The traditional monetary system was one based upon physical currency. Those times are long gone.
Many fear inflation due to an increase in the number of units of currency. While that might have some applicability, it will not be for long. In this digital age, as software takes over more of the economy, we see that the world of atoms is changing.
The impact is already seen in the realm of bits.
What this means is the old model is going out the window. By the end of the decade, the traditional teaching of Economics will be cooked. Nothing in their theories accounts for what we are about to encounter.
In this article we will look at the new paradigm that is forming.
The New Monetary System: Bandwidth and Compute
I maintain the most misunderstood concept is inflation. This is something routinely apply theories that are no longer valid (if they ever were). In the digital era, we see the rapid advancement of technology disrupting everything.
This includes money.
What will be the currency in the future remains to be seen. For now, the US dollar is dominant. It will remain that way for some time. The network effects are simply too great. Since the overwhelming majority of transactions are digital, the effects of the online world are present.
My guess is the future of money has something to do with energy.
What we have now is a flipping of the traditional paradigm. The idea of too much money chasing too few goods and services is no longer valid. Think about that for a second.
This is a physical world concept. It only applies to the world of atoms. When it comes to the digital world, there is no limit.
For example, how many songs can you download from iTunes or Google Play? Unlike record stores, where there was a physical limit, that does not exist. One can order a billion of the same song from one of the stores and the price, afterwards, would still be the same.
Of course, downloading a billion songs (or the same one repeatedly) is impossible. Even with a bot the servers would be too slow. It might take decades.
Hence the limiting factor. It is not units of currency but, rather, bandwidth.
AI
Consider the idea of chatbots. Let us take this concept a bit further.
Most models are free, at least up to a point. After that, there is a purchase per token. The payment plan might not be presented that way yet that is exactly what is happening.
Here again, what is the limit?
One of the companies behind these models will sell one as much as needed. Most are setting things up on a monthly plan, providing certain compute for X payment. We could break this down into a per token cost.
Let us take the reasoning models. X and OpenAI are a few hundred dollars per month. I haven't looked by I would presume there is a limit to the use. This is very intensive in terms of the compute required, eating up the tokens quickly.
How much "research" could one do? There is an obvious time limit, 168 hours per week. What would the cost be on that, round the clock reasoning?
I would guess it could run into the thousands.
Now multiply that by tens of thousands of people. Certainly, few would ever need to do this but we are establishing a cap.
The key here is we are talking about one single use case. How about the feature of image generation? Again, we have free models which provide a certain limit (one or two per day). What happens after that? A charge is incurred.
Again, someone could easily spend hundreds of dollars if he or she so desired.
Companies are rolling out more compute. Each GPU gets faster compared to the previous generation. Algorithms also help with the efficiency of the system. Overall, the entire stack keeps improving.
This is massively deflationary. That means more money is actually required, not less. It is simply impossible to keep up with the pace. Demand is going through the roof.
We have a system that is now constructed on how fast can things operate. The more output per token, the better. Over time, this increases with regards to price. People get more for less.
The pace of demand, however, is exceeding that. Companies are spending billions to design more advanced models which will entice users into doing more. This is the economic model that many are hanging their hats on.
Will it work? That remains to be seen. What we do know is that output is increasing at an exponential rate. We are dealing with unit of cognition coming from these systems.
That means we are basically operating on a near infinite curve. Unless there is a massive slowdown in advancement, we will keep seeing this increase.
Ergo, there is no way for monetary advancement to outpace this.