Acceleration.
It is a word used often in the technology world, especially when it comes to AI. This can, then, translate into economic as evidenced by discussions about the economic singularity.
Some are projecting that, in a few years, the United States will be able to hit a 10% annual growth rate in GDP. This is still a forecast which could completely bomb. However, some are realizing the potential of AI and what it can bring to an economy.
All of this entails scaling. The question is whether this will continue at the present pace or fall off. Many have stated that a decline in inevitable but, thus far, that hasn't happened.
I am on record as predicting the economic singularity happening in the 2030s. My personal definition is a 30% annual growth rate for the global economy.
Here is where exponentials are needed.
The pace of AI acceleration is impossible to quantify. We have the amount of compute entering the market. Tied to this is the efficiency of the GPUs. Then we have data growth, both organic and synthetic.
Algorithm improvement is taking place. On top of this, we have AI working on itself, creating a feedback system of improvement without human involvement.
All of this equates to a world which might be radically different by the end of this decade (or sooner).
Of course, there is a chance this is all blown out of proportion. People like Elon Musk and Sam Altman are talking big games. They are also dropping large sums of money.
This is mirrored by many in the technology world. Is it hype? To me, the answer comes down to scaling.
If AI is scaling at a pace that many are forecasting, then we are dealing world changing technology.
What we are really talking about is the scaling of automation. This is the same process whether we are focusing upon bits (digital world) or atoms. With the latter, robotics are incorporated.
A hot topic is the progress of robotaxis by companies such as Tesla or Waymo. The former has been talking about this for almost a decade without much market impact. That could be changing.
Right now, the number of Tesla vehicles being used in the capacity are miniscule. There are only a few hundred.
This is fuel for the bears who claim this is overhyped (by Elon and the company). What is being ignored is the pace of scaling. Each week more vehicles are being added, at an increasing pace. This is the key.
Simple mathematical extraction leads us to conclude that, even starting with a couple hundred, millions can be reach if the pace of expansion grows over a long enough period of time.
People tend to think in linear terms. The key is remember is the exponential that often comes from convergence. In this instances, we can use the number of vehicles being added to a market along with the increase in the number of cities.
It is a similar process that will be applicable to humanoid robots when they start to roll off the assembly line. At first, it will be an insignificant number.
Convergence is the most powerful force in technology. As stated over the years, the smartphone was really the merging of mobile, compute, and the Internet. It was this breakthrough that altered society. Nothing was really new in terms of the technology (at this level).
AI is, by design, an exponential. There are a number of layers in the stack that are all seeing improvement. Collectively that creates outsized advancement, where progress is measured in months as compared to years. It is likely that this will shift to weeks and eventually days.
The term "bottleneck" is thrown around. These have to be eliminated to scale. By definition, a bottleneck impedes the progress. As long as innovation allows for the surpassing of these, scaling will continue.
If one does present a global issue, then advancement is stalled.
