Mehmet Yurtseven
Mehmet Yurtseven
United Kingdom
Dear copiers and followers, Today I am going to talk about the cost of AI and why I think AI is a wonderful technology but a financial bubble at the same time. AI already delivers real value in three areas. First, it helps people learn faster, especially in coding and technical problem-solving. Second, it helps produce documents, drafts, summaries, and other written material for legal, financial, academic, and educational work. Third, it amplifies existing skills by acting like an assistant that is available at any time. Those are genuine benefits, and they are a major technological step forward. When you ask for the information, when you ask for the interpretation, AI does not find the source and take the information from there. It has a very sophisticated database of the most of the internet, the entire knowledge base of humanity in a sense. It uses this entire database every single time for every single query it receives. The entire knowledge of humanity, which is stored in these models, is going to be very big. They will need lots of computational power to access it. For every single word that these LLMs give you, they go through the entire stack. This is not sustainable. You cannot expect the whole knowledge base to work for every single human being on a $20 subscription indefinitely. My guess is that the industry will eventually become more specialized. Instead of one model trying to be everything for everyone, we may see more targeted systems: technical models for coding and science, professional models for law and finance, educational models for schools, and lighter social or consumer models for everyday use. That would be a more efficient and economically rational direction. In other words, AI may become profitable not by being universal, but by being focused. The infrastructure side also makes me cautious. Data centers, power demand, semiconductor supply, and cooling systems all require huge capital investment. That reminds me of the fiber-optic boom of the 1990s, car boom of the1920s, railroad boom of 1800s. The innovation was absolutely real and transformative every time, but many of the companies that overbuilt the infrastructure did not survive. The technology changed the world; but the investors did not all win. That is basically my position on AI. I think it is one of the most important technologies of all time. I use it, I benefit from it, and I believe it will remain part of the future. This text has been checked and enhanced by AI. But I also think there is a serious risk that today’s enthusiasm is outrunning the underlying economics. So my skepticism is not about the technology itself. It is about whether the current wave of investment will generate returns for the people funding it. In short: AI is real, useful, and transformative. But transformative technologies can still be terrible investments at the peak of the hype cycle. Since our idea is to be part of companies that make money rather than to join the wave and dump the stocks on someone else, I am not planning to invest in these companies anytime soon. I wanted to get this out to you before SpaceX IPO, so you know where I stand in all this. Happy investing, Mehmet
Not investment advice. The author may have financial interests in the mentioned instruments.
1 reply
null
.