Maybe you’ve heard this story before: New tech arrives on the market, hype shoots through the roof, investments pour in, time passes, usage plateaus, hype remorse ensues.
If it sounds like a familiar tale, that’s for good reason: These days it happens all the time. But what you probably didn’t know is that it follows the path of a decades old, seldom-cited tech tenet. Coined in 1973 by researcher and avid futurist Roy Amara, Amara’s Law declares, "We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
Let that sink in.
In a Silicon Valley culture obsessed with the sensation of progress, Amara’s Law takes a backseat to Moore’s Law. The latter, better-known axiom, stating that microchip processing capacity doubles every two years, is linear, straightforward and blindingly optimistic. Amara’s law, on the other hand, describes a cycle. And yet the two intertwine somewhere in the future: While it’s easy to imagine your laptop’s computing power doubling in two years, then again in four, once we travel years down the line, the numbers generated by exponential compounding are quite difficult for the human mind to grasp.
Amara recognized that our ability to understand (and overhype) the near term is related to our inability to properly assess the long term, and his law touches on that problem. It deals with the limits of human perception, a murkier playground than the mathematical certitude of computing. And given the current convolutions and prognostications around generative artificial intelligence, it’s extraordinarily relevant.
Right now GenAI’s flagship tech is ChatGPT. ChatGPT is an all-star example of Amara’s Law in action.
The GenAI text tool launched almost exactly a year ago. It immediately sent seismic tremors rippling from the tech world into popular culture. By January of 2023, ChatGPT had logged more than 100 million users, making it the fastest-growing consumer software application in history. In March, it was valued at $29 billion (!). By the end of Q2 ‘23, a $14 billion tidal wave of investor cash had flowed to GenAI companies and the GenAI-adjacent.
Then came September. Reports came out charting a drop in usage for three consecutive months. Apparently ChatGPT’s use cases were thin. In an interview with a German newspaper, Bill Gates declared an AI plateau. Observers had apparently overestimated its effect, leading to a leveling off of excitement.
Yet the valuation of OpenAI continues to climb—to $80-90 billion as of late September, according to the Wall Street Journal. The cash continues to flow from Sequoia Capital, the oracular investment firm, into the GenAI field. In October of 2022, roughly 16% of Sequia’s new investments were in AI; they’re now at about 60%, specifically veering from early support of foundation model-builders like OpenAI toward startups that harness foundation models. The tech giants are stepping up, too: Amazon and Google both recently invested in Chat competitor Anthropic to a total of around $6 billion. By providing affordable access to Anthropic’s Claude and other foundation models, Amazon’s Bedrock system could enable countless developers and companies to bring their AI ideas to life.
These companies are looking toward the future. They’re extrapolating from Amara’s Law.
Amara’s Law jibes with another principle I find immensely instructive right now. It’s called “skeuomorphism,” and though it comes specifically from the graphic design industry, it carries wide-ranging implications.
Skeuomorphism is that thing where a digital object is designed to appear like its real-world counterpart. The most famous example is the trash can icon on your desktop. There is no trash can on your desktop, but the image of one tells us all we need to know. Send a document there and you even hear the satisfying sound of wastepaper crumpling or hitting the bottom. The style’s popularity ebbs and flows with design trends; app tiles used to appear on our phones with drop shadows, as if you were pushing a three-dimensional, physical button. Years later, we no longer need the reassuring illusion of a button and the drop shadows are gone.
To me skeuomorphism describes the tendency most humans have to interact with new tech in the same way they interacted with whatever existed previously, prior to mastering the fundamentals and innovating within the new form. It takes time to embrace new technology that literally advances faster than our ability to interpolate it. Until then we just apply the new tech like we applied the old tech, which might be useful but doesn’t harness its full potential.
Retrieval-augmented generation (RAG), which might be considered ChatGPT’s core function, is an example of skeuomorphism. It’s extremely useful in breaking down silos of info and data inside organizations, but it’s essentially search on steroids. It’s like Henry Ford inventing a faster horse rather than an affordable automobile.
To my eye ChatGPT is to GenAI what Flash was to the early internet: a compelling advancement that will be surpassed in resolution and practicality soon enough. Users have uncovered some interesting applications, and I imagine the interface will be with us awhile, but it’s not GenAI’s omega function. You could spend a relatively small amount to add ChatGPT to your offering—and the spend is worth the return—but I urge you to continue thinking about the problem you want it to solve. ChatGPT isn’t the ultimate answer. GenAI will continue to pay off in ways we can’t yet foresee.
What we really want from GenAI is not Flash, but YouTube, a game-changing innovation that changes the way we think at a fundamental level.
Amara’s Law invites us to think differently about how we arrive at the future. So how do we get there? How do we get to the highest, best use of GenAI when we’re currently satisfied doing the old thing a new way? How do we get to the long-term evolution that we’re currently underestimating, or even completely failing to envision?
Here are three ways to start thinking differently.
- Jobs to Be Done Theory. Clayton Christensen was the man who introduced the concept of disruptive innovation to the business world. Another of his core ideas is known as Jobs Theory. As explained by the Christensen Institute,
“While conventional marketing focuses on market demographics or product attributes, Jobs Theory goes beyond superficial categories to expose the functional, social, and emotional dimensions that explain why customers make the choices they do. People don’t simply buy products or services; they pull them into their lives to make progress. We call this progress the ‘job’ they are trying to get done, and understanding this opens a world of innovation possibilities.
Such a brilliant framework! So simple yet capable of amazing insight!
Let’s take Amara’s Law to its grandest possible conclusion and think on the largest possible scale: What are the jobs that need to be done right now? End food insecurity, solve the climate crisis, end exploitative labor… the list goes on. When we think about our jobs as entrepreneurs this way, we might end up with meat created without harming animals or man-made diamonds that remove carbon from the atmosphere.
- The Artist’s Way. Yes, there’s a well-known book on this topic intended to help anyone develop their unique mode of expression, but that’s not necessarily what I’m talking about here. Artists like Van Gogh literally saw the world differently than other people, but similarly visionary creators like David Lynch adhere to disciplined routines in order to tap into their wellspring of unconscious ideas. While entrepreneurs like me might believe that the next innovation will solve everything, my impression is that great artists do not fall prey to that delusion. They’re as entranced by the journey as the destination.
- Rely on tech to show us the way. Why not ask ChatGPT and Claude what they see in their future? Why not train them on the above frameworks and query their thoughts? I don’t believe these bots are literally intelligent, and they’re certainly not evolving in a biological sense, but I think they can offer us insights that we humans might not otherwise get to. We already know that working with GenAI benefits from insightful, adaptive human input, so let’s start collaborating with these machines now. The sooner we do, the sooner we can arrive together at their most thrilling, world-changing potential.
[CALLOUT]
Amara’s Life
As someone interested in the early days of computer science, I’m deeply curious about Roy Amara, who was among the very first computer engineers in the world. Details about his life are scant, but I’ve learned that after graduating from MIT he moved to Northern California and contributed to many groundbreaking projects.
While at Stanford Research Institute, he helped develop a first-of-its-kind mainframe computer for Bank of America. ERMA—for Electronic Recording Machine, Accounting—debuted in 1955 and was the prototype for all digital banking and credit-card processing machines that followed.
Amara spent 18 years at SRI, where he worked alongside Doug Engelbart, another hero of the early computer age who deserves his own sidebar!
He joined with Paul Baran, one of the architects of ARPANET, the earliest form of the Internet, and other deep thinkers of the era to launch the nonprofit think tank Institute for the Future. It was during this time that Amara made the proclamation known as Amara’s Law. As president of the IFTF, he launched the Ten-Year Forecast, the flagship research project of the IFTF, which has evolved into an annual conference. He also initiated one of the first studies of global climate change in 1978.
Amara died in Palo Alto in 2008 at 82 years old. The IFTF is still going strong today.
[/CALLOUT]
