Amara's Law
Judge a new technology on a 20-year clock, not a 2-year one
- Difficulty
- Easy
- Time to result
- ~ongoing to results
- Steps
- 6
- Confidence
- 88%
Amara's Law says we overestimate the effect of a technology in the short run and underestimate it in the long run. The mechanism is a mismatch between two clocks. A working prototype arrives fast and gets extrapolated in a straight line, so expectations spike and then collapse when the complementary layers, regulation, cost curves, infrastructure and user habits, fail to arrive on the same schedule. Those slower layers keep compounding after attention has moved on, so the technology quietly exceeds the original forecast a decade later. Naval's calibration on this episode is that the overestimate window is roughly the first ten years and the underestimate window is the following twenty, with a crossover around year fifteen. Using the law well means dating the technology from its real birth rather than its loudest moment, then identifying which slow layer is the actual bottleneck.
Origin
Named for the futurist Roy Amara and cited by Matt Ridley in How Innovation Works. On this episode Naval Ravikant tests it against crypto, adding the ten-year over-, twenty-year under-estimate calibration and the point that crypto's clock started with Bitcoin in 2009, not at the hype bubble.
Core principles
- 01Every technology runs on two clocks: the demo clock and the deployment clock.
- 02Hype prices in the demo clock; compounding happens on the deployment clock.
- 03The bottleneck is almost never the core invention, it is the slow complementary layer.
- 04Disappointment in year five is evidence about the schedule, not about the ceiling.
- 05The long-run upside is not guaranteed, because complex systems never repeat exactly.
How to run it
- 1
Date the technology from its real birth
Find the first working instance, not the moment it became a headline. Crypto dates from 2009, not from the bubble, which makes it far older than most people assume when they call it a failure.
Pro tip Write the birth year down before you read any commentary, so the hype cycle cannot re-anchor you.
- 2
Split the claim into a short-run and a long-run version
State separately what the technology is supposed to do in the next three years and what it is supposed to do in twenty. Most arguments about a technology are two people holding different halves of this pair.
Watch out If you cannot state a specific long-run claim, you do not have a thesis, you have enthusiasm.
- 3
Name the slow layer
Identify the complementary thing that has to exist before the technology pays: infrastructure, regulation, cost per unit, or a habit change. This is what the short-run forecast silently assumed away.
Pro tip Ridley's diagnostic-device example shows regulation is often the slow layer: 20 to 70 months for a licence deters entrepreneurs from entering at all.
- 4
Set the crossover date
Mark roughly fifteen years from the real birth date as the point where the underestimate begins to bite. Before it, expect disappointment; after it, expect the surprise to run the other way.
Watch out The crossover is a prior, not a schedule. Use it to set your review cadence, not to guarantee an outcome.
- 5
Watch the green shoots, not the headlines
During the disappointment window, track whether serious builders are still working on the slow layer. Naval's tell for crypto was decentralised finance and identity or storage plumbing being built quietly after the bubble popped.
Pro tip Builder count during the trough is a better signal than price or press volume.
- 6
Re-underwrite at the crossover, not on news
Revisit the thesis on your own schedule rather than every time the technology has a good or bad week. Ask explicitly what would make this the case where the long-run payoff never arrives.
Watch out History does not repeat exactly. A complex system that reliably repeated would carry no new information, so treat the long-run upside as probable, never certain.
In the wild
Ridley expects crypto to keep disappointing for years and thinks many people will lose their shirts before it delivers, if states allow it at all. Naval agrees on the timing but re-dates the clock: Bitcoin launched in 2009, so crypto is older than the conversation assumes. He points to two green shoots after the bubble popped, a decentralised finance stack for borrowing, lending, derivatives, trading and custody, and crypto plumbing for file storage and identity that independent developers prefer because they cannot be deplatformed. His forecast is that the plumbing gets laid over five years and the results land in the following decade.
→ The same technology reads as a failure on the short clock and as an early-stage build-out on the long clock. Only the second reading generates a decision.
Naval says he has seen the overestimate-then-underestimate pattern in Silicon Valley over and over, naming autonomous vehicles, the internet, mobile phones and crypto. Each arrived with a demo that suggested imminence, went through a stretch where the sceptics looked right, and then, for the ones that survived, quietly exceeded the original claims once the complementary layers arrived. The internet was declared a fad; mobile phones were a luxury toy; autonomous vehicles are still mid-cycle.
→ A repeated cross-technology pattern that lets you predict the shape of the disappointment rather than being surprised by it.
Common mistakes
Dating the technology from the hype peak
Starting the clock at the bubble instead of the first working instance makes a mature technology look young and a stalled one look promising. Get the birth date wrong and every subsequent judgement is off by years.
Treating the long-run upside as guaranteed
The law describes a bias in forecasts, not a law of nature. Complex systems never produce the same result twice, so some technologies genuinely never pay off.
Reading short-run disappointment as a verdict
The trough is the expected part of the curve, not new evidence. Selling or quitting during it is exactly the error the law predicts people will make.
Is it for you?
Best for
Investors, founders and operators deciding whether an emerging technology is dead, early, or genuinely overhyped.
Not ideal for
Short-horizon trading or any decision whose outcome resolves inside a single product cycle.
From the transcript
“You cite Amara's Law which talks about how the effects of innovation are overestimated in the short term and underestimated in the long term.”
“I have absolutely seen that by the way. I've seen that in Silicon Valley over and over. From autonomous vehicles to the internet to mobile…”
“It wouldn't be a complex system if you could easily predict the next step.”
From the episode
Matt Ridley: How Innovation Works, Part 2
Matt Ridley