Old Model, New Tricks

This week marks the 85th anniversary of the state-sanctioned murder of the Soviet economist Nikolai Kondratiev. Kondratiev (sometimes spelt as Kondratieff) was a noted economist whose insights into long-term innovation cycles had angered Stalin because they contradicted Soviet dogma on many key aspects of economic policy. Kondatiev’s chief sin was to challenge the idea that the Soviet economy would continuously expand under central planning. He believed that periods of downturn were inevitable, and observed what he believed were long-term cyclical patterns in commodity prices and economic activity.

Kondratiev advocated for judicious analysis of empirical data rather than ideologically driven policy at a time when anything diverging from the Party line was tantamount to heresy. Most damningly, the economist highlighted risks of Stalin’s unsustainable crash programs following his extensive analysis of historical data on prices of items like wheat, copper, and other raw materials. His in-depth analysis led him to believe that prices did not simply move randomly, but appeared to follow wave-like cycles lasting approximately 50 years from peak-to-peak. In his theory, Kondratiev described how prices rose and fell with periods of economic expansion giving way to downturns before rising again.

For these crimes, Kondratiev was murdered in a secluded wood known as Kommunarka. Kommunarka was the former dacha—or summer home—of Genrikh Yagoda, a former Soviet police chief, and it was located (as many dachas were) on the outskirts of Moscow. Unlike all the other dachas, however, Kommunarka was also the graveyard of thousands of individuals murdered on Stalin's orders, former elite and intellectuals who had been branded “enemies of the people” as part of the Great Purge.

Before his execution, Kondratiev identified two full cycles in the data from 1770 to 1896. Each cycle—which would become known as Kondratiev Waves (or K-Waves)—consisted of an upswing in prices and growth lasting about 25 years followed by a downward phase of declining prices and depressed economic conditions for another 25 years. He believed that these “long waves” of economic activity were driven by technological innovation and the permeation of new technologies throughout the economy. Major technological breakthroughs, he theorised, drove periods of increased productivity, profit, and investment. However, eventually the new technology would fully diffuse through the economy, profits would fall, and a downturn would commence until the next wave of innovation emerged.

The first wave Kondratiev identified occurred from the 1770s to the 1820s, and emerged in Cromford, Britain, where a man named Sir Richard Arkwright pioneered the development of his frame cotton-spinning machine, thereby inventing manufacturing in the process. Early mechanisation would give way to the steam engine, the workhorse of the Industrial Revolution (the second wave). The steam engine rapidly increased production and was the underlying basis for key innovations such as steam-powered textile factories, iron production processes, and steam locomotives.

Such developments are categorised as "General-Purpose Technologies" or GPTs (not to be confused with the famous AI kind, more about which in a moment), and were generally thought to be greater in sum than they were in their individual parts. According to Kondratiev, GPTs are transformative innovations that have a broad and profound impact on multiple sectors of the economy and society as a whole. Simultaneously, the waves create or redefine infrastructures and in doing so, a new paradigm of techno-economic innovation principles are set.

In the instance of the first wave, we see by the establishment of the factory as the centre of production, mechanisation, the importance of productivity, and the existence of local networks. Such developments and societal changes lay the foundations of the next wave (in this case, the development of steam power and railways), which in turn opens up the opportunity of the next wave, and so forth. Since Kondatiev’s death, multiple waves have been identified. Theories vary, and sometimes overlap, but the general consensus of those who followed Kondratiev is that the discovery of long economic cycles challenges the concept that development happens in one singular upward trajectory.

Perhaps the best-known theorist to build on Kondratiev’s work is the economist Carlota Perez. Perez expounded on Kondratiev’s theory of long waves and developed a comprehensive framework analysing the dynamics of technological revolutions over the past 250 years and their socioeconomic impacts. Her extensive research examined the technological, financial, and institutional factors interacting during great surges of development. In Perez’s view, we are still undergoing the fifth Kondratiev wave.

Others are not so sure. In his 2018 book, Navigating the Tech Storm, the futurist Nicklas Bergman wrote that we may already be seeing the emergence of a potential sixth wave. This one, Bergman argues, would cluster around the BANG technologies (Bits, Atoms, Neurons, and Genes). The “N” here refers to technologies which use neuroscience to further technology, also known as AI. Five years on, I believe Bergman was on the mark. Not just AI, but specifically Generative AI, will be the key general purpose technology of this next wave.

There are many reasons for thinking so. According to Perez’s theory, each wave goes through four financial stages. The first is the irruption phase, when there is intense funding of innovation in new technologies, and clusters of new revolutionary inventions appear, new industries are established, and the construction of new infrastructure begins. 2023 is already a record year for investment in generative AI startups, with equity funding topping $US14.1 billion across 86 deals. At the same time, entire industries, from my own business to drug discovery, are being transformed en masse by this technology.

Just last week, Google DeepMind told the world that it had developed a new AI language model, AlphaMissense, capable of predicting whether or not genetic mutations in proteins are likely to be harmless or disease-causing. The language model—trained on protein sequences—analysed nearly 71 million possible genetic mutations and categorised 89 percent of them with high accuracy, outperforming anything that had come before. In this sense then, GenAI does not just represent an incremental improvement on previous AI technologies.

We are entering an age in which content is not curated for individuals but created for them; drug creation is not a case of trial and error but an accelerated process of targeted molecular generation and simulation; AI agents dynamically synthesise and execute plans or tasks toward goals rather than await our input like “traditional” AI systems have done in the past. Then there is the central premise of Kondratiev waves—diminished labour costs. Conversational chatbots alone—just one facet of many when it comes to GenAI—are set to reduce labour costs by some $80 billion by 2026. All these things point to us beginning to enter the next stage of the cycle.

However, whether or not this is the beginning of a new wave remains a question to which we will not know the answer until that answer arrives. But the sense that something is starting is palpable. From the school governors I've spoken to designing new syllabuses around AI, to film directors experimenting with photorealistic worlds via neural radiance fields (NeRFs), to Vatican priests advising companies on their AI ethics, it feels as though there is a gathering consensus that we are on our way to whatever comes next. Breakthroughs in AI, robotics, biotechnology, and renewable energy all point to a new techno-economic paradigm emerging. Government initiatives, demographic shifts, and inter-state competition, meanwhile, all seem to anticipate significant discontinuities on the horizon.

It is also true that the continued expansion of underlying technologies such as semiconductors and lithium batteries suggest that some of the defining technologies, industries, and innovations of this current wave cycle have yet to fully crystallise. It may be that we are still in the fifth wave cycle. Even so, only a year ago, we would have put GenAI, AR/VR, and quantum computing in their embryonic stages. But in June, IBM used a quantum computer to solve a problem that had until then stumped the leading classical methods. That same month, Apple announced their Vision Pro headset. A few months prior, I had founded the world’s first GenAI consultancy, quickly followed by many others. The pace of change is accelerating, so perhaps the speed with which new waves arise will too.

On the anniversary of his death, we await Kondratiev’s sixth wave. For our understanding of the ongoing cyclical passage of progress its arrival represents, we must thank the man whose willingness to confront ideologues in the name of science cost him his life.