AIconomics...a new Malthusian trap?

In the 18th century the English economist Thomas Robert Malthus established a theory that stated that population growth is exponential, while the growth of resources needed for population sustenance is linear, eventually leading to population decline. This event is called the Malthusian trap, or more scary, catastrophe...potentially leading to societal collapse, or at best a significant reduction in population in a not so nice way(think war, famine, etc...).

Last week Nvidia CEO Jen-Hsun Huang stated "Just like we generate electricity, we're now going to be generating AI...through AI factories running 24/7". Mr. Huang basically sees AI becoming a commodity or rather a resource similar to electricity that will be consumed by any or all...who pay for it. AI as a Service is already pervasive today, subscription and consumption(think "tokens") based charging models are common practice.

AI eats electricity for breakfast, lunch, dinner and the munchies

Paraphrasing(ok, butchering) the brilliant Peter Drucker, nevertheless in this statement lies Huang's first conundrum. AI is projected by the International Energy Agency to consume 1,050 TWh in 2026 — equivalent to the energy demands of Germany, I actually think it will be an order of magnitude more. Getting us closer to the first "trap", will there be enough electricity available for the AI demand? Will the increase in demand surpass supply to the point of which the cost of the electricity resources spikes the price of the AI "resource" to such heights that only the very few will be able to afford the ongoing AI usage?

AI as a technology will become ubiquitous, and probably not even be on the forefront of technology conversations in the near future, but will it be available and affordable to all in the future?

AI as an economy model transformer

The current economic model of many countries in the world is built on supply, demand and for many driven on human labor. Supply, in relation to labor, is the number of workers available in the market, and demand, in relation to labor, is the number of workers that a particular company, industry, or entire economy wishes to employ. Employees are the supply, and employers are the demand.

Now today AI is still very much augmenting the workers, rather than replacing them, but it is highly probable that current jobs done by workers will indeed be displaced by AI, both for knowledge and physical labor work. At the heart of economics is that demand is driven by consumers...and consumers are able to consume because of income generated by labor.

Trap #2: no labor, no income, no AI services

If AI is a resource that will be consumed in the future, similar to electricity, there need to be consumers that are able to pay for this resource, but where will the income for consumers come from? If their source of income(labor) is actually displaced by the very thing they want to consume, then there are no means to consume. This is where the need for an "AIconomics" transformational model is needed i.e. the current and thousands year old economics will need to be revisited and rethought.

AIconomics in the age of autonomous AI

Already we are seeing autonomous AI agents becoming available to consumers, no longer is it just about AI assistants "on demand", but rather within enterprises as within the common household, AI agents are slowly taking over tasks normally performed by human workers as "obvious".

And we all want more...

What will constitute labor in AIconomics, what will generate the income needed to consume AI resources, when our dependency on these AI resources are such that our very way of living and society depends on them being there 24/7, doing what we tasked them to do.

New era economists are already working on finding answers to these questions, but the pace of their thinking may not be at the pace of the evolution towards the "trap"

The Holy Technology Trinity

So is the current evolution and the vision of brilliant people like Mr. Huang taking us towards a dystopian future? I don't think so...

The challenges outlined and the risks for falling into a Malthusian trap are real, but IMHO within AI also lies our path to avoiding this trap.

Three technologies have the promise to help us:

AI, Quantum computing and Fusion power..The holy technology trinity.

Many of the challenges and the pace needed to gain answers actually need AI to help come up with answers. Quantum computing holds the promise to accelerate the "time to answer" from AI to, well, superposition "speed"(the answer is already there, just find the right way to ask)...and finally for the needed power to fuel both the AI and quantum computing(I know supposed to be more efficient, BUT still to be proven at scale) we need Fusion power.

Fusion power, the power source that "keeps on giving" today is still not possible, i.e. we need more energy to produce fusion power than it provides. The answer towards solving this fusion power challenge, many scientists think lies exactly in the AI and quantum computing technology advances and the combination will get us to this "holy grail".

In summary...

Yes, there are many possibilities to fall into the Malthusian trap with the rapid AI developments and vision of the likes of Mr. Huang, but the answer to address and avoid these challenges and traps, lie exactly in the technologies that may lead to the trap.

The technology trinity will be needed to come up with a new AIconomics and possibly future societal model, but in the beginning...it will still be humans who will come up with the right questions and AI assisted answers.

Kindest regards,

Yves.

#AIconomics #Technology_trinity #AI #Quantum_computing #Fusion_power