Dopamine as energy allocation

Here's how I think about dopamine: it's not a signal for pleasure. It's your brain deciding where to spend its limited energy budget.

While others think about dopamine as pleasure, I've started thinking about it as energy allocation. I think this applies to all things related to motivation and habits.

The question dopamine answers isn't "did that feel good?" but rather, "should I spend energy on this?"

The monkey and the juice

The classic dopamine experiment is Wolfram Schultz's work with monkeys. A light turns on, juice comes out, the monkey drinks the juice.

Blaise Agüera y Arcas, a Google VP and researcher, describes what Schultz's lab found in his book "What Is Intelligence?":

When the monkeys first stumbled upon an action producing the sugary drink, the spiking rate of these dopamine neurons rose. Once the monkeys figured out the association between the visual cue and the reward, extra dopamine was no longer released when the treat came, but was released earlier, when the visual cue was presented. [...] If, following the visual cue, the treat was withheld, then activity of the dopamine neurons subsequently decreased — that is, they went quiet relative to their background rate.

Agüera y Arcas, Blaise. What Is Intelligence? (Antikythera) (pp. 222-234). Function. Kindle Edition.

Once the juice is fully predicted, the dopamine response drops to zero. If dopamine were about pleasure, it would fire when the monkey drinks. Every time. But it doesn't. It fires for the unexpected prediction of reward, not reward itself. Unexpected light, not expected juice.

A rat experiment from the same book makes this even more vivid (and perhaps makes you think about the life of these little fellows):

When the dopamine-producing neurons of rats are destroyed, the rats become passive and starve to death, even if food is "literally under their noses." If food is placed into their mouths, they eat it with evident pleasure, but no matter how hungry they get, without dopamine, they aren't spurred to action.

Agüera y Arcas, Blaise. What Is Intelligence? (Antikythera) (pp. 222-234). Function. Kindle Edition.

The rats still experience pleasure. What's broken is the allocation system that says "spend energy moving toward food." They like food fine — they just can't spend energy going to get it.

As Agüera y Arcas puts it, "the most apt feeling to associate with dopamine is probably anticipation" — the feeling that something good is within reach.

Not pleasure. Anticipation.

A prediction chain is the sequence your brain learns: cue leads to action leads to reward. The light comes on, dopamine says "I know this path pays off — execute, spend motor energy." If the juice is already predicted after the light, there's no new information, and therefore no new energy to spend.

What engagement actually feels like

Let's talk about engagement and BJJ. BJJ is a sport full of dopamine: sweeping an opponent with a move I've practiced, the focus when I'm in a scramble I can't predict, and the "ha didn't see that coming!" moment after getting caught in a submission — they all just feel like "I'm engaged." But they're actually three different signals, all answering the same question: "Where should I spend energy?"

The first is the spike at a cue you've learned. You see the opening, you know the path. "Execute. Spend motor energy." The sweep works because your brain has learned the prediction chain and is telling you to go. In fact, perhaps not even "your brain" but rather your nervous system — in your whole body.

The second is what happens when the outcome is genuinely uncertain — a 50/50 scramble. Dopamine doesn't spike and fall here. It ramps up and stays elevated. You're thinking harder because your brain is allocating resources to an unresolved problem.

The third is the prediction error. You expected the sweep to work but got swept instead. Oops! "My model was wrong. Update."

From the inside, all three feel the same. They all feel like "I want this." But they're three different allocation decisions.

It's similar with the guitar. When I'm working on a passage that's just beyond my ability — where I can feel myself getting closer to nailing it but I'm not there yet when I play back my playing — it's hard to put it down. All three signals are firing. When I go back to something I've already mastered, this engagement evaporates. The song hasn't changed. My fingers haven't changed. My prediction error is zero. I'm literally going through the motions.

Working on something engaging fires all three signals simultaneously. That's why real engagement feels so different from entertainment.

What keeps you locked in

Sustained drive comes down to two things: you value something highly AND you keep encountering things you can't predict about it. When both are true, dopamine keeps spending energy — or rather literally, it keeps you spending energy on it.

That combination — high value plus persistent unpredictability — is what keeps you locked in.

Think about what happens when only one is high. A comfortable job you've mastered has high value but it's fully predicted — you coast. No drive. Scrolling through random TikToks? Lots of unpredictability, but you don't care enough about any of it to follow through with any significant action.

But both high? Building something that's working but not figured out. BJJ with a partner at your level. Solving a puzzle at the limit of your capabilities. You can't stop because your brain keeps spending energy on a problem it scores as valuable for you but that it hasn't solved yet.

Not all uncertainty is equal

The meaningful distinction isn't boring vs. exciting — it's whether the uncertainty is learnable or just random.

Your brain should ideally spend energy on problems where effort actually improves your predictions. If it can't learn anything, the energy is wasted.

But there's an important nuance to this: random isn't necessarily not-addicting. Slot machines prove that.

When your brain sees maximum uncertainty, it stays engaged even if there's no learning signal. No matter how many pulls, you never get better at predicting the outcome of a slot machine.

The question to ask yourself in similar situations is: "Am I getting better at predicting my rewards from this?" If yes, you're learning and therefore using dopamine appropriately. If not, you're... well, just a monkey pulling a lever and hoping for juice.

Modern dopamine hijacking

Once you see dopamine as energy allocation, the modern hijacking of our dopamine becomes as clear as the reason why it's not satisfying.

Junk food is the primitive hijack — artificially concentrated flavors creating unnaturally strong, and wrong, prediction signals about your nutrition. We "love" things like sugar not because they're good for us, but because our allocation system evolved in an environment where calorie-dense food was always worth pursuing.

Apps, easy games and slot machines exploit prediction error by making reward structurally unpredictable. The system never converges, so engagement never stops. Cal Newport describes this mechanism in Digital Minimalism:

Many people have the experience of visiting a content website for a specific purpose — say, for example, going to a newspaper site to check the weather forecast — and then find themselves thirty minutes later still mindlessly following trails of links. [...] Every appealing headline clicked or intriguing link tabbed is another metaphorical pull of the slot machine handle. Technology companies, of course, recognize the power of this unpredictable positive feedback hook and tweak their products with it in mind to make their appeal even stronger. [...] The notification symbol for Facebook was originally blue, to match the palette of the rest of the site, "but no one used it." So they changed the color to red — an alarm color — and clicking skyrocketed.

Newport, Cal. Digital Minimalism (pp. 18-19). Penguin Publishing Group. Kindle Edition.

The red notification badge isn't informing you. It's generating a prediction error that your brain’s allocation system can't ignore.

Then there's Balatro — a roguelike poker card game I've been playing — which is interesting exactly because it hijacks all three signals at once. Strategic depth gives you real prediction chains to execute. Variable reinforcement keeps uncertainty high. And occasional massive positive (or negative) surprises trigger huge learning signals — the “oh that’s wild!” moment.

It's not exploiting one pathway — it's hijacking the entire allocation system. It's also wildly popular — Balatro sold over 5 million copies in its first year, which is unsurprising when you understand the dopamine mechanics at play and, therefore, the challenges in resisting it.

Infinite scroll is another prevalent example. It presents uncertainty that's never resolvable by design (like a slot machine), burning your attention on problems you can never get better at. It feels like rest, but it isn't. It's thousands of micro-reps training your system toward shorter and shorter loops and minimal energy spend.

Junk reps

Here's where this connects to something I've written about before: easy and hard routines.

Every time you reach for easy dopamine, you're training your brain that good rewards come cheap, that is, with low energy. That recalibrates your whole system. If a scroll gives you a dopamine variation for zero effort, the effort required for something hard — reading, building, practicing — feels impossibly expensive by comparison. Your brain won't allocate energy toward it.

Scrolling isn't rest. It's junk volume — just like the exercise that makes you weaker instead of stronger. The good news is you can restore your dopamine thresholds the same way you restore your palate: stop eating junk. When you cut out hyper-processed food, normal food starts tasting good again. Same thing happens here — cut the easy unpredictability and hard things start feeling worth doing again.

Redirecting your system

Willpower alone isn't enough here. You're trying to override the system that controls your energy allocation — the thing that literally decides what feels worth doing. You might win a few rounds with sheer discipline, but the system doesn't get tired. Yet, you do.

The better move is to stop fighting your reward system and start redirecting it. Redesign the prediction chains so your brain pulls you toward the right actions instead of against them.

In practice, one thing that this means is sequencing. Hard thing first, reward after. Not as punishment — but as a newly programmed prediction chain for your brain. Read before you scroll (if you're gonna scroll... that is). Train before you play. Build before you consume. Do it consistently enough and the hard thing stops being the obstacle — it becomes the cue that predicts the reward.

It also means cutting the shortcuts. If your brain can get a dopamine variation for very little effort, you’ll need a big variation to spend a lot of effort, which you’ll rarely do for the day-to-day rewards. That's not weakness — that's allocation working exactly as designed! So remove the free option. Delete the app. Leave the phone in another room. You're not fixing a character flaw. You're adjusting inputs to a system that works this way and that we were all born with.

And the rewards themselves matter. If your prediction chain ends in a reward of “suffer," your brain will never learn to prefer it — no matter how disciplined you are. You have to work with rewards your dopamine system actually recognizes. Movement, sunlight, real food, genuine social connection, creative flow — things your brain is programmed to pursue. You can’t trick your way to misery (if you do pushups, your reward is more pushups!). You have to build chains that end in something genuinely rewarding... at least the way your brain sees it.

The same system that keeps you locked into infinite scroll can keep you locked into deep work, if you design its inputs right. It all runs on the same mechanics. Which means your brain is not broken because you’re demotivated, your motivation is just pointed in the wrong direction. And you can change that.

Your dopamine system is going to direct your brain to spend energy somewhere. Train it. Spend your energy building the future you want.