The era of ‘moving history’ One economist responded: ‘radical agnosticism’

There is a speed faster than the speed of evolution of artificial intelligence, i.e. sound or light: The speed of generating new artificial intelligence experts And its results in the most diverse fields. This phrase is in one of the memes that have been circulating in recent weeks. Keep up with the discussion about ChatGPT4 and generative AI in general. In another, Homer Simpson declares himself a “communications specialist,” hides behind a private soldier, and poses as an “artificial intelligence specialist.”

Argentine creative traveled weeks ago to South by Southwest (SXSW) Austin, the world’s largest innovation festival (with nearly 300,000 visitors) commented on the dismay it made him see how Generative AI became a kind of “lightning rod” that monopolized all keynote presentations: “Nobody could escape it, and there were better or worse metaphors, different degrees of depth, but one can’t help but think that no matter how genius they are, they’re talking about a phenomenon that only started unfolding three months ago,” he explained.

The American economist Tyler Cowen offered a similar warning, writing, among his endless activities (he reads several books a week), comments on the very popular blog. Marginal Revolution. Meanwhile, the miserable teacher (the majority) or the optimist and in an age of objective “radical uncertainty”, Quinn proposes a logical position: that of “radical agnosticism”.

The truth is that no one at the beginning of the printing press had any idea what changes this invention would bring about. At the dawn of the fossil fuel era, no one had any idea of ​​the coming transformations. No one is good at predicting the medium or long-term effects of radical technological changes. Nobody, neither you, nor Sam Altman (leader of OpenIA), nor the next-door neighbor, Quinn says, is in conversation with LA NACION.

“ChatGTP’s penetration speed was the highest ever for a new service: it went from 1 million to 100 million users in 60 days”

“So when someone predicts an existential dystopian scenario with artificial intelligence, I don’t think that counterarguing on their own terms is an appropriate response. Radical Agnosticism,” the economist continues.

For Quinn, it’s decades past kind of “story bubble”, With relative stability, without major wars or too radical changes. Our minds are not prepared to live in a ‘moving history’, as has been the case with the majority of human evolution, and this generates an enormous amount of uncertainty. We are used to thinking that we are sailing on a sea of ​​turbulent waters, but that at some point we will reach a different port than the one we started in, but with calmer waves. This is wrong because The waves will be bigger and more intense.

Just as generative AI has completely supplanted Web3 (it’s believed to come along with decentralization) as the backbone of innovation festival talks, something similar has happened with controversy among economists.

Nobel Prize in Economics Paul Krugman made a noise by emphatically asserting that artificial intelligence will not have a significant impact on the economy of developed countries, at least for a decade. Krugman’s argument is that productivity tends to lag with the advent of technologies, because companies have to adapt their new processes, and that takes time.

The classic example is electrification or PC sizing: there is a long period in which companies are exploring how to get the best out of what is new and in which infrastructure is being built. Electricity was introduced in the United States in 1892 and the impact on productivity began to be felt from the 1920s, for example.

“Away from ‘narrow AI’, which is focused on specific tasks, the current goal is to generalize its use in different tasks and create new concepts.”

But not everyone agrees with Krugman’s There are more economists who say that “this time is different”. because? First, since generative AI is a technology that is already on companies’ doorsteps, no new infrastructure is needed to deploy it (or at least, there is no need for any very significant infrastructure).

The best example of this ChatGPT’s oft-cited transmission speed, the highest ever recorded for a new service, Which went from 1 million to 100 million regular users in just 60 days.

Another strong reason put forward by those who argue with Krugman is this Companies are already convinced they have to jump headlong into this new ocean. There’s no “wait to see” like there was last year with the metaverse or Web3, where the dilemma was played on expectations.

Here’s a “real-time future” that’s really happening, the typical enthusiasm cycle (“Hype Cycle,” from consulting firm Gartner) has been compressed and the dynamics look more like “The ketchup effect” Cited a few weeks ago in this section: When there’s a little sauce left in the pot, we tap on the bottom and everything comes out at once, making a mess. This metaphor belongs to Microsoft Vice President Jun Maeda: “Artificial intelligence has been around for a long time. And, Just as we would with ketchup, we’d shake the bottle, and see if anything came out. Now everything suddenly came out “It fell all over and we’re badly stained by the AI,” said Maeda.

Similarly, the discussion among economists is bogged down not only because of the speed of change that is taking place but also because of the difficulties of measurement and lack of parameters.

This was not always the case. For example, in the 1980s and 1990s, chess became a perfect mirror in AI studies, and it is relatively easy to measure: well-defined rules that require a complex computational structure, but with a limited number of possibilities.

Until two years ago AI efforts have focused on developing systems that will excel at a specific task: Play chess, image recognition, language translation, etc. These models are known as “narrow AI”, or narrow. The current goal is raised, and for the first time since the introduction of the concept of artificial intelligence in 1959, It generalizes its use in various tasks and creates new concepts: Artificial General Intelligence (“AGI”).

The playing field is not only in constant motion, it is also getting wider. Artificial intelligence in its new version is what experts call innovation “General Purpose Technology”, So was electricity and the mass use of a personal computer or the Internet at the time. Here the cultural, social, work-related, and existential aspects come into play, reinforcing the “radical uncertainty” to which Quinn proposes, in response, “radical agnosticism”.