Technology adoption rate

Existing attempts to describe the history of technology were dominated by story telling and by analyzing the context of innovation. A certain person has invented a new machine at a certain year and this new technology has influenced the society. For example the steam engine has started the industrial revolution while the computer has powered the digital age.

What was missing is a mathematical equation to predict how big the influence of a certain technology is and how long it takes until an innovation will become available in the mainstream. One possible attempt for a mathematical description of new technology is the "technology adoption rate" which should be explained next.

The adaption rate measures how many people using a technology at which year. A good example are the number of internet users [1]:

1993, 14 Mio
1994, 23 Mio 
1995, 45 Mio
1996, 78 Mio
1997, 121 
1998, 189
1999, 282
2000, 415
2001, 502
2002, 665
2003, 781

This table shows, that the in the 1990s only small percentage of the overall population had access to the World wide web and it took some years until around 2003 to reach a larger audience. In other words, that internet wasn't a success from day 1 but it took around 10 years until the mainstream audience was interested in this technology.

The technology adoption rate measures this timespan, in case of the internet, the timespan was roughly 10 years, compared to the adoption rate of the steam engine this was very fast because the steam engine took around 100 years until a larger amount of people had access to the technology.

The advantage of the adoption rate is, that it allows to compare different technology on a mathematical scale. Instead of describing in text what a steam engine is or what the impact of the color television was, its enough to compare only the years until a new innovation has attracted the masses. The general trend is, that the adoption rate has become faster. After the invention of the Internet the next powerful technology was the introduction of the smartphone. Similar to the internet this technology has reached hundred of millions of users. The difference is, that the adoption rate for the smartphone was only 7 years compared to 10 years for the internet.

In case of the chatgpt website, the adoption was even faster than for the smartphone. The adaption rate for chatgpt was only 4 years. A possible explanation why newer technology is available faster has to do with the requirements before new technology can be deployed. Before it was possible to connect every household to the electrical grid, there was a need to build the entire grid. Hundreds of kilometers with cables were needed all over the world to connect every city and every village. Building such an infrastructure is expensive and takes many decades. The result was that the adoption rate for electricity was around 50 years.

In case of chatgpt the precondition was more comfortable. Thanks to the computer boom, the internet boom and the smartphone boom, all the households were already equipped with modern computer technology including access to electricity and the internet backbone. The only new innovation was the activation of a single website which gives all the world access to artificial intelligence. There was no need to deploy new cable or buy another hardware, but existing devices are powerful enough to render a large language model frontend in the browser.

From a mathematical standpoint an intelligence explosion is the observation that the technology adoption rate has been reduced. A faster adoption rate means, that a product or service is available in a shorter timeframe which maximizes the impact for the society. In stead of question in which year a certain AI product is available for beta testers and engineers the question is, how long does it take until the product is available for the masses.

Compared to the adoption rate of AI and the internet, the adoption rate for the Computer was much longer. The estimation is that it took 20 years upto 25 years until the computer was available for 50% of the population. Technically the first home computer, the Apple II, was introduced in 1977. But this doesn't mean, that in 1977 all the households in the U.S. were equipped with such technology. Even in the late 1980s most households had no access to a computer. It took until the mid 1990s, before a larger population was using a computer in their home. There are multiple reasons available why the timespan was so long. Technically the microchip was available in 1977 and technically all the households in the U.S. had access to electricity. The problem was that in the beginning homecomputer were very expensive and they were complicated to use. During the 1980s computers were only a niche products for computer nerds and for scientists but not for the mainstream population.

With the advent of the graphical user system, namely Windows 3.0 the situation changed drastically. The amount of annual sold PC skyrocketed in the 1990s which is equal to a high adoption rate. IN the mid 1990s, computer technology was available in most households and was used routinely for entertainment, work and education.

The advantage of slow adoption rate of 20 years upto 25 years is, that the population can adapt to new technology over a longer timespan. Even if the computer was perceived as revolutionary technology it was not able to change the society drastically. Most part of the life were working great without Apple or IBM computers, and the dominant technology in private homes were the color television, the printed book and the radio.

In contrast the situation for chatgpt is the opposite. The adoption rate of AI Technology is faster which means, it affects more people in a shorter timespan. This indicates that the population gives chatgpt a high priority. Lots of efforts in spend hours and spend money are redirected to artificial Intelligence. The faster adoption rate for AI proofs that the singularity theory is valid. That means, there is an intelligence explosion and that new technology is introduced faster than in the past. Perhaps it makes sense to explain the term "intelligence explosion".

In contrast to common assumption, intelligence explosion has nothing to do with an AI Supercomputer which becomes more intelligent, but the term is referencing to technology in general. Its about technology like the printed book, the internet, chatbots or robots which are introduced to the world with a high adoption rate. Intelligence explosion means, that these technology is available for everyone and that every year, new technology is added. There is not a single generative AI available but dozens of them like: midjourney, chatgpt, gemini, dall-e, copilot, deepseek. So the term intelligence explosion means simply that the collection of all the ai technology grows quickly.

Even if chatgpr has become a great success, there was some limiting factors available in the early beginning. In the first month before the website was seen as a breakthrough technology, many users in the world didn't even know that chatgpt exists. In the first months it was sees as website for computer science students in the english speaking world. In addition the chatgpt server was not able to fulfill all requests. It took some months until the problems were solved and the brandname was popular in the mainstream population.

[1] Internet users worldwide, https://www.internetlivestats.com/internet-users/