Why your brain treats a product launch like a hit of dopamine — and why the crash that follows is the whole point.
In 1997, Wolfram Schultz stuck electrodes into the brains of macaque monkeys and squirted juice into their mouths. What he discovered reframed our understanding of desire, disappointment, and why your timeline was insufferable in the spring of 2023.
The experiment was elegant. Give a monkey an unexpected sip of fruit juice, and a cluster of neurons in the midbrain lights up — dopamine cells firing in a burst. The monkey got something good it didn’t see coming. But Schultz kept going. He paired the juice with a tone: play the sound, deliver the reward. After a few repetitions, something shifted. The dopamine burst migrated. It no longer fired when the juice arrived. It fired when the tone played — when the monkey first expected the juice. The reward itself became neurologically boring. The anticipation became the drug.
Then came the cruelest finding. Schultz played the tone and withheld the juice. The monkey’s dopamine didn’t just return to baseline. It dropped below baseline — an active depression signal, the neurons firing in a pattern that encoded something specific: I expected a reward, and it didn’t come. Not just the absence of pleasure. The presence of punishment. The monkey’s brain treated a broken promise worse than no promise at all.
Schultz had discovered what neuroscientists now call dopamine prediction error — the brain’s core teaching signal. Dopamine doesn’t track reward. It tracks the gap between what you expected and what you got. Better than expected: burst. As expected: silence. Worse than expected: crash. This insight was so foundational that in 2017, Schultz, Peter Dayan, and Ray Dolan shared the Brain Prize — a million euros from the Lundbeck Foundation — for showing how this mechanism underpins learning, addiction, depression, and decision-making across virtually every domain of human behavior.
Including, it turns out, how we adopt technology.
If you’ve spent any time in enterprise tech, you’ve seen the Gartner Hype Cycle — that smooth, confident curve charting how technologies move from breakthrough to buzz to backlash to boring utility. Gartner publishes over 130 of these annually, tracking 1,900 innovations through five named phases: Innovation Trigger, Peak of Inflated Expectations, Trough of Disillusionment, Slope of Enlightenment, Plateau of Productivity.
The model has plenty of critics. An empirical analysis by Steinert and Dedehayir examining 40-plus technologies from 2003 to 2009 found that Gartner’s placements often diverged from actual market visibility, relying on analyst judgment rather than standardized metrics. The curve shape itself has no mathematical derivation — it’s a vibes diagram dressed up in research methodology branding. Only about 20% of technologies tracked ever complete the full five-phase journey. Most die in the trough or stall indefinitely.
And yet the model persists. Not because it’s scientifically rigorous, but because it captures something psychologically real. Anyone who lived through the blockchain craze of 2017, the metaverse frenzy of 2021, or the generative AI explosion of 2023 recognizes the emotional arc: novelty, overenthusiasm, disappointment, grudging acceptance, quiet productivity. The curve feels true, even when the specifics don’t always line up.
Here’s why it feels true: the Gartner Hype Cycle isn’t a model of technology adoption. It’s a model of collective dopamine prediction error, operating at population scale.
Lay Schultz’s three-state model over Gartner’s five phases and the correspondence is almost uncomfortably precise.
Innovation Trigger = First Unexpected Reward. A new technology appears. Nobody predicted it. GPT-3 generates coherent paragraphs. DALL-E makes pictures from words. The collective brain encounters an unexpected reward, and dopamine neurons across millions of people fire in concert. This is genuine positive prediction error — reality exceeded expectation. The excitement is not irrational. It’s the correct neurochemical response to novelty.
Peak of Inflated Expectations = Dopamine Transfers to the Cue. This is where Schultz’s tone-before-juice finding becomes critical. Just as the monkey’s dopamine migrated from the juice to the tone that predicted juice, collective excitement migrates from the technology working to the announcement that the technology exists. Product launches, keynotes, demos, and breathless headlines become the dopamine event. The actual product experience is almost beside the point. By the time ChatGPT reached 100 million users in January 2023, the dopamine was firing on the idea of what generative AI would become — not on what it could actually do that Tuesday. Seventy-nine percent of enterprises planned GenAI projects. Five percent had production use cases.
Trough of Disillusionment = Negative Prediction Error. This is the cruelest phase, and Schultz’s monkeys explain exactly why. When the predicted reward fails to materialize — when the hyped technology doesn’t transform your business, when 95% of enterprises report no meaningful ROI on AI investments by August 2024, when “the great AI hangover” sets in — the collective brain doesn’t simply return to neutral. It drops below baseline. Dopamine neurons fire in that active depression pattern. The gap between what was promised and what was delivered generates a signal that feels worse than never having expected anything at all.
This is why failed technologies aren’t just forgotten — they’re treated with contempt. Google Glass didn’t fade from memory; it became a punchline. The metaverse didn’t quietly underperform; it became a symbol of corporate delusion. Negative prediction error doesn’t produce indifference. It produces active resentment. The brain is punishing you for having believed.
Slope of Enlightenment = Recalibration. Slowly, expectations reset. The hedonic treadmill — that well-documented psychological phenomenon where humans return to a stable happiness set point regardless of what happens — does its work. People stop expecting transformation and start expecting utility. Prediction errors shrink. Dopamine signaling stabilizes. Second-generation products arrive that do less but deliver more. The technology begins to be evaluated on its own merits rather than against inflated fantasies.
Plateau of Productivity = Dopamine Silence. And here’s the punchline that makes the whole framework click. When a technology reaches the plateau — when it works reliably, when it delivers what people expect — it generates zero prediction error. No burst. No crash. No excitement at all. Dopamine neurons go quiet.
This is not failure. This is the goal.
Think about electricity. TCP/IP. Indoor plumbing. Container shipping. These are among the most transformative technologies in human history, and they bore you completely. You don’t get a dopamine hit when you flip a light switch. Your brain has fully predicted the reward. The prediction error is zero. And that neurological silence is the sound of a technology that actually works.
If collective dopamine dynamics explain the hype cycle, they also explain why the cycle seems to be speeding up. Virtual reality first appeared on Gartner’s curve in 1995, peaked immediately, then spent twenty-one years — from 1997 to 2016 — grinding through the Trough of Disillusionment. Two full decades of negative prediction error before reaching the Slope of Enlightenment.
Generative AI went from Innovation Trigger to Trough in roughly three years. 2022: trigger. 2023: peak. 2025: trough. By the time you read this, AI agents — the next phase — are already cresting their own peak, preparing to tumble into their own trough.
The acceleration tracks perfectly with prediction error dynamics. Social media creates what researchers have called filter bubbles that amplify hype signals. When every node in your network is firing on the same stimulus, positive prediction error doesn’t just sum — it compounds. The initial surprise gets reflected and amplified through millions of social connections, each one adding its own dopamine-flavored enthusiasm to the wave. The peak gets higher. But the crash, when it comes, gets proportionally deeper, because the gap between collective expectation and collective reality widens with every amplification cycle.
The macro-neuroeconomic argument — what some researchers are calling the “Dopamine Collapse Hypothesis” — suggests this is happening not just to individual technologies but to the entire digital economy. Competitive markets reward firms for maximizing engagement stimuli, which progressively desensitize the collective reward system, which requires ever-stronger stimuli, which accelerates the cycle further. It’s the hedonic treadmill applied to civilization.
There’s a practical implication buried in all this, and it’s one that Warren Buffett has been exploiting for decades without the neuroscience vocabulary.
“Be fearful when others are greedy, and greedy when others are fearful.”
Translated into prediction error terms: buy during collective negative prediction error. The trough is the moment when the population-level dopamine signal has dropped below baseline — when the technology is being evaluated not on its actual capabilities but through the neurochemical lens of broken expectations. The contempt people feel for a technology in the trough isn’t a rational assessment. It’s an active depression signal generated by the gap between what was promised and what was delivered. And since that gap is almost always distorted by the inflated peak that preceded it, the trough almost always overcorrects.
Amazon’s stock lost 90% of its value between 1999 and 2001. The internet was in the trough. The technology hadn’t changed. The collective neurochemistry had.
If you can recognize which phase of the prediction error cycle you’re in — really recognize it, not just intellectually but viscerally — you gain something rare: the ability to discount your own dopamine. To notice when your excitement about a technology is the burst of a positive prediction error rather than a sober assessment of value. To notice when your contempt for a failed promise is the crash of a negative prediction error rather than evidence that the technology is worthless.
Here’s the thing about Schultz’s monkeys that nobody talks about in the popular accounts: the most successful monkeys — the ones who learned fastest, who adapted most efficiently — were the ones whose dopamine responses eventually went quiet. Not because they stopped caring. Because they stopped being surprised. They had built an accurate model of the world. Prediction matched reality. Error signal: zero.
The next time you feel nothing about a technology — no excitement, no contempt, just a shrug — pay attention. That neurological silence might be the most important signal in the room. It means the technology has crossed from hype into reality. It means your brain has recalibrated its expectations to match what’s actually being delivered. It means the prediction error has resolved.
The most useful technologies in your life are the ones you’ve stopped thinking about. The boring ones. The ones that just work.
Your dopamine doesn’t care about those technologies anymore. That’s how you know they matter.
The boring phase is where trust gets built
When the dopamine fades and a technology has to prove itself on merit alone, what survives is verifiable track record. For AI agents, that means cryptographic provenance — not hype, not promises, but a chain of receipts that proves what an agent has actually done. Because the Plateau of Productivity runs on trust, not excitement.
Verify our provenance chain · Public chain data · pip install agent-trust-stack-mcp