It looks like these are tough times for Silicon Valley tech companies working in the artificial intelligence space. This comes as more industry investors doubt AI’s ability to deliver the massive returns they seek. Western AI shares have fallen 15% since peaking last month.
Image source: Copilot
Against this backdrop, more and more experts are questioning the limitations of large language models (LLMs), which are the basis of generative algorithms such as ChatGPT. Big tech companies have spent tens of billions of dollars creating LLMs and various AI tools based on them, and projections of future spending look even more impressive. Despite this, the most recent statistics indicate that only 4.8% of US companies are using AI to create products and services, down from 5.4% at the start of the year. The number of companies using AI is expected to double over the next year, but this is not enough.
What is happening with AI technology can be described by the term “hype cycle,” which was popularized by Gartner and is well known in Silicon Valley. After the launch of any promising technology, a period of irrational euphoria and excessive investment begins, after which a serious decline occurs. This leads to unrest among market participants and investors due to the slow adoption of technology and difficulties in making a profit. However, eventually the technology begins to be used en masse, facilitated by huge investments in the initial stage and the infrastructure created by the investments.
The initial stage of technology development, accompanied by huge investments, is certainly useful, as past examples prove. In the 19th century, Britain was gripped by a railway fever, and in pursuit of high profits, investors poured huge amounts of money into the industry, thereby creating a stock bubble. Then came the collapse, after which the railway companies, using the funds raised at the first stage, nevertheless built a railway, connecting numerous settlements with it. This helped transform the economy and contributed to the development of rail transport. The history of the Internet followed a similar scenario. In the 90s, new technology caused euphoria, and futurists predicted that in just a couple of years people would be shopping online. In 2000, the market crashed, leading to the collapse of many Internet companies around the world. However, by that time, telecommunications companies had already invested billions of dollars in laying fiber optic cables, which later became the basis of the infrastructure for the modern Internet.
Although the AI field has not yet experienced a similar collapse, some experts are confident in the future global dominance of artificial intelligence. “The future of AI will be like any other technology. There will be expensive construction of giant infrastructure, then a huge decline as people realize they don’t know how to use AI productively, and then a slow revival as they figure it out,” says economic commentator Noah Smith.
It is possible that the AI field will develop differently. For several decades now, artificial intelligence has been going through periods of boom and bust, accompanied by rises and falls in the activity of researchers and investors. In the 60s, there was a lot of excitement around AI, including the emergence of Eliza, the first version of a chatbot. Then there was a lull for several decades until the advent of generative neural networks in 2020.
There are many other technologies that have managed to avoid the hype cycle. For example, cloud technologies from their emergence to mass distribution have followed a fairly straight line without significant ups and downs. Solar energy and social media have evolved in similar ways. At the same time, there are many technologies that, having overcome the first stage of hype, reached a point of decline, but have not entered the lives of people around the world, at least not yet. A striking example of such technologies can be considered web3 or carbon nanotubes. In the past, it was also predicted that every person would have a 3D printer at home and 3D printing technology would become a part of everyday life, but this also did not happen.
The study showed that only a small part of innovative technologies undergo development through the “hype cycle.” Only about a fifth move from innovation to rapid growth, decline, and then widespread adoption. Many technologies begin to be used widely without such leaps. In other cases, technologies go from boom to bust and never come back.
When it comes to artificial intelligence, it can still create a revolution if some of the tech giants make a breakthrough in this area, and businesses and companies can realize the benefits offered by AI-based technologies. One of the biggest challenges for developers right now is to show that AI actually has something to offer the real economy.