blog podcast

AI and Change

AI has been a new development in society that caught the public’s eye because of OpenAI’s smart marketing and rapid release of multiple impressive language models. It didn’t take long until all possible future scenarios were being played through. This text is my take on why AI won’t be as revolutionary as many think, at least not in the sense it’s being predicted by many, at least not the large language models. I might very well be wrong, still it’s such a privilege to be allowed to think and express yourself that it would be a waste not to.

Historically there has been some great products that have been bought by big corporations and that have then slowly deteriorated. One example is Skype that after being sold to Microsoft ended up being integrated with all kinds of weird systems. The consequences were that the stability dropped and the smoothness of the product got lost.

The reason for this is not because Microsoft lack smart people, it has loads of them, and they are well paid. The thing is that all of these smart people are not paid to give users the most valuable tool. They are paid to generate revenue. Even though these goals sometimes overlap, companies don’t develop things that aren’t part of their strategy to advance their own interests.

There is an intrinsic law to how products develop: they develop in a way to forwards the interests of the companies that created them. When you apply this logic to AI you’ll realize that commercial AI will forward the interests of the companies that is creating it. Take for example Google, the growth of AI is threatening their business model with the search engine. They need to find a way to move development away from disrupting search engines. Take Microsoft, they are relatively safe, but if AI makes single people very productive it would threaten the comparative advantage Microsoft has by having loads of money and a huge workforce to solve problems. Both of these companies are developing AI, but you can be pretty sure that at least for the products they’ll make available from AI, they’ll do their best to not hurt any other parts of their business.

What is the greatest opportunity with AI? In my thinking it’s the ability to influence people’s minds. You get a tool that people ask all kinds of problems that they have, and by influencing these answers you get to partly control the way people observe and think of the world. One should remember that the large language models like ChatGPT etc. are not reasoning, they are just statistically generating text depending on the words you have written. The flavor and content of this text is not based on some sentient intelligence (yet), it’s based on whatever texts you fed the model with when you did training. This sounds like something that could have a huge impact on society, right? Yes but one should remember that this is not different from Google at the moment. It’s also a place people go to seek information and ask questions, and by influencing the search results it’s already possible to influence what information is being consumed by large portions of the world.

In summary regarding AI. I don’t believe big companies will have an interest in creating the most useful version of AI possible and I don’t think AI introduces something new that would give its creators unchecked power because of influencing people. After all, what many young entrepreneurs have already learned, having the ideas (which is what AI helps you with) is easy, what’s difficult is to make them a reality.