Where is the AI revolution?

AI was supposed to change everything, and it hasn't so it must all be BS

AI can’t even get its hype cycle correct. See the correct graph later in this post.

An interesting question is this: if AI is supposed to change everything, where is all of the change? Where is my AI-powered coffee maker? Where is my AI-powered doctor? And where is my AI-powered wealth manager?

If you believe all of the breathless hype coming from venture capitalists and technologists, AI is an instantaneous state change: there is the time Before AI and the time After AI. Except, the world is not that binary. The adoption of AI will be a continuous process, which will unfold over a decade or more. Anyway, here’s an article from the Economist:

Today some companies are once again ramping up [capital expenditures], to seize what they see as the enormous opportunity in AI. This year forecasters reckon that Microsoft’s spending (including on research and development) will probably rise by close to 20%. Nvidia’s is set to soar by upwards of 30%. “AI will be our biggest investment area in 2024, both in engineering and compute resources,” reported Mark Zuckerberg, Meta’s boss, at the end of last year.

The article goes on to note that, in spite of these technology companies ramping up spending on AI technology, many companies outside of the technology sector are not so smitten with it. How do we reconcile these two apparently contradictory postures? The Economist concludes:

These trends suggest one of two things. The first is that generative AI is a busted flush. Big tech firms love the technology, but are going to struggle to find customers for the products and services that they have spent tens of billions of dollars creating. It would not be the first time in recent history that technologists have overestimated demand for new innovations. Think of cryptocurrencies and the metaverse.

The second interpretation is less gloomy, and more likely. The adoption of new general-purpose technologies tends to take time. Return to the example of the personal computer. Microsoft released a groundbreaking operating system in 1995, but American firms only ramped up spending on software in the late 1990s. Analysis by Goldman Sachs suggests that while only 5% of chief executives expect AI to have a “significant impact” on their business within one to two years, 65% think it will have an impact in the next three to five. AI is still likely to change the economy, but with a whimper not a bang.

This strikes me as directionally correct, and it’s worth taking a bit of time to unpack why it is correct. There is Amara’s Law, which states, roughly¹: We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run. Here’s a handy graph which illustrates the same principle:

Sourced from here. This graph is more conventionally known as the Gartner Hype Cycle, but it also serves as a good graphical representation of Amara’s Law.

Let’s make this a bit more concrete. New general purpose technologies seem like they ought to be rapidly adopted by everyone, everywhere, all at once. Consider the computer. Early spreadsheets such as Lotus 1-2-3 were vastly superior to paper-based spreadsheets, so you’d have to be crazy not to throw out your carbon in favor of silicon! This argument is attractive to technologists and venture capitalists. The shiny new thing is superior to the tried and true old thing, for reasons, so you’d have to be crazy not to adopt the shiny new thing.

And—the technologists and venture capitalists are frequently right: the shiny new thing is superior to the old thing. It has better features, more capabilities, and it can make one’s work more efficient. And yet—companies are frequently loath to adopt new technologies. Companies are complex things, ultimately managed by people. And people are illogical. They’re emotional. They’re averse to change. They don’t like to spend money on uncertain payoffs.

Overcoming all of this institutional inertia takes time. It takes time for companies, especially companies for which technology is not a cultural identity or competitive advantage, to adopt new things. Integrating AI into new products and workflows just isn’t something that can be done with the flip of a switch, no matter how much technologists and venture capitalists assert otherwise.

So here’s the resolution for our paradox. AI technology is rapidly improving, but the application of it proceeds more slowly. The world is a complex place, and the technologists’ and venture capitalists’ hyperventilation about radical change! can’t change any of that. The revolution will be AI, but it will unfold over a much longer timeline than the salivating denizens of Silicon Valley would like.

1. Quote sourced from: https://www.oxfordreference.com/display/10.1093/acref/9780191826719.001.0001/q-oro-ed4-00018679