible. (Location 893)
We hear tales of geniuses like the Wright brothers who invented machines that could fly, these can-do achievers who strived heroically against all odds to change the world forever. (Location 894)
But the Wright brothers can’t claim credit for the endless chain of human innovation that precedes them. (Location 896)
but to recognize that we were only a stepping stone away from flight given past innovations. (Location 906)
Great invention is defined by the realization that the prerequisites are in place, laid before us by predecessors with entirely unrelated ambitions, just waiting to be combined and enhanced. (Location 907)
Just as nature and Picbreeder are both stepping-stone collectors, so is the tree of human innovation, growing ever-outward, towards computers, Internet, cars, and planes—everywhere you might want to go and at the same time paradoxically nowhere in particular. (Location 912)
When we unleash search from the trap of the objective, liberating it from the requirement to move only towards where we hope to arrive, it becomes a kind of treasure-hunter that finds needles in the haystack of what’s possible. (Location 917)
from finding the perfect partner to creating the next great invention, when objectives are ambitious, the only reward you’re likely to receive is deception. That’s what you get for traveling by the false compass. (Location 920)
your passion, (Location 926)
Passion is what drives you to that point, and then one day you might realize that you are only one stepping stone away from being rich. (Location 926)
On the contrary, a single-minded preoccupation with money is likely exactly the wrong road to abundant wealth. (Location 928)
We should be concerned by the disconnect between how the world is supposed to work and the way it really does work. (Location 932)
But this philosophy leads to absurdities if taken literally. You can’t evolve intelligence in a Petri dish based on measuring intelligence. (Location 934)
Human intelligence did evolve, but it was never set as an explicit objective for evolution. (Location 939)
exploration swing free, that the farthest frontiers are conquered. (Location 944)
The problem is that the ambitious objectives are the interesting ones, and the idea that the best way to achieve them is by ignoring them flies in the face of common intuition and conventional wisdom. More deeply it suggests that something is wrong at the heart of search (Location 946)
The world’s greatest compass can get us lost while a mysterious form of ignorance turns out to be surprisingly powerful. (Location 948)
It doesn’t take much effort to state what you want to be, where you want to go, or what you want to accomplish. The problem is getting there. (Location 971)
Perhaps the effort invested in specifying clear objectives would be sometimes better spent in identifying promising stepping stones, but not specifically those leading to the objective. (Location 973)
So maybe there’s a different way to think about search. Instead of worrying about where we want to be, we could compare where we are now to where we have been (Location 974)
Our preoccupation with objectives is really a preoccupation with the future. (Location 977)
We can instead compare to the past, which is a lot easier than the future because we know it—the past already happened. (Location 983)
Instead of judging our progress towards a goal, the past allows us to judge our liberation from the outdated. (Location 986)
The point is that novelty can often act as a stepping stone detector because anything novel is a potential stepping stone to something even more novel. (Location 988)
Interesting ideas are those that open up new possibilities. (Location 990)
The important point is that novelty (and interestingness) can compound over time by continually making new things possible. (Location 995)
novelty the reward is an endless chain of stepping stones branching out into the future as novelty leads to further novelty. Rather than thinking of the future as a destination, it becomes a road, a path of undefined potential. (Location 996)
ratcheting processes that build stepping stone upon stepping stone, branching and diverging ever outward to everywhere and nowhere in particular. (Location 999)
It was worth exploring the path opened by the Alien Face because it held potential (Location 1001)
How do we know where we’re going? But that’s exactly the point. (Location 1005)
Rather than relying on a false compass, novelty only asks us to compare where we are with where we’ve been. (Location 1014)
Deviating from the past is simpler and richer with information because we can look at the whole history of past discoveries to inform our judgment of current novelty. So it’s not unreasonable to believe that novelty is a meaningful engine for progress. (Location 1016)
So it’s no coincidence that the concept of interestingness comes up naturally when discussing novelty. When an idea feels genuinely novel, that’s often enough to make us curious. (Location 1020)
The reality is that we humans have a nose for the interesting. We understand that if we take the interesting path, it may yet lead somewhere important, even though we might not know where. (Location 1025)
Behind any serendipitous discovery there’s nearly always an open-minded thinker with a strong gut feeling for what plan will yield the most interesting results. (Location 1040)
“In the fields of observation chance favors only the prepared mind [ (Location 1044)
To investigate how promising it is to search without objectives, we’ll design a kind of computer program, or algorithm, that has no objective. (Location 1049)
The first step towards testing such a program is to decide on what’s called the domain (Location 1068)
simply try to do something different than it has done in the past. (Location 1087)
The only way for the robot to do something novel is now by not crashing (Location 1101)
At the same time, it might seem that this apparent success is simply a result of “trying everything” (which computer scientists call exhaustive enumeration (Location 1107)
The reason it’s more interesting than that is that novelty search tends to produce behaviors in a certain order (Location 1110)
For objective-driven search, what we usually expect is bad behavior before good behavior. (Location 1113)
The problem is that the order from bad to good, while attractive, is unfortunately naive. (Location 1115)
there, the only order that leads to success is from bad to worse before getting better. (Location 1116)
Although it might appear that being more novel is “better” than being less novel, it all depends on what you’ve seen so far—someone else with different experiences might come up with exactly the opposite judgment. (Location 1123)
The reason is that as soon as a novel behavior is discovered, it quickly becomes less novel as similar behaviors are discovered. (Location 1125)
novel depends completely upon time and context. (Location 1126)
novelty is a relative measure. (Location 1127)
Objective-driven search is much different: A car with a more efficient engine is always more efficient than a car with a less efficient engine, no matter when either of them was invented. (Location 1129)
But it does in fact provide a more interesting ordering: from simple to complex (Location 1132)
When all the simple ways to behave are exhausted, the only new behaviors that remain to discover are more complex. (Location 1135)
And most often those very first behaviors will tend to be simple, just as simple inventions are stepping stones to more complex inventions. (Location 1137)
There’s much that could be said about the nature of simplicity, but for our purposes the key is that it requires no information or knowledge about the world. (Location 1139)
Eventually doing something genuinely novel always requires learning something about the world. (Location 1144)
Because eventually you have to acquire some kind of knowledge to continue to produce novelty, it means that novelty search is a kind of information accumulator (Location 1146)
The longer the search progresses, the more information about the world it ends up accumulating. (Location 1147)
Novelty search is a special kind of non-objective search, but others, such as natural evolution, clearly exhibit this signature dynamic as well. (Location 1150)
it shares with novelty search the key property that what it ultimately produces is not its objective. (Location 1152)
That’s why increasing complexity is almost inevitable if evolution is to continue. (Location 1156)
But they can also be viewed as the inevitable tendency of a search with no final objective to accumulate information about its world. (Location 1159)
Then it will become a part of evolution’s accumulated inventory of information. (Location 1164)
The ability to learn and adapt over our lifetime has propelled the evolutionary information accumulator to a recent extreme. (Location 1168)
But what we observe again is that a search without a clear objective (evolution in this case) accumulates information as it moves from the most simple single-celled organisms to the most complex animals. (Location 1169)
None of these were its objective, but they are all being discovered. (Location 1174)
For novelties to be continually found, eventually the robot would have to discover that the world is made of walls and doors, and that robots crash into walls but fit through doorways. (Location 1182)
One reason objectives are reassuring is that they can filter an overwhelming space of possibilities to only a few practical options. (Location 1184)
In other words, objectives can weed out clearly inferior ideas, which could eliminate pointless effort we might otherwise invest in irrelevant activities. (Location 1186)
The world provides its own constraints. Many behaviors that are possible to imagine (like flying) would not be considered by novelty search because they’re impossible (Location 1189)
In general, many behaviors that we can imagine collapse to the same actual behavior in reality. (Location 1192)
This idea that the physics of the world limits the behaviors considered by novelty search also explains its tendency to accumulate information about the world: (Location 1198)
are the ones that respect how the world actually works (Location 1199)
best way to create novelty is to exploit the way the world really works and accumulate information about it. (Location 1202)
As we review some of these arguments next, you might be reminded of the discussion from the previous chapter on why survive and reproduce isn’t a traditional objective. (Location 1209)
When we think of objectives, we usually assume that our objective is satisfied when we achieve what we set out to achieve. (Location 1212)
novelty doesn’t work that way. (Location 1214)
but in that case becoming a banker was not your objective. (Location 1215)
With novelty, what you become or what you achieve is never your objective. So there’s something different about novelty. (Location 1215)
But—here’s the second problem—being novel is an elusive and slippery concept. (Location 1218)
And convergence means that many potentially interesting directions will not be explored. (Location 1235)
diverge while collecting new stepping stones along the way. (Location 1237)
term divergent thinking is associated with creativity and innovation. (Location 1238)
they are known for fearless and surprising discoveries that others tend to miss. (Location 1239)
Experiments in simulation are common in the field of AI because the simulated robot can try new behaviors over and over again very rapidly and with no risk of damage. (Location 1247)
In other words, the hope for novelty search is that good ideas could be stepping stones to something interesting. (Location 1257)
You may have an interesting idea and then after thinking about it for a while realize that it opens up other interesting ideas. (Location 1263)
So if the robot falls on its face, that’s good as long as the robot never fell down before in the same way. What do you think a biped robot looking for novelty would eventually end up doing? (Location 1294)
In the Czech Republic, Peter Krcah found that novelty search helped to solve deceptive problems when searching not only the behaviors but also designing the bodies of simulated robots (Location 1307)
but instead evidence that in general novelty search may sometimes yield a better result than searching for a specific objective. (Location 1313)
Maybe you simply can’t always get what you want when you want (Location 1323)
imagined. Ultimately, there may be futility at the heart of search, as the next chapter will explore. (Location 1324)
though we won’t sacrifice our optimism in exchange. Instead, we’ll discover a source of optimism that embraces the uncertainty of the far future rather than fearing or denying it. (Location 1359)
This kind of possible failure illustrates that it’s important not to expect too much: Novelty search is not a solution to all our problems—even though it might sometimes work better than pursuing a particular objective. (Location 1373)
Instead, all we’d need to do if we’re actively pursuing an objective is keep around a variety of alternative stepping stones, just to be safe. (Location 1385)
The more ambitious our objectives, the more deceptive they become as well. Deception is just a very nasty beast. And when you’re dealing with deception, by definition the objective is a false compass. (Location 1394)
going around the wall is likely to be cut off prematurely. The problem is that the only time the objective actually pushes the search in the correct direction is at the very end (Location 1408)
Learning to run towards a goal that is directly in front of you is not what one would call a challenge. (Location 1410)
The more troubling lesson is that objective-driven search is sometimes embarrassingly incompetent. (Location 1428)
we’re left with the stark reality that nothing can reliably reach particular target objectives. (Location 1441)
In fact, it turns out that improving a search process to reach certain objectives will always hurt its performance on a different set of problems. In short, you really can’t win them all. (Location 1446)
We can reliably find something amazing. We just can’t say what that something is! (Location 1452)
great discoveries are possible if they’re left undefined. (Location 1453)
But the more important lesson of non-objective search is that it’s a powerful treasure hunter (Location 1455)
instead of finding a particular treasure that you might have in mind, as it diverges through the search space it finds many treasures, (Location 1456)
To highlight the irony here, the Car never would have been discovered if someone had set out to find it. (Location 1461)
This strange paradox, where trying is a curse and not trying is a blessing, (Location 1462)
It means that ambitious goals can’t reliably be achieved by trying—unless they are one stepping stone away, where they come within reach. (Location 1463)
The treasure hunter is an opportunistic explorer—searching (Location 1465)
To be a treasure hunter, you have to collect as many stepping stones as you can, because you never know which one might lead somewhere valuable. (Location 1466)
One implication of Picbreeder and novelty search is that we can actually build systems based on the non-objective principle. (Location 1478)
but once we escape the myth of the objective, the possibility of building treasure-hunting systems becomes intriguing. (Location 1479)
That’s not the treasure hunter—consensus is exactly the cultural tendency that we need to escape. (Location 1483)
No, the way to unleash the treasure hunter is actually through separating people from each other, like in Picbreeder, where people only interact by taking off from where someone else left. (Location 1485)
the system as a whole ends up lacking a unified objective because people’s objectives differ. (Location 1488)
Interestingly, the internet provides an exciting opportunity (Location 1491)
If you were buying a chair then a set of available models would greet you. (Location 1495)
instead of choosing one model to buy as normal, you could request a set of variations of your favorite option. (Location 1496)
but one generated based on the model you chose just before. (Location 1500)
the magic of the treasure hunter begins to come out: The site becomes a stepping stone collection, a stockpile of discoveries in a growing database of chair designs, each one possibly a stepping stone to even more appealing chairs. (Location 1506)
Manufacturing processes may have to catch up (perhaps aided by the rapid progress and increasing interest in three-dimensional printing machines), (Location 1513)
no longer be restricted to the designs of expert designers. (Location 1515)
thought experiment raises an intriguing question: Would you trust this kind of catalog? Would you trust a catalog in which essentially nothing is designed by an expert and everything is “discovered” by its users? What would you expect the collection of articles to be like? Could they possibly be worth your money or attention, or would they look like the second-rate dabbling of aimless amateurs? (Location 1517)
The explorers are still not furniture designers and also have no unified design goal among each other. So the question remains: What do we expect to come of such an experiment? (Location 1528)
The main point is that the treasure hunter can be harnessed to create and discover innovative concepts that otherwise would never exist. (Location 1543)
But—the silver lining is that we can still find hidden treasure in distant lands by departing without a destination in mind. So we shouldn’t mourn too deeply for the myth of the objective. (Location 1549)
First in this chapter we’ll make the general case for how the growing obsession with objectives can harm society. (Location 1578)
“The more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor.” (Location 1588)
And the end result isn’t students with meaningful knowledge and skills, but ones who are good at memorization and test-taking (Location 1593)
then a drive to increase objective performance is likely to produce deception, preventing the best possible result from being discovered. (Location 1598)
The economy could be stuck in a Chinese finger trap—a decrease might be needed to produce a much larger increase. (Location 1601)
The problem is that because GDP is such a simple measure, it doesn’t truly capture what really makes for a healthy economy. (Location 1605)
A more poisonous and extreme form of Campbell’s law is the problem of perverse incentives (Location 1608)
Ultimately, the number of venomous snakes in India actually increased (Location 1611)
And the same thing happened in Hanoi but with rats—leading not to fewer rats, but to rat farms [64]. (Location 1613)
paying workers for each fragment of dinosaur bone they find leads to smashing whole bones [66]; (Location 1616)
with bonuses for higher earnings leads to short-term profits but long-term disaster (Location 1617)
Instead, we’re examining those pockets where people are striving for innovation, discovery, and creativity. (Location 1635)
we’re interested in the distant shore on the far side of the mist-cloaked lake. (Location 1636)
But while we can find creativity and innovation almost everywhere, it’s helpful to explore particular instances to show how non-objective thinking can bring specific insights that could directly impact society. (Location 1638)
Instead it’s the more sinister problem of deception—the steps that lead to great invention aren’t likely to resemble great invention. (Location 1649)
lot of common assumptions, like the supposed danger of aimlessness, start looking more uncertain if you accept that the objective compass is a myth. (Location 1656)
And now, after seeing so many examples of great discoveries without explicit objectives, we have some idea why that is. With this perspective, success stories like that of Steve Jobs make perfect sense—in his own words: (Location 1658)
your strategy is as valid as taking the pre-med route with all the required classes. (Location 1665)
It’s just that in our culture “exploring” without a clear objective seems so flawed and misguided that not even Steve Jobs had the words to say it positively. (Location 1666)
The assumption is that increasing performance indicates progress towards the ambitious goal of near-perfect performance in the future. (Location 1686)
The problem is that the stepping stones to genuinely fantastic classroom-wide performance are likely completely unrelated to any common educational metric: Tests try to force everyone to better resemble the desired fantastic outcome (i.e. objective), which we’ve seen is often a formula for getting stuck. (Location 1698)
In the same article, DeMarco writes that although metrics allow for control, strict control is only important or appropriate when working on a project with little chance of major impact (Location 1710)
measurements are great when you have a modest goal, (Location 1712)
Surprisingly, accuracy doesn’t necessarily help increase performance in a pursuit driven by objectives. (Location 1727)
The important observation is that most of them look nothing like a skull. (Location 1733)
Unfortunately, we already know where this effort will lead. If we’re lucky, it may lead to some interesting places, but where it won’t lead is the Skull (Location 1743)
So accuracy does nothing for us when we have such a fundamentally broken compass. The result is just a better estimate of what remains a distraction. (Location 1750)
Tying the argument back to education, there’s little doubt that the problem of achieving a nation of exceptionally well-educated students is much more complex than the problem of breeding the Skull (Location 1754)
is the drive towards uniform standards (Location 1762)
So if students in one area are getting a worse education, or teachers in a particular school are not performing well, the problem will become clearly visible. (Location 1769)
Even if the standard is well-chosen and reflects current best practices, those current best practices will likely be rooted in the mythology of objectives. (Location 1787)
If we believe that measuring and encouraging progress through standardized tests is a broken compass in general, why focus on applying one particular version of that broken compass more uniformly? (Location 1790)
it also undermines making future discoveries. (Location 1797)
Enforcing uniform standards means converging to a single standard. (Location 1797)
As a result, it becomes more likely that future standards and tests will be limited to tweaks of the imposed status quo—because it is the only one applied and explored by teachers in their classrooms. (Location 1799)
foster deep learning and understanding of materials. The same story is reflected by the evolution of the Picbreeder Skull in Chap. 3 Without a diversity of users following their own varying ideals within the space of pictures, it is unlikely an image like the Skull would ever be uncovered. (Location 1806)
Or better yet, if they allowed researchers to follow their own intuitions about what might lead to the Skull (Location 1811)
When it comes to education, we’re essentially trying as a society to search for effective approaches to educating people. (Location 1815)
Perhaps the effort invested in assessment would be better spent on trying different ideas without so much emphasis on accurate measurement. (Location 1823)
Through this process society becomes a treasure hunter for teaching methods. (Location 1827)
In that sense Finland’s system follows more the spirit of non-objective search. Finland also is a world leader in education, significantly outperforming the US [76].) (Location 1830)
Sometimes being too objective is dangerous. (Location 1834)
And the treasure hunter is about collecting stepping stones. So when we’re dealing with societal efforts like education, we might be able to make good progress if we as a society help expose each other to potential stepping stones to new ideas. (Location 1844)
The members of the review panel then assess their colleague’s work on several measures, such as curriculum completeness, innovation, and student performance. (Location 1849)
From the perspective of avoiding deception, the appeal of this peer-driven assessment approach is that people and schools are not being compared to where we want them to be (which is the objective fallacy), but rather are assessed based on where they are (which is the philosophy of the treasure hunter). (Location 1855)
Great things are accomplished in the long run not because they were the objective, but because they were not. (Location 1868)
The next chapter continues in this spirit of liberation with an examination of objectives and their damaging effects in the pursuit of innovation. (Location 1869)
the first European explorer to sail into New York harbor, the price for fame was steep—he was killed and possibly eaten by natives while exploring an island in the Lesser Antilles (Location 1917)
Hundreds of years later, after narrowly missing out on being the first explorers to reach the South Pole, Robert Falcon Scott and his men ran out of food on their attempted return to civilization, tragically freezing to death [79]. (Location 1919)
There’s something enthralling about the raw danger and untapped potential of the unexplored. (Location 1922)
But those who penetrated the unknown (and returned still breathing) helped to expand the horizons of human knowledge. (Location 1923)
In fact, new ideas and technologies have the power to reshape our world and society entirely. (Location 1927)
Because scientific progress is currently so prominent in our culture, we’ll begin with it. (Location 1929)
Sometimes we forget how quickly the world is being transformed through scientific progress. (Location 1930)
although the 300 year interval between them is a only a wink within the 40,000 years of human history. (Location 1933)
“Well, for two years after World War II, your grandpa performed relief work in Paraguay, helping to build up sanitation. We corresponded by letters – mostly handwritten – and it would often take two weeks for one of his to reach me.” (Location 1934)
But while scientists aren’t eaten by cannibals and don’t typically freeze to death when their experiments fail, the truths they uncover still hold real potential to turn the world upside-down. (Location 1940)
Improvements in technology shrink the world, make our lives easier, and cure once-fatal diseases. (Location 1943)
But such experiments are often expensive. (Location 1946)
This means that scientists seeking to explore must first find funding for their projects. (Location 1947)
One main difference between science and education is that we would never be happy with a school that is a complete and utter failure, but in science it is commonplace and expected for individual projects to fail. (Location 1950)
The problem is that there exist more scientific projects than can possibly be paid for, which means that risk and reward must be carefully balanced. (Location 1952)
But the lesson from Picbreeder is that the most interesting discoveries can’t be predicted in advance. (Location 1957)
it’s an area where exploration is essential but individual failures aren’t as dangerous. (Location 1960)
Groundbreaking ideas are often also risky. So while some may succeed, many others will fail. (Location 1965)
must take some risks to have the hope of furthering the most innovative ideas. (Location 1968)
But there’s trouble lurking behind the apparent common sense—the main effect of the system is to reward consensus (Location 1974)
the reason so many images have been discovered is that users don’t agree on which images are better. (Location 1977)
The problem is that when individuals with opposing preferences are forced to vote, the winner often represents no one’s ideals (which perhaps explains the nearly-universal frustration people have with politics). (Location 1981)
people don’t agree on what the most interesting stepping stones are. (Location 1983)
compromise between opposing stepping stones. (Location 1984)
Instead of allowing each person to discover their own chains of stepping stones, the system squashes a diversity of opinions into a generic average. (Location 1988)
It’s possible that anti-consensus may be more interesting than bland agreement. (Location 1989)
After all, attracting a unanimous vote in science could be a sign of nothing more than echoing the status quo. (Location 1990)
On the other hand, an interesting idea is likely to split votes. (Location 1992)
unknown are questions whose answers remain uncertain. (Location 1993)
That’s why the opinions of experts should diverge in such uncharted territory. (Location 1993)
Just think, which project is likely more revolutionary, one that receives excellent, excellent, poor, poor, or one that receives excellent, excellent, excellent, excellent (Location 1995)
Darwin’s theory of evolution was dismissed by many experts when it was first introduced [83]—a good sign! (Location 2000)
For all these reasons, some of our resources should go towards rewarding disagreement rather than consensus. (Location 2003)
The agreed path is an objective choice because people agree with its destination (Location 2005)
Only in non-objective search is the idea of unifying on a single destination discouraged. Only then can being interesting attract resources and funding. (Location 2009)
Science is among the greatest explorations of humankind—that we would reward primarily consensus in deciding where to step next is as stifling to discovery here as in any creative endeavor. (Location 2011)
Science needs to be a treasure hunter and a stepping stone collector. (Location 2014)
but not for creative exploration. (Location 2015)
Rather, the point is that disunity among research groups and within science as a whole can actually drive progress. In this way, disunity’s power can help us better structure scientific and other creative endeavors. (Location 2021)
For example, if you believe in objectives, you might also think that the structure of scientific progress is predictable. (Location 2024)
So grand projects to cure particular diseases through direct focused research may not always be the best approach—if you actually want to cure that disease. (Location 2027)
as a byproduct it may enable unintended breakthroughs in seemingly-unrelated areas. In fact, (Location 2029)
But this was an uncertain pronouncement that only proved possible because it was just at the edge of current technological possibility (that is, it was one stepping stone away when it was conceived). (Location 2040)
similar line of thought is that there is predictable structure in how scientific projects will impact the world. (Location 2050)
it’s possible to optimize towards projects with breakthrough impact by consistently funding projects that currently seem to offer the most potential for impact. (Location 2051)
As a result, predicting impacts is not always possible, and the attempt to do so discounts serendipity’s important role. But (Location 2066)
Ultimately the aims for science as a whole are to uncover deep and transformative truths—but it may not matter at all if any particular project is transformative. (Location 2069)
As a system it may discover a robot able to reach the goal, but it only works when robots are not graded by their individual ability to reach the goal. (Location 2072)
The stepping stones to the most important scientific discoveries may not themselves seem important, and the stepping stones to the most transformative technologies may reveal no hint of transformation themselves. (Location 2076)
government wishing to promote only research agendas that it deems important at the time, or that provide clear short-term benefit to the country. (Location 2088)
Who knows where the next great commercializable technology will come from? While this outlook may sound pessimistic, it makes the world of science a much more interesting place. (Location 2107)
it’s just that we don’t know what exactly will lead to important discoveries. Like the surprising value of disunity in science, there may also be wisdom in funding seemingly unimportant—though clearly interesting—scientific experiments. (Location 2110)
As usual, you might ask how we can so smugly idolize stepping stones without knowing where they lead. (Location 2113)
But given that objectives are often a false compass in any case, maybe the idea of success shouldn’t always be the focus. (Location 2123)
And as we’ve seen, ambitious objectives almost never work out. (Location 2129)
because those are the only ones that can realistically be assured—the ones that are but a single stepping stone away. (Location 2131)
While some level of risk is necessary for progress, at the same time those paying the bills generally don’t want so much risk that resources are simply being wasted on crackpot projects. (Location 2134)
On the other hand, in areas likes business investing, the hope is for more short-term payoff. (Location 2136)
that the only good objectives are on nearby stepping stones. (Location 2138)
So the innovative solution must be discovered before approaching the investor. (Location 2139)
Realistic objectives, which tend to be the province of investing, tend to be exactly those that are one stepping stone away. (Location 2152)
But that doesn’t mean that business can’t be innovative—an innovative business idea reveals a nearby stepping stone that we didn’t previously realize was there. (Location 2154)
Incremental steps into the blind alley of the unforeseen can reap huge rewards. (Location 2162)
The stepping stones are yet to be discovered. In the end, the businessperson tends to look for nearby stepping stones before seeking funding, while the scientist ideally requests funding to follow a hunch that an interesting stepping stone is nearby. (Location 2164)
In the long run, it’s the accumulation of stepping stones that leads to the greatest innovations. (Location 2167)
But as with all great processes of discovery, the revolutions are rarely the objectives of the stepping stones that lead to them. (Location 2168)
Investors recognize this principle even if it’s not explicitly stated. In short, if you’re looking to invest in visionaries, find those who wander in nearby shadows. (Location 2169)
For artists and designers, the concept behind an idea is often more important than the goal (Location 2172)
The roof has to keep the rain out. (Location 2175)
It turns out there’s an interesting parallel here between these kinds of “objectives” and the constraints on organisms in natural evolution. (Location 2176)
are more like constraints on creativity than typical objectives that are pursued for their own sake. (Location 2179)
So even though the constraints do matter, the overall search is still driving towards uncharted spaces. (And if evolution interests you, this topic is discussed in greater detail in the first case study after the book’s conclusion.) (Location 2181)
the greatest explorations have no goals. (Location 2200)
The result is beautiful and thought-provoking. Such artwork is not only valuable for the way it suggests past violence and mellowing, but also for new related art forms it may inspire in the future. (Location 2201)
objectives run almost everything. They’re not always wrong, and even when they are wrong replacing them is no simple matter. (Location 2209)
Before we take this final step, it’s important to recall that all along we’ve been talking about ambitious objectives. (Location 2271)
Houses should still be constructed according to the blueprint, software should still be designed to meet its specification, and there’s no harm in following the recipe when you cook your next dinner. (Location 2273)
On the contrary, it’s critical to realize that just because you don’t start out searching for what you ultimately find does not mean that you are necessarily searching blindly. (Location 2287)
Perhaps that’s why those who do succeed brilliantly are often cloaked in mystery and rewarded with devoted respect. (Location 2301)
The answer is to become the treasure hunter (Location 2306)
countless treasures buried deep in unmarked locations. All these treasures are worth finding even though none of them may be something you’re seeking in particular. (Location 2307)
For that to happen, you have to learn to search for clues. (Location 2310)
The point is that novelty search is a good example of a treasure hunter—it pushes towards new discoveries without any overarching goal. (Location 2317)
To put it another way, our tastes are more refined than only craving novelty and nothing else. (Location 2320)
one clue to it might be interestingness (Location 2322)
Ideas do become less interesting as they become less novel. (Location 2328)
Even just accepting that this treasure-hunting approach makes sense is important. It opposes so many of the messages we receive today from our culture that say you need an objective to achieve anything worthwhile. (Location 2337)
When you’re standing on the edge of possibility looking out over the unknown, objectives become false beacons, but interestingness is different. (Location 2350)
Contrary to popular belief, great inventors don’t peer into the distant future. (Location 2353)
but a true innovator looks nearby for the next stepping stone. (Location 2354)
The successful inventor asks where we can get from here rather than how we can get there (Location 2354)
they concentrate on the edge of what’s possible today (Location 2356)
This set contains all of our sciences, our technologies, and our arts. (Location 2357)
There are delicate moments in history when our set of capabilities steps over an invisible line and exciting new possibilities snap into view. (Location 2359)
But within months millions were sold. As the leader of Apple, Steve Jobs had noticed that both society and technology had progressed to a point where a commercially viable tablet became possible (Location 2373)
objective-driven companies often dwindle without a product for years or decades. (Location 2381)
the success of Apple illustrates that deciding where to go based on where you are is often wiser than deciding where to go based on where you want to be (Location 2383)
None can transform the future into the present. (Location 2385)
For example, flying machines were an ambitious objective for centuries before the Wright brothers achieved flight. (Location 2387)
It turns out that for them bicycles were stepping stones to flying machines. (Location 2392)
You could say that they saw in the present an echo of the future, rather than trying to drag a preconceived vision of the future into the present. (Location 2392)
Letting go of objectives is also difficult because it means letting go of the idea that there’s a right path (Location 2396)
When there’s no destination there can’t be a right path (Location 2399)
we should judge our projects for their potential to spawn more projects. (Location 2400)
then the only important thing about a stepping stone is that it leads to more stepping stones, period. (Location 2401)
The worst stepping stone is one that leads nowhere beyond itself, no matter how nice it may feel to stand upon it for the moment. (Location 2402)
The more stepping stones we find, the more opportunities there are to depart to somewhere greater. (Location 2404)
So if you’re wondering how to escape the myth of the objective, just do things because they’re interesting. (Location 2415)
If you don’t have a clear objective, then you can’t be wrong, because wherever you end up is okay. (Location 2416)
A great achievement is one that leads to more great achievements. (Location 2417)
The prevailing philosophy is to chain exploration, to enslave it to our objectives. (Location 2426)
To achieve our highest goals, we must be willing to abandon them (Location 2431)
If that isn’t mind-boggling enough, modern science tells us that the unguided process of natural evolution is what created and organized this massive team of cells that make us who we are. (Location 2455)
But evolution is unique because it continually makes new from old, a kind of perpetual recycling machine. (Location 2464)
it can completely recreate an entire ecosystem. The more you study evolution, the more incredible it seems. (Location 2465)
It turns out that a slightly-broken copying machine can lead to the appearance of intelligent design. Imagine (Location 2469)
The twist comes if the self-copies are slightly imperfect (Location 2472)
The result of this kind of imperfect copying is that you’ll end up with a variety of different machines, because each machine will make something slightly different than the last. (Location 2472)
So in the end, we could say that Darwin discovered that the process of evolution is itself the mysterious creature-maker. (Location 2481)
but on how to view evolution, and what ingredients of evolution cause it to be so creative and powerful. In other words, different interpretations can disagree on which forces are most responsible for driving evolution. In particular, a key question is, how important is natural selection? (Location 2487)
Gould thought that chance and history in evolution were perhaps more important than natural selection, (Location 2495)
natural selection was more important than chance (Location 2496)
they disagreed over its interpretation (Location 2497)
optimizing survival and reproduction through natural selection, or as a more divergent process that happens to accumulate different forms of life? (Location 2499)
And more than that, the phrase does a good job of capturing one driving force behind evolution: Creatures that are better adapted to their particular niche will tend to out-reproduce their less-adapted peers. (Location 2511)
The whole notion of a competition for survival suggests a competitive search for ever-more-fit organisms, which leads ultimately to the objective of “the fittest,” or the best reproducer. (Location 2518)
increasing improvement. For similar reasons, “survival of the fittest” is also responsible for the popular view of natural evolution as a bloody competition among organisms. (Location 2522)
measure up objectively. As Tennyson put it, “Nature, red in tooth and claw.” Focusing on competition in evolution leads to the outlook that nature always strives to improve itself; but—here is the interesting question—to what end? (Location 2524)
still believe that evolution is progressive, moving towards some sort of objective perfection, a kind of search for the über-organism. (Location 2526)
evolution objectively leaves room for different interpretations. The most naive is to imagine that evolution is set up explicitly to search for humans as its objective. (Location 2529)
If every organism from the very first cell on Earth was selected for its similarity to humans, humans would never have evolved. (Location 2533)
The stepping stones to intelligence aren’t intelligent themselves. (Location 2539)
In fact, evolution is just one of many major scientific discoveries that took humans off our pedestal as intrinsically special. (Location 2544)
The real point here is that evolution doesn’t treat us humans specially. There’s no hint in evolutionary theory that we’re elevated over any other creatures. (Location 2549)
As we briefly discussed in Chap. 4, most people with a scientific background would say that survival and reproduction is evolution’s objective. (Location 2553)
acknowledges an evolutionary force driving towards greater survival ability: natural selection. (Location 2556)
Importantly, natural selection provides an explanation for why organisms are so well adapted to their environments: (Location 2559)
Selection is constantly trying to improve how an organism fits with its surroundings, like a puzzlemaker trying to whittle a piece’s edges to better fit with its neighbors. (Location 2560)
In short, evolution’s objective is to keep increasing fitness. This view of evolution fits neatly within the objective paradigm and also makes intuitive sense. (Location 2565)
More deeply, if we view evolution not only as trying to achieve survival and reproduction, but as searching for the organism that survives and reproduces the best (that is, the most fit organism), then we’re still left with contradictions. (Location 2579)
Evolution begins to look more like the unfolding processes of Picbreeder or human innovation than it does objective-based achievement. In this way of thinking (as noted in Chap. 4), when we look at the most inspiring achievements of evolution, such as photosynthesis, flight, or human-level intelligence, it’s interesting that none were ever its objective. (Location 2597)
Fundamentally, the objective-based explanation dangerously ignores that the most interesting products of evolution become just a side-effect of the drive towards the objective (high fitness). (Location 2601)
A different way to think about survive and reproduce is as a constraint (Location 2608)
reproduce on Gentle Earth as well, which means that the same lineages would be explored. (Location 2629)
But the positive side of this tradeoff would be that many interesting organisms that just happened to be unlucky and did not survive on real Earth would be given a second chance on Gentle Earth. (Location 2632)
The point of this thought experiment is to illustrate that selection is not really a creative force: It focuses and optimizes, and ultimately restricts exploration. Selection’s grand achievement is only to ensure that resources (which are finite on normal Earth) are spent on organisms that can reproduce on their own. And this kind of restriction isn’t the same as exploration. (Location 2637)
So “survive and reproduce” can be viewed as constraining the stepping stones that evolution explores further, and not an objective in itself. (Location 2640)
The problem with competition as a general explanation for creativity is that it usually drives towards everything converging to the best. (Location 2647)
because it drives towards divergence (Location 2651)
towards a multitude of varying solutions to life’s problems. (Location 2652)
Interestingly, it’s not by competition that natural evolution leads to diversity, but often by avoiding competition. (Location 2654)
Because it then becomes the first organism to live in this new way, the competition is less intense for the lucky newcomer and it can reproduce more easily. (Location 2655)
Nature is always seeking to expand. (Location 2660)
Additionally, a new niche often makes founding newer niches possible, and such newer niches often lead to even newer ones. (Location 2661)
The waste products of one organism may become the food of another. (Location 2663)
On the other hand, mutations that don’t much affect fitness have more uncertain fates, because natural selection will be mostly indifferent towards them. These kinds of neutral genetic changes evolve not by natural selection, but according to aimless genetic drift (Location 2674)
Another force that allows serendipity to act in evolution is exaptation, which occurs when a feature of an organism evolved for one function proves useful in an entirely different context. (Location 2682)
An intriguing pattern that we observe in many different kinds of search processes is that interesting discoveries for one purpose often to prove useful in the future in a variety of unexpected ways. (Location 2690)
but rather as escape from competition. (Location 2692)
In other words, when we view evolution as a search for optimality or efficiency, we’re equating it with more boring searches, (Location 2698)
definitive best path to the library can be found—and (Location 2700)
viewing evolution as a competitive search for optimal organisms suggests that evolution itself will someday converge to an ultimate uber-organism. But (Location 2701)
important to avoid the inference that contrasting interpretations must be incompatible. (Location 2703)
The formula for the universal law of gravity is interesting, but the explanation behind it is more profound: Matter attracts matter, a simple and elegant truth. (Location 2707)
Imagine if we could bottle up the power of evolution and apply it wherever we like, to cars as easily as to robotic pets. (Location 2710)
An important tool in understanding natural processes is abstraction (Location 2714)
good place to begin abstracting is with the perspective that “survive and reproduce” is a constraint rather than an objective. (Location 2723)
while an organism that doesn’t reproduce will act like an evolutionary dead end. (Location 2725)
all successful organisms in evolution now become equals. (Location 2726)
we only care whether it does or does not survive and reproduce. (Location 2727)
interesting to look at evolution through the perspective of this bottleneck: It leads to the idea that nature has found an incredible diversity of methods to start from one cell to do nothing more than ultimately produce another. What happens in between is life. (Location 2734)
In fact there is a hint of Rube Goldberg machines in the awesome contortions taken by life to start from one cell only finally to produce another of the same type. (Location 2736)
But the signature of Rube Goldberg machines is that a simple task is achieved in an overly complicated way. (Location 2745)
And now, years later, you stand as a trillion-celled tower of reproductive inefficiency. (Location 2752)
There are many ways to meet the minimal criteria of life, whether a beetle, a bird, a buffalo, a barnacle, or a businessman. (Location 2759)
Instead of competition, in this view the engine behind evolution’s creativity is searching for many ways to do the same thing. (Location 2761)
In this way, the novelty search algorithm described in Chap. 5 is like natural evolution with the accelerator pressed to the floor—always rewarding the most novel with more offspring. (Location 2772)
when the simple ways to survive are exhausted, (Location 2775)
Seen this way, evolution is a special kind of non-objective search: a minimal criteria search. (Location 2777)
Evolution is just accumulating all the different ways to survive and reproduce. (Location 2779)
But they aren’t discovered because they are better or more optimal—they are simply the stepping stones reachable from where the search last stood. (Location 2781)
However, competition may be the least interesting of the evolutionary forces because it tends to diminish diversity. (Location 2792)
It’s more a honing force that optimizes creatures within a particular niche, or in limited ways across niches, like when gazelles adapt to run from lions. (Location 2794)
So for evolution to produce impressive products it’s critical that competition doesn’t overpower the drive towards creativity. (Location 2799)
including the drive to escape competition by founding new niches. (Location 2801)
In other words, each organism is not in competition with every other organism. (Location 2807)
A human being competing globally with bacteria on the yardstick of reproductive fitness would not measure up (the bacteria reproduce more often and more quickly). (Location 2812)
but only produces limited effects between niches (Location 2814)
encourages the founding of new niches to escape competition. (Location 2816)
Instead, the idea is to recast evolution in a different light in the hope of deeper understanding. (Location 2821)
In fact, we also formalized this refined abstraction as another algorithm called novelty search with local competition (Location 2821)
A key insight from thinking non-objectively in this chapter is that although evolution can be seen as a competition, out-competing other creatures on the “objective” of surviving and reproducing is less important than escaping from competition to form new niches. (Location 2827)
It’s how stepping stones accumulate and the potential of the system as a whole ratchets upwards. (Location 2829)
Before the Cambrian explosion, life was limited to small simple niches, similarly to how Picbreeder begins with simple pictures, or the behaviors of maze navigators in novelty search are at first simple, or human innovation starts with the wheel and not the internal combustion engine. (Location 2848)
From this point of view, we can see the fossil record itself as the original story of serendipity, written not with words but through dead organisms immortalized in rock, a billion-year testament to the rich power of the non-objective treasure hunter. (Location 2853)
more about exploration than about bloody competition; more like constructing novel Rube Goldberg machines than aspiring towards perfection. (Location 2858)
At its heart, science is the search for knowledge. (Location 2898)
discoveries of today are the stepping stones to the unimaginable technologies of the future. (Location 2900)
scientists specialize in one or two smaller areas of scientific knowledge, called disciplines (Location 2906)
That way, the scientists in the community can keep up to date without reading every paper ever written. (Location 2918)
The ideas that are accepted for publication in journals and conferences are important because they shape the direction of future research—they’re the stepping stones that tend to be explored further. So the interesting (and significant) problem is how best to decide which ideas should be published and which should be discarded. The most common solution (Location 2924)
So if a community isn’t cautious, it can stagnate as the gatekeepers play favorites and hold back progress. (Location 2942)
After all, objectives are a favorite tool of society’s gatekeepers—the people in power who decide where to invest resources and what to ignore. (Location 2963)
actually conducts the search for AI algorithms. As we’ve discussed throughout this book, the key idea behind search is to follow a gradient: A bread-crumb trail of some kind, a path of increasing intensity. (Location 3013)
from bad to good on some performance measure. In novelty search, it’s the gradient of novelty. (Location 3015)
A technical term for a rule of thumb that guides search is a heuristic (Location 3024)
The first, which we’ll call the experimentalist heuristic, follows the rule of thumb that an algorithm’s promise is given by how well it performs. (Location 3025)
better than existing algorithms in benchmark tasks. (Location 3027)
When scientists benchmark AI, the idea is that a smarter algorithm should solve a problem faster than a weaker algorithm. (Location 3029)
main gradient in AI research is the theoretical heuristic. This heuristic suggests that algorithms are better if they can be proven to have desirable properties. (Location 3030)
If an algorithm doesn’t improve performance or doesn’t come with guarantees, it will be hard for it to pass the gatekeepers—and so it will probably never become known to the AI community as a whole. (Location 3035)
Recall that AI is a search for algorithms that search—a meta-search. (Location 3059)
These kinds of sweeping rejections prune away a large part of the search space, the space of all AI algorithms. (Location 3083)
The comparison only distracts us from what makes Weird interesting in the first place—which may be a better place to focus. (Location 3087)
But benchmarks do little to illuminate why one research direction is more or less interesting than another. (Location 3098)
The experimentalist heuristic is driven by such a simple objective that few AI researchers would actually employ an algorithm based on such a naïve heuristic today. (Location 3108)
What this scenario shows is that the experimentalist heuristic often asks us to compare the performance of two distant points in a giant search space. (Location 3121)
While researchers blaze trails toward future innovation, the practitioner wants to solve real-world problems right now. (Location 3130)
The practitioner is more an AI customer than an AI inventor. (Location 3132)
If OldReliable is 5 % better in a benchmark that resembles the problem the practitioner is trying to solve, then she should probably use OldReliable over Weird (Location 3135)
So what’s interesting for the practitioner—who is frozen at one moment in time—isn’t the right compass for an innovator looking towards the future. (Location 3141)
Judging by performance is a good idea for practitioners but is shaky at best for researchers—because of deception. (Location 3143)
Its core idea is that algorithms with more proven guarantees (Location 3145)
Perhaps surprisingly, it suffers from the same defect as the experimentalist heuristic when it’s used to guide the search for AI algorithms. (Location 3152)
If we can prove that a particular algorithm will succeed under certain conditions, the resulting guarantee can’t be revoked. So in the phrase “theoretical heuristic,” the uncertainty of heuristics may seem to clash with the invincibility of theorems. (Location 3158)
The word “theoretical” applies to an individual algorithm by asking whether there are good guarantees for that algorithm. (Location 3161)
“heuristic” applies to the search through AI algorithms, suggesting that algorithms with lots of proven guarantees may often make good stepping stones. (Location 3162)
No, overall we should be focusing on exploring the space of all AI algorithms and uncovering promising stepping stones. (Location 3166)
how theorems are used to guide exploration through the space of AI algorithms. (Location 3167)
But a theorem about one particular algorithm doesn’t promise anything about future algorithms that may be invented by researchers searching the space of algorithms. (Location 3174)
It means that the community becomes restricted only to those algorithms that honor the same growing set of assumptions, blinding the meta-search to every path forward that breaks the assumptions. (Location 3180)
So to believe that algorithm OldReliable is a promising stepping stone because of its theorem only makes sense if there’s something about this particular theorem that suggests the algorithm may lead to other promising algorithms. (Location 3186)
The assumption is that the more promising theorems you have about an algorithm, the closer you are to the objectives of (Location 3193)
Putting too much faith in the theoretical heuristic just delays (or prevents) exploring the ideas that are raised by Weird (Location 3198)
While theorems are interesting, it’s not clear that performance guarantees (or any other guarantees) are the right kind of information to guide search on their own—especially through a space as vast and complicated as the great room of all AI algorithms. (Location 3205)
Their measurements decide what ideas are good enough to share. (Location 3224)
A good algorithm isn’t one that performs better, but one that leads us to think of other algorithms. (Location 3234)
drop a new idea for worse performance alone? The community is engaged in a search after all, and a search’s function is to find things. The experimentalist and theoretical heuristics are about finding fewer things because they prune out many interesting algorithms. (Location 3241)
The problem is that performance-squeezing algorithms just aren’t good stepping stones. (Location 3248)
are those that lay the foundation for future trailblazers. (Location 3249)
Science isn’t a track meet. This competition often distracts from the real purpose of the AI community: (Location 3251)
Can the AI community run some kind of “novelty search” over AI algorithms? Yes, something like that is possible, but it would require an overhaul to put into practice. (Location 3262)
Isn’t it strange to say that we must protect ourselves from potential cranks by following rigid objective rules? (Location 3274)
The real issue is that the experimentalist and theoretical heuristics, as with many objective measures in society, become excuses not to use our brains, even for the experts. (Location 3281)
And if it performs better, they can simply demand that its creators compare it to many more algorithms before finally giving it the green light. (Location 3283)
This means that they’re forced to think about the core idea and decide whether it’s interesting (Location 3299)
The greatest ideas are unlike those that came before them. (Location 3306)
Every stepping stone has a unique story of discovery. (Location 3307)
And every human being who found one has a story too. (Location 3307)
real expert in any field can think with an open mind and doesn’t need a rigid heuristic to make a decision. (Location 3308)
it’s wrong to assume that the stepping stones will resemble where they ultimately (Location 3313)