“The answer was, for the wind-up toy, ‘Energy makes it go.’ And for the boy on the bicycle, ‘Energy makes it go.’ For everything ‘Energy makes it go.’ Now that doesn’t mean anything. Suppose it’s ‘Wakalixes.’ That’s the general principle: ‘Wakalixes makes it go.’ There is no knowledge coming in. The child doesn’t learn anything; it’s just a word.” — Richard Feynman, “What Is Science?”
“These are the signs of mysterious answers to mysterious questions: First, the explanation acts as a curiosity-stopper rather than an anticipation-controller. Second, the hypothesis has no moving parts—the model is not a specific complex mechanism, but a blankly solid substance or force.” — Eliezer Yudkowsky, Mysterious Answers to Mysterious Questions
1. Feynman’s complaint
Richard Feynman’s complaint is one of the best warnings ever issued against fake understanding. A textbook asks what makes a toy move, what makes a bicycle move, what makes things happen, and the answer is always the same: energy. Feynman objects that this teaches nothing. You might as well say Wakalixes. The child has not been given a mechanism, a picture, a causal model, or a way to anticipate anything. They have just been handed a word. (Goodreads)
This lands directly on Eliezer Yudkowsky’s Mysterious Answers to Mysterious Questions. The problem with a mysterious answer is not that it sounds spiritual rather than scientific. The problem is that it does not constrain anticipation. It has no gears. It does not tell you what must happen next, what cannot happen next, what observations would distinguish one model from another.
And Duncan Sabien’s Gears in Understanding gives the positive version of the same lesson: “Physics is (I think) the system of Gears you get when you stare at any physical object’s behavior and ask ‘What makes you do that?’ in a Gears-seeking kind of way.” That is the right move. Not verbal anesthesia, but mechanism.
So far, so standard.
But my reaction to Feynman’s complaint was not the standard one.
It did not make energy seem shallow. It made energy seem so deep that merely naming it was obviously inadequate.
2. Definition: Energy is mana.
By this I do not mean that energy is spooky, supernatural, or exempt from law.
I mean that energy plays, in reality, the role that mana plays in fantasy: it is the general-purpose spendable substrate behind physical transformation. It can be stored, transferred, concentrated, and released. Nothing gets done without paying in it. And greater control over it makes more things possible.
Fantasy invents mana because stories want a universal medium by which the world can be changed. Physics already has one. It is called energy. The difference is that fantasy mana is usually narratively convenient, while real energy comes with bookkeeping. Conservation matters. Entropy matters. Storage matters. Losses matter. Mechanism matters. The bill always comes due.
That is why Feynman’s complaint matters so much. He was right that “energy” can become a fake explanation. But precisely because it is so fundamental, energy can also become invisible through over-familiarity. We say the word too early and think we have understood something. We have not.
To say “energy did it” is a little like saying money explains an economy. True in one sense; useless in the sense that matters. What matters is where it was stored, how much was available, how it was transmitted, what controlled its release, what losses occurred, and what informational structure let it be aimed here rather than there. That is the beginning of understanding.
3. Intelligence is control
In my earlier post, I argued that intelligence is adaptive control of energy through information. The point was not that intelligence is just power. A hurricane has power. A wildfire has power. A bomb has power. None of these are especially intelligent. The key difference is control. Intelligence is not raw expenditure but directed expenditure. It is the ability to recruit energy, store it, allocate it, and deploy it in ways that are sensitive to feedback and useful for future goals. This section is the backbone of what follows.
That framing connects naturally to Yudkowsky’s The Second Law of Thermodynamics, and Engines of Cognition, which explicitly ties cognition, thermodynamics, and the physical cost of knowledge together. As he puts it, “The First Law of Thermodynamics, better known as Conservation of Energy, says that you can’t create energy from nothing,” and later, “Engines of cognition are not so different from heat engines, though they manipulate entropy in a more subtle form than burning gasoline.” Minds are not outside physics. They are one of the ways physics learns to steer itself locally through information.
Information is spell syntax.
By which I mean: information is what lets energy be directed rather than merely released. It is what distinguishes explosion from construction, fire from metabolism, lightning from computation, and brute force from intelligence.
That phrasing is playful, but I mean it seriously.
4. Definition: Magic is hidden control of energy.
By this I do not mean violation of physics.
I mean: an event feels magical when the effect is visible, but the source, storage, transmission, routing, or control of the required energy is hidden, surprising, or illegible to the observer.
This is the stronger version of Clarke’s law: "Any sufficiently advanced technology can be indistinguishable from magic". I want to cash out why. It looks magical when the observer cannot see the energy path and cannot see the informational control structure that makes the effect possible.
A room filling with light is magic until you understand generators, grids, wires, switches, bulbs, and the infrastructure behind them. A voice crossing an ocean is magic until you understand encoding, amplification, transmission, and networks. A machine answering questions is magic until you understand semiconductor fabrication, training runs, data centers, power infrastructure, and interfaces. The effect is not what makes it magical. The hidden causal depth is.
That is why magic and fake explanation can feel superficially similar while actually being opposites. A fake explanation has no gears. Magic has lots of gears; you just cannot see them.
5. Definition: Infrastructure is the wand.
By which I mean: magical-seeming effects usually depend on hidden systems that store, route, focus, and stabilize power on demand.
The “magic” of a smartphone is not in one tiny glowing rectangle. It is in the fact that an enormous lattice of extraction, manufacturing, power generation, satellites, fiber optics, wireless standards, software, logistics, and computation has been compressed into an interface so smooth that the underlying bill vanishes from view. That is one of the deepest ways the world becomes magical: not by violating causality, but by hiding it behind reliability. This is an inference from the earlier definitions rather than a quotation, but it follows directly from Clarke’s law plus the LessWrong gears framing.
Once you notice this, a lot of “impossible” things stop looking metaphysically impossible and start looking expensive, precise, and infrastructure-heavy. Turn lead into gold? Pull water from air? Grow food in sealed buildings? Heal injuries that used to cripple? Some of these may remain impractical or sharply constrained under known law. But many things people file under magic are better described as problems in energy, control, information, materials, and logistics. The issue is often not whether reality permits the effect at all, but whether one can lawfully pay the cost.
6. Your threshold for magic is low
This is where Expertium’s Intelligence Is Not Magic, But Your Threshold for ‘Magic’ Is Pretty Low fits perfectly. The point there is not that intelligence breaks physics. It does not. The point is that humans call something “magic” long before anything supernatural would be required. Our threshold is low because our intuitive models of what lawful energy plus good control can do are shallow.
That is exactly right.
The deep mistake is not believing in magic.
The deep mistake is confusing “I cannot see how this is paid for” with “there is no lawful way to pay for it.”
One of those is a statement about the world. The other is a statement about your imagination.
7. The merely real is more magical
Yudkowsky’s Mundane Magic captures an emotional stance I strongly agree with: “part of the rationalist ethos is binding yourself emotionally to an absolutely lawful reductionistic universe … and pouring all your hope and all your care into that merely real universe and its possibilities, without disappointment.”
I would only add one thing.
The merely real is not less magical than fantasy. It is more magical, because it has to actually work.
Fantasy can say “mana” and move on. Reality has to provide the reservoir, the wire, the battery, the metabolism, the turbine, the feedback loop, the waste heat, the semiconductor, the factory, the supply chain. Fantasy gets atmosphere for free. Reality has to pay cash.
That is why I like saying the definition out loud:
Energy is mana.
Not because it mystifies energy, but because it extracts the structural truth hidden in old fantasy language and then demystifies it properly. What fantasy grasped dimly is that changing the world requires some spendable substrate. What physics adds is that this substrate is lawful, conserved, measurable, transferable, lossy, and inseparable from mechanism.
8. A glimpse into the future
Once you start thinking from base principles about intelligence, energy, and magic, something changes.
A large class of “impossible” things no longer look impossible in the old sense. They start looking like questions of budget, precision, storage, routing, control, and patience. Not: can reality do this at all? But: can we gather enough energy, direct it precisely enough, and build enough intelligence and infrastructure to make it routine? This is the same step Clarke points toward when he says "the limits of the possible are found by venturing “a little way past” the impossible", and it is the same emotional move Yudkowsky points toward in “Mundane Magic.”
That gives a glimpse of a future that is much easier to miss if you stay mentally anchored to our current, very low utilization of energy and our still-crude powers of control. Food, shelter, health, abundance, even prosperity for everyone, stop looking like purely moral slogans or static slices of a fixed pie. They start looking, at least in large part, like engineering questions. Questions about whether we can access enough energy, waste less of it, direct more of it productively, and build the intelligence and institutions required to do so at scale. This is an inference, not a claim that the cited authors explicitly make in these words. But it follows from conservation, infrastructure, and the control-based view of intelligence laid out above.
None of that means every dream is easy. None of it means physics is permissive in every direction. None of it means abundance is just waiting for us if we say the word “energy” loudly enough. Feynman’s point still stands. “Energy” is not an explanation. It does not save you from gears. It demands gears.
But if you really absorb the deeper point — that intelligence is lawful control, that energy is the spendable substrate of transformation, and that magic is what hidden control looks like from the outside — then wonder stops being the enemy of realism. Wonder becomes a way of seeing reality more clearly. The world is not disenchanted when you understand that its miracles have energy paths. It is enchanted at a deeper level, because now you can start to see where the future’s magic will come from.
And that, I think, is the real unlock.
Not the childish hope that the laws of nature will make an exception for us.
The adult realization that the laws of nature are already astonishingly generous, if we learn how to work with them.
The future will not be built by denying magic.
It will be built by recognizing that much of what we call magic is really energy plus control plus intelligence — and then learning to wield more of it, with more precision, for more people, in more beautiful ways.
1. Feynman’s complaint
Richard Feynman’s complaint is one of the best warnings ever issued against fake understanding. A textbook asks what makes a toy move, what makes a bicycle move, what makes things happen, and the answer is always the same: energy. Feynman objects that this teaches nothing. You might as well say Wakalixes. The child has not been given a mechanism, a picture, a causal model, or a way to anticipate anything. They have just been handed a word. (Goodreads)
This lands directly on Eliezer Yudkowsky’s Mysterious Answers to Mysterious Questions. The problem with a mysterious answer is not that it sounds spiritual rather than scientific. The problem is that it does not constrain anticipation. It has no gears. It does not tell you what must happen next, what cannot happen next, what observations would distinguish one model from another.
And Duncan Sabien’s Gears in Understanding gives the positive version of the same lesson: “Physics is (I think) the system of Gears you get when you stare at any physical object’s behavior and ask ‘What makes you do that?’ in a Gears-seeking kind of way.” That is the right move. Not verbal anesthesia, but mechanism.
So far, so standard.
But my reaction to Feynman’s complaint was not the standard one.
It did not make energy seem shallow. It made energy seem so deep that merely naming it was obviously inadequate.
2. Definition: Energy is mana.
By this I do not mean that energy is spooky, supernatural, or exempt from law.
I mean that energy plays, in reality, the role that mana plays in fantasy: it is the general-purpose spendable substrate behind physical transformation. It can be stored, transferred, concentrated, and released. Nothing gets done without paying in it. And greater control over it makes more things possible.
Fantasy invents mana because stories want a universal medium by which the world can be changed. Physics already has one. It is called energy. The difference is that fantasy mana is usually narratively convenient, while real energy comes with bookkeeping. Conservation matters. Entropy matters. Storage matters. Losses matter. Mechanism matters. The bill always comes due.
That is why Feynman’s complaint matters so much. He was right that “energy” can become a fake explanation. But precisely because it is so fundamental, energy can also become invisible through over-familiarity. We say the word too early and think we have understood something. We have not.
To say “energy did it” is a little like saying money explains an economy. True in one sense; useless in the sense that matters. What matters is where it was stored, how much was available, how it was transmitted, what controlled its release, what losses occurred, and what informational structure let it be aimed here rather than there. That is the beginning of understanding.
3. Intelligence is control
In my earlier post, I argued that intelligence is adaptive control of energy through information. The point was not that intelligence is just power. A hurricane has power. A wildfire has power. A bomb has power. None of these are especially intelligent. The key difference is control. Intelligence is not raw expenditure but directed expenditure. It is the ability to recruit energy, store it, allocate it, and deploy it in ways that are sensitive to feedback and useful for future goals. This section is the backbone of what follows.
That framing connects naturally to Yudkowsky’s The Second Law of Thermodynamics, and Engines of Cognition, which explicitly ties cognition, thermodynamics, and the physical cost of knowledge together. As he puts it, “The First Law of Thermodynamics, better known as Conservation of Energy, says that you can’t create energy from nothing,” and later, “Engines of cognition are not so different from heat engines, though they manipulate entropy in a more subtle form than burning gasoline.” Minds are not outside physics. They are one of the ways physics learns to steer itself locally through information.
Information is spell syntax.
By which I mean: information is what lets energy be directed rather than merely released. It is what distinguishes explosion from construction, fire from metabolism, lightning from computation, and brute force from intelligence.
That phrasing is playful, but I mean it seriously.
4. Definition: Magic is hidden control of energy.
By this I do not mean violation of physics.
I mean: an event feels magical when the effect is visible, but the source, storage, transmission, routing, or control of the required energy is hidden, surprising, or illegible to the observer.
This is the stronger version of Clarke’s law: "Any sufficiently advanced technology can be indistinguishable from magic". I want to cash out why. It looks magical when the observer cannot see the energy path and cannot see the informational control structure that makes the effect possible.
A room filling with light is magic until you understand generators, grids, wires, switches, bulbs, and the infrastructure behind them. A voice crossing an ocean is magic until you understand encoding, amplification, transmission, and networks. A machine answering questions is magic until you understand semiconductor fabrication, training runs, data centers, power infrastructure, and interfaces. The effect is not what makes it magical. The hidden causal depth is.
That is why magic and fake explanation can feel superficially similar while actually being opposites. A fake explanation has no gears. Magic has lots of gears; you just cannot see them.
5. Definition: Infrastructure is the wand.
By which I mean: magical-seeming effects usually depend on hidden systems that store, route, focus, and stabilize power on demand.
The “magic” of a smartphone is not in one tiny glowing rectangle. It is in the fact that an enormous lattice of extraction, manufacturing, power generation, satellites, fiber optics, wireless standards, software, logistics, and computation has been compressed into an interface so smooth that the underlying bill vanishes from view. That is one of the deepest ways the world becomes magical: not by violating causality, but by hiding it behind reliability. This is an inference from the earlier definitions rather than a quotation, but it follows directly from Clarke’s law plus the LessWrong gears framing.
Once you notice this, a lot of “impossible” things stop looking metaphysically impossible and start looking expensive, precise, and infrastructure-heavy. Turn lead into gold? Pull water from air? Grow food in sealed buildings? Heal injuries that used to cripple? Some of these may remain impractical or sharply constrained under known law. But many things people file under magic are better described as problems in energy, control, information, materials, and logistics. The issue is often not whether reality permits the effect at all, but whether one can lawfully pay the cost.
6. Your threshold for magic is low
This is where Expertium’s Intelligence Is Not Magic, But Your Threshold for ‘Magic’ Is Pretty Low fits perfectly. The point there is not that intelligence breaks physics. It does not. The point is that humans call something “magic” long before anything supernatural would be required. Our threshold is low because our intuitive models of what lawful energy plus good control can do are shallow.
That is exactly right.
The deep mistake is not believing in magic.
The deep mistake is confusing “I cannot see how this is paid for” with “there is no lawful way to pay for it.”
One of those is a statement about the world. The other is a statement about your imagination.
7. The merely real is more magical
Yudkowsky’s Mundane Magic captures an emotional stance I strongly agree with: “part of the rationalist ethos is binding yourself emotionally to an absolutely lawful reductionistic universe … and pouring all your hope and all your care into that merely real universe and its possibilities, without disappointment.”
I would only add one thing.
The merely real is not less magical than fantasy. It is more magical, because it has to actually work.
Fantasy can say “mana” and move on. Reality has to provide the reservoir, the wire, the battery, the metabolism, the turbine, the feedback loop, the waste heat, the semiconductor, the factory, the supply chain. Fantasy gets atmosphere for free. Reality has to pay cash.
That is why I like saying the definition out loud:
Energy is mana.
Not because it mystifies energy, but because it extracts the structural truth hidden in old fantasy language and then demystifies it properly. What fantasy grasped dimly is that changing the world requires some spendable substrate. What physics adds is that this substrate is lawful, conserved, measurable, transferable, lossy, and inseparable from mechanism.
8. A glimpse into the future
Once you start thinking from base principles about intelligence, energy, and magic, something changes.
A large class of “impossible” things no longer look impossible in the old sense. They start looking like questions of budget, precision, storage, routing, control, and patience. Not: can reality do this at all? But: can we gather enough energy, direct it precisely enough, and build enough intelligence and infrastructure to make it routine? This is the same step Clarke points toward when he says "the limits of the possible are found by venturing “a little way past” the impossible", and it is the same emotional move Yudkowsky points toward in “Mundane Magic.”
That gives a glimpse of a future that is much easier to miss if you stay mentally anchored to our current, very low utilization of energy and our still-crude powers of control. Food, shelter, health, abundance, even prosperity for everyone, stop looking like purely moral slogans or static slices of a fixed pie. They start looking, at least in large part, like engineering questions. Questions about whether we can access enough energy, waste less of it, direct more of it productively, and build the intelligence and institutions required to do so at scale. This is an inference, not a claim that the cited authors explicitly make in these words. But it follows from conservation, infrastructure, and the control-based view of intelligence laid out above.
None of that means every dream is easy. None of it means physics is permissive in every direction. None of it means abundance is just waiting for us if we say the word “energy” loudly enough. Feynman’s point still stands. “Energy” is not an explanation. It does not save you from gears. It demands gears.
But if you really absorb the deeper point — that intelligence is lawful control, that energy is the spendable substrate of transformation, and that magic is what hidden control looks like from the outside — then wonder stops being the enemy of realism. Wonder becomes a way of seeing reality more clearly. The world is not disenchanted when you understand that its miracles have energy paths. It is enchanted at a deeper level, because now you can start to see where the future’s magic will come from.
And that, I think, is the real unlock.
Not the childish hope that the laws of nature will make an exception for us.
The adult realization that the laws of nature are already astonishingly generous, if we learn how to work with them.
The future will not be built by denying magic.
It will be built by recognizing that much of what we call magic is really energy plus control plus intelligence — and then learning to wield more of it, with more precision, for more people, in more beautiful ways.