107 - Chasing Greatness: Shoot for the Moon but Only If You Can Train the Monkey First (Part 2 of 2)

06-10-2023

Career DesignPersonal Development

Henry VIII ruled over England with an iron fist for thirty-six years. For much of the second half of his reign, their king’s foul mood concerned Parliament. Word went that Henry VIII’s famed temper was down to leg ulcers. Painful and untreatable, they irked the King of England to tyranny. A sour Henry coped by eating and drinking. Being a monarch, he probably had the pick of the spoils. To top it off, his leg didn’t let him burn those calories off. Henry ballooned. It probably didn’t help his state of mind that he didn’t have a male heir to the throne and he kept picking wives to help him produce one.

>>>>> gd2md-html alert: inline image link here (to images/image53.png). Store image on your image server and adjust path/filename/extension if necessary.
(Back to top)(Next alert)
>>>>>

alt_text

Pic: Iconic portrait of Henry VIII from 1536 by Hans Holbein the Younger. Labeled a work of propaganda by some for the brushing over of the leg injuries incurred by Henry earlier in the year.

Now imagine if Henry could walk into a clinic as you and I do today and get his leg ulcers fixed for good with laser surgery. Or the royal GP could prescribe him a course of antibiotics for his infections. The marvels of modern medicine that we don’t even think about for a moment today were in the 1500s the ones we could not imagine even if we tried to.

What were the stepping stones between leg ulcers and gene editing? Could we have predicted the greatest advances that got us from Henry VIII to Jennifer Doudna? Could we have set them as ambitious objectives and made a plan to accomplish them?

In the first half of this essay, I offered a no, chiming in on the insightfully written Why Greatness Cannot be Planned by AI researchers Kenneth Stanley and Joel Lehman. For the simple reason that the future is recombinant. The ingredients for it are still in the making. We don’t know what we’ll need, much less for what.

Yet, it is also true that freethinking ambition needs patrons and deep pockets. Even science, the purest search for knowledge, cannot be unlocked without grant proposals to funders on which scientists burn months justifying line items for expense.

Progress needs measurement. Ideas need objectives. Gatekeepers need assurance. And even if somehow such wrinkles were to be smoothed out—maybe we had a Medici family as our sympathizers—there’s a snag in the fabric that’s hard to ignore. It is that we are socially conditioned to be uncomfortable without a yardstick for measurement. We are so used to judging and being judged for our efforts that an open-ended search is disorienting, even pointless, to many of us.

So how do we pursue greatness if we don’t know and can’t see the exact path to it? What do we do with never-done-before ideas? What is another way to develop world-changing ideas if we can’t plan them as projects on the calendar?

Here are the highlights from this week’s piece:

  1. Even the best lack the good sense to quit when the time is right. So, they define the breakthrough point in advance and make a precommitment to quit if you fall short.
  2. The breakthrough point is both a state and a time. You don’t have unlimited time and you cannot compromise on the state.
  3. Chasing greatness is not a once-in-lifetime chance. It is a repeating option and it demands that you’re ready to walk away if you get stuck.
  4. A moonshot factory stands to be reduced to ashes after every failed project. But the ashes are nothing but moonshot compost from which sprout new ideas.
  5. Uncertainty is a constant companion of those with deep ambition. The worst thing to do while chasing greatness is to pretend to be more sure than you are.

The case for X

X, Alphabet’s innovation lab and Google’s sister company, is called a moonshot factory. Teams at X are paid to pursue designed-to-fail ideas (like turn seawater into fuel) in their search for radical impact. They are also expected to be willing to give up the pursuit if—and they decide the if early on—they fall short of the breakthrough point.

The breakthrough point is both a state and a time, and falling short implies that what you’re building is not in state A by time B, where both A and B are pre-defined. Before the start of any moonshot project, the project leader makes a precommitment to quit if the team falls short of the breakthrough point.

In Astro Teller, X’ers are led by someone whose responsibility as Captain of Moonshots is to make the world a better place. Teller outlines what they do at X in this pithy quote:

You can’t pre-business-plan a moonshot any more than you can paint a masterpiece by color-by-numbers given to you by committee.

This kind of roomy yet uncomfortable scale could lead to chaos. Only, as the portfolio for X bears out, it is our most proven vehicle for a method that nurtures creative madness.

Let’s pick apart the machinery that powers the world’s most riveting moonshots.

Train the monkey or shut shop

A few weeks ago in this newsletter I published this passage as part of a piece on the value of solving the hardest thing first when on world-changing pursuits:

‘Let’s say you’re trying to teach a monkey how to recite Shakespeare while on a pedestal,’ says Astro Teller who heads X, Alphabet’s moonshot factory. ‘How should you allocate your time and money between training the monkey and building the pedestal?’

‘The right answer,’ Teller continues, ‘of course, is to spend zero time thinking about the pedestal. But I bet at least a couple of people will rush off and start building a really great pedestal first. Why? Because at some point the boss is going to pop by and ask for a status update — and you want to be able to show off something other than a long list of reasons why teaching a monkey to talk is really, really hard.’

In Teller’s memorable metaphor, the monkey is the bottleneck and training it is the breakthrough for the project. An ambitious venture will present multiple challenges but one above all else has the power to make or break the project. Find it, solve it, power on. Or quit.

What sort of a claim to greatness is quitting? What can you achieve by giving up?

Seems like a lot. In over twelve years of operation, X has been an incubator of hundreds of moonshots, a number of which like Waymo (self-driving cars), Verily (ML-powered disease management), and Mineral (climate-resilient food system) have graduated to independent businesses.

>>>>> gd2md-html alert: inline image link here (to images/image54.jpg). Store image on your image server and adjust path/filename/extension if necessary.
(Back to top)(Next alert)
>>>>>

alt_text

A three-step recipe to quit something you dearly love (and do something great)

Undergirding the monkeys-and-pedestals thinking is the belief that the pursuit of greatness is a repeating option. That you will have more crazy ideas to go for than you have time, money, or energy to throw at. By giving up something that is undoable, at least in the prevailing state of the world, you free yourself up to pursue something else as big, or bigger. This is the opportunity-cost mindset that the moonshot takers at X have cultivated.

One gets an understanding of how X’ers do this by studying the X ways of working. Among the tenets, there are three that stand out for the purpose of our discussion. Of these we already know the first one:

1️⃣ Tackle the monkey first.

Figuring out how to tackle the monkey will not come as a step-by-step plan. Here are two more tenets to help manage the process of pursuing far-out ideas.

2️⃣ Focus on the outcome, not the output.

Anyone who prides themselves on their problem-solving knows that it can be hard to give up a problem they’ve put their heart and soul into. But just because something is hard to do doesn’t mean it is worth doing. There’s no point in turning seawater into fuel if it cannot be done at a cost that makes it a viable alternative to fossil fuels.

X’ers stay the course by being passionately dispassionate. They keep their skepticism close and ‘learn to look dispassionately at the results, so we’re ready to change course or walk away when that is the right thing to do.’

But how easy is it to be passionately dispassionate when sunk costs accrue, commitment escalates, and the siren song of a just-round-the-corner breakthrough rings loud. This is where their relationship with failure comes to the fore. The third and final ingredient in the recipe shines light on this.

3️⃣ Embrace ~~failure ~~learning.

Eugénie Rives, leader of Project Move, a robotics moonshot, says:

We were developing robotics and AI solutions for the logistics industry and the next day I would have to tell my team that we were shutting down the project. We were weeks away from unveiling ourselves to the public and deploying to our first customers. But that afternoon, I’d made the incredibly difficult decision with X leaders to kill our moonshot.

Weeks away from deployment of a multi-year project, Rives decided that the monkey was intractable. She continues:

The process of winding down my own project was the most painful in my career, yet, I’ve never learned so much. It made me realize how important it is to have intellectual honesty when you’re working towards bigger goals. It made me a stronger and more resilient leader. And it helped me understand just how fertile the ashes of killed projects can be.

Moonshot Composting - the shared soil in the path to greatness

Moonshot Composting is something you probably haven’t heard about.

Nonetheless, the idea behind it would be deeply familiar. We all have failed and used the remains of the experience to make something useful later in life. How do we do that when the stakes are high?

That is exactly what is going to happen in a small village in Tamil Nadu in the coming months. The unsuspecting villagers--those least able to pay--are going to receive high-speed Internet. All because of a stratospheric failure. Literally.

In 2012, a team at X asked something crazy:

Could we connect the remotest parts of the world with high-speed Internet delivered through stratospheric Internet-beaming balloons?

That project, called Loon, failed. In disbanding Project Loon, Teller wrote:

_We hope that Loon is a _stepping stone to future technologies and businesses that can fill in blank spots on the globe’s map of connectivity. To accelerate that, we’ll be exploring options to take some of Loon’s technology forward. We want to share what we’ve learned and help creative innovators find each other — whether they live amidst the telcos, mobile network operators, city and country governments, NGOs or technology companies.

_Stepping stone. _It’s that same phrase that Stanley and Lehman, the two authors of the first theory, used.

World-changing ideas don’t always work out. Scratch that. _Most _world-changing ideas fail. But their failure can be serendipitous. Their residue is rich fertilizer for the soil of creativity for the next moonshot. Failure is learning.

While working on Loon, the leader of the team, Mahesh Krishnaswamy, had an epiphany: Could we deliver on the ambition of that same moonshot not in the skies but on the ground?

In 2016, Project Taara kicked off to connect those least able to pay to the world by laser beams. And one of the earliest pilots of that moonshot is now happening in the village where Krishnaswamy used to spend his summers.

>>>>> gd2md-html alert: inline image link here (to images/image55.jpg). Store image on your image server and adjust path/filename/extension if necessary.
(Back to top)(Next alert)
>>>>>

alt_text

If it works out, the technology will travel to congested cities because ground-based Internet infra of fiber optics is cumbersome and expensive. All because of a whopping failure.

What’s wrong with how most corporations run moonshots?

Project Loon failed. From its ashes took root Project Taara. And now it’s going to change the world. Krishnaswamy and his team didn’t see this coming but they also didn’t give up on learning from the experience. They found a way to compost failure. Look beyond this singular instance of exaptation, and you may see that the ruins of tens of abandoned projects at X carry the lushest insights gathered over years. Such an ecosystem is uncommon in corporate life.

When shooting for the stars, it helps to look beyond pass-fail. Annie Duke, decision strategist and author, suggests in her book _Quit _that we do the mental math differently. Not in terms of how short of the target you ended up but how far you landed from your starting point.

In most organizations trying to take a new product/service idea from 0 to 1, the pass-fail nature of goals is the yardstick in use. Executives, especially if they’ve been successful before, have a hard time sufficiently accounting for uncertainty in planning. Because executives don’t sufficiently appreciate uncertainty and account for it, they are not alert to learning because of it. Instead they focus on metrics and annual projections, and ignore the dancing monkey that refuses a word of instruction. And because they don’t learn, or they don’t learn quickly enough, they keep building pedestals until the clock runs out.

Failed projects demand a wave of restructuring. Teams change, personnel change, and a new project kicks off. The remnants of learning from past projects are lost, or even discarded in disregard of Chesterton’s fence. The new team begins, or chooses to begin, from a blank page. There’s little ‘moonshot compost’ to feed the growth of the next project.

Perpetuate such behavior, and we see ambitious corporations that promise the moon to investors and shareholders continually downgrading expectations. Down from revolutionary new tech, down from double-digit annual growth projections. It is the curious corporations that steal a march over them.

Perhaps, paradoxically, it is the most curious organizations that are truly ambitious.

Can greatness be planned or is it unplannable?

Last week I raised a toast to aimless wandering. To the virtues of following our curiosity. To using stepping stones as a unit of innovation and to continue building bigger edifices of human progress.

This week I reflect on how _greatness needs to be planned to the breakthrough point. _Define the bottleneck that needs to be removed to realize the full potential of our best ideas, and be ready to scrap the whole thing if that’s not possible. And if that does come to pass, it’s not wasted time and energy. Put it down as a valuable experiment, the lessons of which gives the next team of inventors a headstart.

Where the two theories unite is that they are both mechanisms to deal with uncertainty. While one proposes embracing it and being okay with going where serendipity pulls you, the other suggests _accepting _uncertainty and then defining the range within which you’re prepared to face it and beyond which you’re ready to walk away.

Humans share a visceral response to uncertainty. Uncertainty triggers an aversion to loss. Never mind the transformative potential of a radical idea. We can and do look past it all the time. We are deeply guided by our worries. And here we have two proven ways of dealing with uncertainty, which are in essence saying the same thing: accept uncertainty, embrace it, make the most of it.

It matters less whether we pick the path of curiosity or monkey-training for the pursuit of our deepest ambitions. What matters is that we enjoy the thrill of the chase.

PIECE 2

This newsletter is called Curiosity > Certainty.

If that suggests to you that being curious is a big deal around here, that’s probably right. But how easy, if at all, is following one’s curiosity? I for one have struggled. This week I write about why it might be a good idea to give yourself permission to follow your curiosity.

Give yourself permission—that may sound like a strange turn of phrase. But it isn’t, really. You can point to the stack of bills, to the safety of a stable income, to your comfortable lifestyle and say that those are the things that tip the scale. Where does permission come in at all?

Hear me out.

Last year I read a fascinating book. It questioned the value in setting objectives. It argued against what it called objective-oriented search. In doing so, it suggested something you don’t hear much about: Greatness cannot be planned. It is unorchestrated. It is inherently uncertain.

Now before you call me a lover of chaos, here are two facts:

  • The book was written by two AI researchers, one of whom is with OpenAI.
  • I’m just coming off a 12-week plan (April-June) at work that is the very definition of structured pursuit.

So why is it a good idea to allow yourself to follow your curiosity?

‘Start with the end in mind’ is great advice but….

when you start with the end in mind, your curiosity depends directly on the difference between where you are AND the end point. This gap between the present and the desired future defines any objective-oriented search. It makes you constantly look ahead and measure. The closer you are to the end, the less curious you are.

Quoting from _Why Greatness Cannot be Planned: The Myth of the Objective, _authored by AI researchers Kenneth Stanley and Joel Lehman:

Microwave technology was not first invented for ovens, but rather was part of magnetron power tubes that drove radars. Only when Percy Spencer first noticed the magnetron melt a chocolate bar in his pocket in 1946 did it become clear that microwaves are stepping stones to ovens.

Turns out Amazon Web Services has shades of the microwave-oven origin story.

Around 2000, Amazon was growing fast, hiring big, but not building applications any quicker. They were taking months just to build out the compute or storage component for applications. They had teams building their own resources for individual projects, with no thought to scale or reuse. That’s when the smart people at Amazon realized they needed a set of common infrastructure services everyone within could access.

And you would think one day someone particularly bright got the idea for AWS and its transformative potential. Not quite.

A TechCrunch piece points out that the core systems in AWS developed organically ‘without any real planning’ over 2000-2003. No one realized in those early years that they had the building blocks of a business that would become AWS. Andy Jassy, AWS CEO, who was there from the start, says: ‘In retrospect it seems fairly obvious, but at the time I don’t think we had ever really internalized that’.

Katalin Karikó’s requests for funding for her novel idea for making synthetic mRNA to fight disease were rejected time and again. Both public and commercial funders thought it was a pipe dream. Few saw mRNA as a stepping stone to turning the human body into an ‘on-demand drug factory’. Along came COVID.

Different things led to these groundbreaking innovations coming to fruition but none of the ingredients came about with the dish in mind.

The difference between looking at the past and looking at the future

_Objective-driven search _compares the present with the future. Because the future is defined, we work to reduce the gap between where we are and where we want to be. It’s the goal we’ve set out to achieve. The driving force behind this kind of search is how soon before I can start cashing in—ideal weight, plum job, market share?

Non-objective search has no path mapped out to the destination. There’s no destination. Possibilities emerge with each step we take. The whole point of our journey is to keep trying new things with no goal in sight. The driving force?—I wonder what’s coming next.

A study of the evolution of our biggest inventions from their fundamental components bears this out. The path from vacuum tubes to modern computers was long. Vacuum tubes didn’t make people think of computers for over 100 years.

Why didn’t scientists simply set their objective to build a computer that’s as fast as one today?

The first computer ever built, ENIAC, was two million times slower than today’s computers. Imagine if they had in 1946 defined the configuration of a modern computer and greenlit a project to build it. What all did they need to know about just to be able to set themselves a target–graphics cards, microprocessors, operating systems? What evidence did they have at hand to be able to think this was within the realm of possibility?

In 1946, modern computing speed seemed impossible, hence it was not and could not have been an objective.

Which brings me to the second insight about non-objective search: it looks to the past to decide what’s possible.

If objective pursuits adjust direction by looking ahead, non-objective exploration learns from looking back.

The problems with a false compass - Can not trying be better than trying?

Shreyas Doshi, product management consultant who has worked with Twitter, Stripe, and Google, explains why product initiatives sometimes fail at successful firms. He suggests that those involved have too much ‘need to manufacture metrics & milestones to show straight-line progress and demonstrate certainty during an inherently uncertain journey’.

The seduction of straight-line progress en route to the accomplishment of a neatly defined yet audacious objective is what Stanley and Lehman call a false compass in their book.

A maze is a good search space to illustrate the deception of the false compass.

Researchers Stanley and Lehman tested two algorithms:

  1. where the goal was to get a wheeled robot to learn to find its way out of a maze (called objective search), and

  2. to have the same robot try only ‘new’ behaviors in the maze with no specific objective (called _non-objective _or novelty search).

They repeated the experiment 40 times for each algorithm. In 39/40 non-objective tries, the robot solved the maze. Only in 3/40 objective tries, the robot solved the maze.

>>>>> gd2md-html alert: inline image link here (to images/image56.jpg). Store image on your image server and adjust path/filename/extension if necessary.
(Back to top)(Next alert)
>>>>>

alt_text

>>>>> gd2md-html alert: inline image link here (to images/image57.jpg). Store image on your image server and adjust path/filename/extension if necessary.
(Back to top)(Next alert)
>>>>>

alt_text

In the objective search, when the robot came closest to the goal it found itself at a dead end. But it continued with the behavior because that was ‘good’ by the objective function of the algorithm (how far it was from the exit). The robot kept ramming into a dead end.

In the non-objective search, the algorithm simply tried to come up with new paths for the robot each time. For this, the robot did more things that seemed exploratory. The book explains:

So, if the robot goes around a wall it’s never gone around before, then chances are that a slight modification of that behavior might go even further. On the other hand, if the robot does something it’s done many times before, like crash into a wall, then that behavior is ignored and not explored further.

This points to two startling findings:

1️⃣_Not trying _(to reach a goal) succeeds more often than _trying _does. The robot made its best discoveries (solving the maze) when making discoveries was not its objective, but trying new things was.

2️⃣New ideas are the stepping stones to newer ideas. You know this from pursuing creative thought: You may have an interesting idea and then after giving it some thought realize that it opens up other interesting ideas.

PREVIOUS IDEA FOR THE WEEK

If learning were a product and every learner were nothing but a user, an undeniable Aha moment would be every time a learner spots a meaningful pattern among otherwise disparate things.

Spotting a pattern is like making a secret discovery. It hits you high on the pleasure meter, but when it also simplifies the world for you and for others it has consequential benefits.

These are three such Aha moments I’ve experienced recently. In this essay, I hope to connect the dots for you.

But before I dive in, a question:

Once we reach a certain level of expertise in a given discipline and our knowledge is expansive, the critical issue becomes: how is all this stuff navigated and put to use?

The answer, I believe and it is backed by research, is in better pattern recognition. We store information not as separate individual pieces but as a common collective—a cluster of similar or linked entities. An expert in a field has spent years building a network of deeply interconnected knowledge. Such that she needs to keep less in her working memory when she encounters a problem. She can invoke the appropriate pattern, like an API call, to the situation.

You know about and use these patterns all the time. Perhaps you’re familiar with the term mental models. You apply a set of proven principles when the situation fits a mold.

  1. ‘We cannot calculate our important contents, adventures, and great loves to the end.’
  2. ‘When I broke my hand in that Super-Heavyweight Finals match, time slowed down in my mind—or my perception became so sharpened, so focused on the essential, that I processed necessary information much more quickly than usual.’

Dopamine (& Nor/epinephrine) Lead to Time Overestimation;

Serotonin & Time Underestimation; Decreased Frame Rate

Fun “Feels Fast” BUT Is Remembered as Slow; Boring Stuff “Feels Slow,” Recall As Fast

Trauma, “Over-clocking” & Memories; Adjusting Rates of Experience

  1. ‘Soon enough, learning becomes unlearning. The stronger chess player is often the one who is less attached to a dogmatic interpretation of the principles. This leads to a whole new layer of principles—those that consist of the exceptions to the initial principles.’

Enjoying the article?

Check out our free ebooks and companion resources to aid your learning and development!