The Energy of Human-Scale Systems
The self-sustaining flame is a way to think about the properties of intrinsic self-reinforcing forces within ecosystems. But what is that force? It turns out that the engine at the core of ecosystems helps us understand the engines at the core of all human-scale systems: individuals, teams, and ecosystems. The core factor is energy.
Newton’s first law tells us that objects at rest tend to stay at rest, and objects in motion tend to stay in motion, unless acted upon by an outside force. We can understand the actions of humans and human-created systems via a parallel to these fundamental laws of the universe.
In the idealized world of physics, the underlying substrate is a perfect, frictionless, infinite plane. In the plane of human existence, on the other hand, it’s more akin to a viscous swamp of molasses. Things only happen with movement, and movement takes a constant influx of energy to overcome the friction. More concretely, you can imagine an object having a force continuously applied to it to move it to a given goal state.
That intrinsic energy can come from individuals, teams, or whole ecosystems. In some cases these sources of intrinsic energy have their output multiplied with a self-reinforcing feedback loop. At that point the system can be thought of as an engine.
Once you recognize these engines, ideas that previously seemed like no-brainers will become obviously unworkable because of hurricane-force headwinds, and ideas that previously seemed unremarkable will turn out to be powerful because of having strong winds at their back. Discovering these engines and incorporating them into your strategy is an important category of insight. Engines are a powerful force, and to do things sustainably and at scale requires identifying them and harnessing their energy.
Intrinsic energy comes from within an individual, team, or ecosystem. It is in some ways a “free” source of some limited amount of energy.
Physics tells us that there are no sources of free energy, which appears to break this metaphor. But it turns out the sources of intrinsic energy stack on one another, and ultimately come from the same underlying source. Ecosystems ultimately derive their energy from the actions of high-functioning teams, teams ultimately derive their energy from high-functioning individuals, and ultimately individuals derive their energy from, well, food. In that way, this framing is not just a metaphor, but in some ways literally true.
The energy of these systems comes from food, but the direction that energy is applied comes from the accumulation of a multitude of human decisions. Those decisions are influenced primarily by incentive structures, and those incentive structures can be themselves influenced by other neighboring systems. Human systems that are sufficiently strong create a power differential that changes the local incentive structure (“if you can’t beat ’em, join ’em”) akin to a gravitational field.
For individuals, the intrinsic energy is fundamentally the human drive to grow, thrive, and succeed. It’s widely acknowledged that when individuals are in their best conditions they can achieve far more than they could in a normal state. This is known varyingly as “finding your passion”, identifying your “highest and best use”, or even in some circles as “following your bliss.” The thing they all have in common is the individual being in a flow state. In the best cases, being in this flow state doesn’t take mental energy, but actually helps give it, allowing the individual to sustain it in perpetuity. This is why a good way to optimize your personal impact is to identify the types of tasks that give you energy, and maximize the proportion of your time spent doing those. Likewise, arguably the most important metric for a manager to maximize is the percentage of time their reports are in their flow states.
For teams, intrinsic energy comes from a high-functioning group that has diversity (in all ways, including diversity of experiences, skillsets, and personalities), a shared mission (which helps align incentives), and autonomy. A good way to improve the intrinsic energy of a team is to help them respect and value one another, or to give them a clear purpose they can all work towards. A bad way to improve the intrinsic energy of a team is to identify a shared enemy, activating the powerful us-vs-them dynamic to motivate motion. This is a bad tactic because from a slightly broader perspective it is obviously toxic (what if that shared enemy is another team within the company?), and that negative energy tends to spiral out of control and infect other team dynamics over sufficiently long time scales. This pattern might sound rare, but in practice it’s disturbingly common because of how easy it is to unintentionally fall into the us-vs-them mindset, and how quickly it spirals from there as trust erodes.
For ecosystems, the intrinsic energy is the sum total of all of the individual teams’ local incentives to grow, thrive, and profit. These individual teams are often in competition: within a large company they might be in different divisions; in the broader economy they might be competing companies. But all of their individual incentives create an emergent system that can achieve powerful outcomes that no individual team had any intent of doing. The core mechanisms of capitalism and democracy harness this energy.
Intrinsic energy can only come from within — otherwise, it’s extrinsic energy, which must come from somewhere. It is almost meaningless to talk about “creating” intrinsic energy. Instead it’s best to think about discovering it, and then facilitating or nurturing it to reach its full potential. Sometimes trying to grow intrinsic energy is a bit hard (for example inspiring people on a project), and sometimes it’s impossible (for example when incentives are directly opposed).
But nurturing of intrinsic energy has only a linear multiplier effect. To truly have an engine, you need a self-reinforcing feedback loop.
Feedback loops are the magical properties that make engines almost unlimited sources of power. Without a feedback loop, you have garden variety motivation and execution — powerful, but limited.
A self-reinforcing feedback loop, in this context, has the property that the more energy that moves through it, the faster it goes. This means that energy compounds at an ever-increasing rate, which makes them extremely powerful. When nurtured correctly they create repeated order-of-magnitude increases in output.
This property makes them easy to spot: some success metric that is growing at an accelerating rate for a sustained period of time. Because of the compounding effect, the scale of the chart doesn’t matter much. As long as the fundamental source of intrinsic energy doesn’t run out (and you don’t approach your invisible asymptote), relatively soon the metric will be orders of magnitude greater than what it was before. This property makes them sneak up on teams who aren’t looking for them; when they’re small they’re easy to dismiss, and by the time you’ve noticed it’s too late to do much about it.
For individuals, the feedback loop is the increase in new mental tools and accumulated experiences that compounds, allowing them to tackle ever larger and more ambiguous problems.
For teams, there are a number of feedback loops. One is about a diversity of perspectives working together well, giving them the tools to tackle large, multifaceted problems that no individual could tackle on their own. High-functioning teams are often able to deploy collaborative debate due to a culture of mutual respect. In high-functioning teams individuals inspire and empower each other, complementing each other’s skills. Healthy teams also tend to attract other great people.
Ecosystems have a number of different types of feedback loops, and all of them are more powerful than those for teams or individuals. The strongest is the type of ecosystem with a two-sided market (e.g. a developer / end-user ecosystem), which have straightforward, massively powerful self-reinforcing effects. But there are also single-sided ecosystems with powerful feedback loops powered by classic network effects (e.g. a social network). Finally, the weakest ecosystem loop is single-sided effects with a product whose value for users is sufficiently greater than the friction to use it, which allows it to grow organically through word of mouth and increasing engagement.
Feedback loops, like intrinsic energy, can be nurtured and improved. A feedback loop is fundamentally enabled by a series of incentives that line up into a loop. But once you have that, there are a number of accelerants that can help linearly affect the underlying flow rate of the loop, which then compounds. Accelerants can include distribution, features that boost engagement, marketing, adding moments of extra delight, extra incentives, etc. But accelerants don’t matter without the fundamental enabler of the loop.
When you combine a source of intrinsic energy with a functioning feedback loop, you have an engine. Engines are effectively sources of unlimited, sustainable energy. Harnessing their output is one of the best ways to accomplish things at massive scale.
All engines are powerful, but in particular a healthy ecosystem engine is one of the most powerful human systems in the world. The base amount of energy is massive — the combined output of all of the participants — and the feedback loops compound strongly. A powerful ecosystem is a gravity well, strongly bending neighboring incentives into it, pulling ever more energy into it at an accelerating rate.
Creating an engine from scratch is basically impossible. Engines are a summation of inter-dependent incentives, which makes them highly emergent systems. These kinds of systems are orders of magnitude more challenging to properly analyze than non-emergent systems. Even tiny changes in their configuration have chaotic results, with only an infinitesimal fraction of possible configurations leading to a stable equilibrium. The result is that any engine crafted in a lab will likely implode or just fail to start. In almost all cases you have to start from a seed that already exists —that is, discover an engine already in equilibrium — and then grow and evolve it from there continuously.
The Siren’s Call of Heroism
Highly effective individuals often fall into the trap of heroism. They’re so good at doing things on their own that whenever they see a problem, their first instinct is to dive into the breach and fix it themselves. The reason this problem is so pervasive is because often their heroics are the best way to fix the problem — but only in the short-term.
In the long-term, heroics prevent more powerful, sustainable forces from taking hold. Heroics lead to things that never get better on their own, meaning they will forever take that heroic effort to make them work. For example, if you power through “office politics” that is actually due to a poorly aligned org structure, that office politics will never go away (and will likely intensify as us-vs-them sets in). Similarly, if you’re micromanaging your reports — if you aren’t nurturing their intrinsic motivation by giving them autonomy and opportunity — then you’re basically just getting a small constant-factor on your own intrinsic energy, a factor that never gets better. But heroics aren’t a good idea for many other reasons.
Heroics comes from your own intrinsic energy, applying it as an external force onto another system. You can imagine heroics being necessary to counter an opposing force on the way towards your goal.
If the other system is another linear intrinsic force then this approach isn’t particularly sustainable because the opposing force will keep rising to meet your force. If you give up at any point or get distracted, the other force will overwhelm you and push you back. This means that it’s especially difficult to do other things, since a portion of your energy and attention are perpetually in use.
Heroics is kind of unsustainable when applied to another linear intrinsic force. But it’s impossibly unsustainable when applied against an engine, or even worse, to an ecosystem engine. That’s because those engines provide large sources of effectively free energy, often orders of magnitude greater than what any individual can provide. To oppose them directly with a source of energy that is orders of magnitude smaller is pure folly.
One way to deal with an opposing force is to not try to counter the force head on. Instead you come at it from an angle in an attempt to divert the momentum instead of stop it. This is far more effective per unit energy, especially if you can divert the momentum to a region of generally good outcomes.
But this technique of applying force at an angle is a more general insight than just countering an opposing force. When used correctly it’s the secret to scaling.
Motive Force vs Guiding Force
So far we haven’t distinguished between how we use our individual intrinsic energy to affect outcomes. Energy is just energy, force just force. But it’s helpful to think of two categories of force: motive force and guiding force. Motive force helps move in the general direction of a goal, which must be continuously applied to overcome the high friction of the molasses swamp. Guiding force is rotational force that affects the direction of the object. An effectively-applied guiding force is often referred to as a nudge.
In a car, motive force would be the engine, and guiding force would be the steering wheel. The powerful thing about guiding force is that it only needs to be applied when the momentum of the object drifts off course. If there are no countervailing forces pulling your momentum continuously off course, you can often apply guiding force only for short, intermittent bursts.
A constant size motive force has constant-size impact. But a constant guiding force, when applied to powerful engines, can have unbounded impact, in proportion to the motive force provided by the engine. The mechanism is similar to how the martial art aikido uses the attackers’ own momentum against them.
The secret to scaling is to maximize the amount of your energy you use as guiding force, by harnessing the motive force of existing engines.
Nudges come from many sources. For individuals it’s helpful feedback, or ensuring good opportunities are in reach. For teams it can be asking the right questions, helping define a north star, or good old-fashioned behind-the-scenes influencing in 1:1s. For tech ecosystems it’s carefully setting up rules and designing APIs to create guardrails that divert a good amount of energy from developers’ short-term incentives towards a collective long-term goal.
There are some downsides to applying energy via guiding force over motive force. Often the momentum of an engine means that you’re only able to affect it in meaningful ways over sufficiently large time horizons — and sometimes you simply don’t have enough runway. In addition, it’s fundamentally a precision trade-off. When you’re applying the motive force yourself, you can precisely move an object to its goal state. But when you’re applying guiding force to an engine, it’s much harder to be precise.
But in practice this trade-off is worth it. It turns out that goals states are much larger than they might otherwise appear, and it’s easy to get to “good enough”. Plus, the massive scale you can sustainably accomplish by harnessing an engine’s energy more than makes up for the trade off in precision.
When and where to nudge
Nudging is not free, and there are often almost infinite possible nudges that you could apply. How do you know where to apply your energy?
The main question to ask yourself is “if this system continues on its current trajectory, will it converge to a good-enough solution in a reasonable time-frame with acceptable efficiency?” If the answer to this question is “yes” then your guiding force is likely best used elsewhere. But keep an eye on it in case conditions change.
It’s important not to aim for an unrealistic level of precision or efficiency. A healthy engine has more than enough energy to spare that worrying about some inefficiency isn’t worth it. As you approach perfect precision or efficiency, the amount of energy required to achieve it approaches infinity. That is, the amount of energy necessary to get the system to 80% efficiency is only a fraction of the energy required to get the system to 95% efficiency. As long as the system is converging on a good outcome, it’s OK to not sweat the details.
Be as patient as your runway allows. Things get increasingly uncertain in the long term, but if your goal region is large enough and you have enough runway, it’s nearly always best to allow engines to drive to the right outcome. Engines that move under their own power get stronger and stronger, which makes them even more powerful tools in the future if you don’t intervene. Think of it as an investment.
The bang-for-buck of your nudges goes down quickly as you get farther afield from your “sphere of influence,” reflecting the fact that systems are more likely to accept your nudges if you have earned credibility or otherwise have power over them. It’s important to factor that into your cost/value trade-off as you consider possible courses of action, because sometimes a less-effective but cheaper, close-by nudge is better, especially if you have the runway to wait for it to converge. The net result is you’ll end up disproportionately asserting guiding energy in your sphere of influence.
It is possible to apply too much nudging energy. Your nudges will be the most effective when they are nearly aligned with the underlying system’s momentum. Nudges that are perpendicular to the system’s momentum are OK, but only if used sparingly. Often systems will seek to route around a sufficiently strong or obvious nudging force (and nudges applied from farther afield are more likely to be viewed with suspicion). When systems react in this way, the outcome is almost always bad: being blown off course needlessly, the system generating more momentum towards a bad outcome, or even worse dampening the intrinsic energy of the system, for example by encroaching on individuals’ autonomy and deflating their motivation.
The extreme of too much nudging energy is when you apply a force to directly counter an intrinsic force of energy — that’s called applying stop energy. Systems generally don’t take kindly to stop energy and will seek to route around it, now with a convenient shared enemy to motivate them. In addition, they’ll now be inclined to see all future attempts of yours to nudge as bad-faith efforts to stop them, severely limiting your future ability to help. Finally, there’s a good chance that they’re seeing something you’re not, and that you’re wrong. Applying stop energy to impede a good thing from happening is morally bad, and in ambiguous environments it can be difficult to distinguish between good and bad things. If you see a thing that appears to be on a collision course or headed for a cliff, see if you can divert them away from that calamity with a small nudge. If you can’t, figure out how bad it is, from the broadest possible perspective, for it to go over that cliff. If it’s not that bad (just opportunity cost, for example), then it’s probably best to just leave it. Worst case they’ll learn something about how to avoid that cliff in the future, and there’s always a chance that they were right and you were wrong. Trying to fix everything — especially things that are far afield and thus extremely costly to fix — is another form of unsustainable heroism.
Engines are powerful invisible forces that you need to understand in order to affect them. But there are also other invisible forces at play, the equivalent of the laws of physics. For example, when there’s a natural dividing boundary between two teams (e.g. physical location, company division, team culture), over time the two teams will tend to clash unless energy is applied to help them work together harmoniously. As another example, people in general don’t like thinking through ambiguous or abstract things, which means that in practice teams tend to focus too much on their short-term goals even if they conflict with their long-term goals. Another example is Goodhart’s Law, “When a measure becomes a target, it ceases to be a good measure” — that is, systems will do precisely what they are incentivized to do, which is probably not exactly what you intended. By understanding these other invisible forces, you can identify even better nudges. If you know that one of these other invisible forces will pull the engine in the right direction over time, you might be able to save energy and let it do the work for you. Conversely, you might discover that a system that seems like it’s converging on a good outcome is affected by one of these forces of gravity and will require a consistent corrective nudge to stay on track.
Harnessing the power of engines
Detecting and harnessing the power of these invisible forces is hard. It takes careful study, thoughtful introspection, and wading deep into ambiguity. Sometimes it’s tempting to just ignore them, especially in the heat of the moment. But like it or not, these forces are powerful and omnipresent. Harnessing them — figuring out how to work with them and not against them — is the only sustainable way to scale. When you’re designing a rocket, gravity is kind of a bummer. But if you acknowledge its existence — if you figure out how to work with it and not against it — you can slingshot a rocket to the farthest reaches of the galaxy.