Amara's Law
A reader's summary of the futurist adage that we overestimate a new technology's short-run effect and underestimate its long-run one — Roy Amara's murky attribution, the Gartner Hype Cycle that formalized the same shape, and why it resurfaces in every argument about whether an AI slump is temporary.
The law at a glance
Amara's Law holds that people systematically misjudge new technology on two different timescales at once: they expect too much of it too soon, and then, once the initial hype fades and disappointment sets in, they stop expecting enough of it over the longer run — missing the effect that eventually does arrive once the technology, its supporting infrastructure and its adoption curve have all matured together.
Origin
The line is attributed to Roy Amara, a computer scientist who spent two decades at the Institute for the Future, a Palo Alto forecasting organization he led as president from 1971 to 1990. No dated paper, speech transcript, or interview has been identified as the original source; the quote is known entirely through secondhand citation by colleagues and later writers who credit Amara without pointing to a fixed original text, which means the exact wording in circulation has likely been smoothed by decades of paraphrase.
History and context
The law gained a durable second life through Gartner's Hype Cycle, a graphical model introduced by analyst Jackie Fenn in 1995 that plots a technology's visibility over time through a technology trigger, a peak of inflated expectations, a trough of disillusionment, a slope of enlightenment and a plateau of productivity. The curve is essentially Amara's two-part claim drawn as a chart, and Gartner's repeated application of it to successive waves — the internet, mobile, big data, blockchain, virtual reality, and now generative AI — kept Amara's name attached to a live, recurring commercial framework rather than letting the line fade as a one-off quotation.
Main ideas
The line itself
"We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run" — a single sentence describing a two-part forecasting error rather than a mechanism for why the error happens.
Attributed to a futurist, not a single publication
Roy Amara (1925-2007) was a computer scientist and futurist who served as president of the Institute for the Future, a Palo Alto forecasting think tank, from 1971 to 1990; the quote circulates under his name but no one has traced it to a specific paper, speech, or date he wrote or said it in.
No documented original source
Unlike laws with a dated founding document, Amara's Law is known only through secondhand attribution — colleagues and later writers citing Amara as the source — which puts it in the same evidentiary category as other widely repeated adages whose wording has been smoothed by decades of retelling rather than fixed by a citable original.
The Gartner Hype Cycle graphs the same claim
Gartner analyst Jackie Fenn's 1995 Hype Cycle model — technology trigger, peak of inflated expectations, trough of disillusionment, slope of enlightenment, plateau of productivity — is, in effect, a plotted version of Amara's two-part claim: a short-run overshoot followed by a long-run undershoot before the real, durable effect arrives.
It travels as a rebuttal, not just a description
In practice the law is invoked less to describe a pattern after the fact than to defend a technology against present disappointment — "give it time, Amara's Law" — functioning as a placeholder for patience whenever a hyped technology's near-term results underwhelm.
It gets cited for both halves selectively
Boosters reach for the underestimate-the-long-run half during a slump and the overestimate-the-short-run half is quietly dropped once a technology is finally delivering, which means the same law ends up doing service for whichever half of the claim currently flatters the speaker's position.
Critique
- It isn't falsifiable as stated. Any current disappointment can be read as evidence the technology is still in the underestimated long run still to come, and any current success can be read as proof the long run has arrived — the law offers no test that could show a given technology simply wasn't going to pay off after all.
- The attribution problem undercuts its authority. A law repeated for its wisdom rather than traced to a documented argument invites the same scrutiny as Cunningham's Law: a memorable secondhand quote treated as settled folklore rather than as a claim anyone actually tested.
- Survivorship bias in which cases get cited. The internet and mobile computing are the standard success stories invoked for the law's long-run half; technologies that were also overhyped in the short run and never delivered a long-run payoff — many since-abandoned Web 2.0 platforms, for instance — are simply not brought up as counterexamples.
- It says nothing about which technologies qualify. The law offers no criterion for distinguishing a technology merely in its trough of disillusionment from one that was correctly overestimated at every timescale, which is exactly the distinction anyone citing it in a live argument is trying to settle.
Impact
Amara's Law is now a stock reference in commentary on generative AI, cited by optimists during periods when deployed systems underdeliver on the previous year's promises, and it underwrites the entire premise of Gartner's continued Hype Cycle placements for AI sub-technologies. It also names, more cleanly than most alternatives, the same misjudgment that Rich Sutton's The Bitter Lesson documents from the inside of AI research itself: Sutton's essay is a case study of researchers repeatedly underestimating how far raw compute and scale would go, decade after decade, precisely the long-run undershoot Amara's Law predicts — except Sutton names the specific field and the specific researchers, where Amara's Law stays a general-purpose maxim usable about any technology at all.