How and where artificial intelligence will expand
All the next paradigm shifts that artificial intelligence will bring. Analysis by Paul Wick, manager of the Threadneedle (Lux) Global Technology fund of Columbia Threadneedle Investments
We have been trying to replicate human intelligence in computer systems for over 30 years. The interesting aspect of this process today is the confluence of three fundamental elements: computing power, connectivity and storage. Looking at processors, we went from millions of transistors to tens of millions, then hundreds and finally billions: thus computing power increased exponentially. In terms of connectivity, today it is possible to purchase switches capable of supporting 25.6 terabits per second and 51.2 terabit switches will be available as early as next year. Finally, at the same time, we have witnessed an increase in memory capacity.
AI has been a hot topic for several years now, but last year saw a turning point when companies finally had the computing power and data needed to make the most of the performance of deep neural networks. In late November 2022, OpenAI released ChatGPT for public use, receiving extensive media coverage and attracting investor interest. ChatGPT itself opened the curtain, bringing everyone's attention back to Artificial Intelligence: just a week or two following the release of the application, it was discovered that Microsoft had invested ten billion dollars to acquire a controlling stake in OpenAI, raising many doubts. AI may seem new, but that's not the case for us; it is a sector that we have been studying and investing in for many years, with the help of Columbia Threadneedle Investments' large team of analysts and researchers dedicated to the technology sector.
Internet, cloud computing and mobile devices
We believe that the advent of generative AI (and AI in general) is set to compete with the advent of the Internet and mobile computing as significant trends in the technology market. With respect to these two areas, however, AI is not necessarily a delivery driver. Think about the Internet and mobile devices: they allow you to play, buy and sell online, view advertisements, etc. All innovations that represented great developments both in terms of distribution and new technologies. But generative AI, in our opinion, will be a revolutionary tool, making the Internet easier to use and much more powerful. The latter will constitute the next transition platform. After PCs, the Internet, the mobile era and cloud computing, we are now witnessing the exploit of AI. Similar to some of these technologies, which have had a lasting influence on the economy, we expect AI to have a global economic impact of approximately $7 trillion over a 10-year timeframe.
Alongside purely technological developments, the speed of innovation of AI-related companies is also impressive: the frequency of new product releases is only increasing, encouraging the creation of ever better products for end users. Microsoft, for example, has released several “co-pilots” that could implement productivity in code generation. Other areas where companies could use AI include customer service, IT support services, content creation, fraud detection, supply chain optimization and predictive maintenance.
The advent of AI is also redesigning the ways in which patients are diagnosed, treated and monitored in the healthcare sector; a front where Nvidia is at the forefront. Additionally, large language models should help accelerate the discovery of life-saving drugs, and over time we foresee more advanced use cases in autonomous driving and robotics.
A rapidly evolving technology
Already today, the areas of concrete application of AI are obviously numerous (Figure 1). But, in our opinion, the most innovative development has not yet occurred. We are convinced that one day someone, perhaps a young university student, will ask themselves "How can we exploit all this?".
It is certainly difficult to predict the speed of change and technological innovation, but so far it can be said that this has been faster than most investors expected: ChatGPT reached 100 million users in a very short time and Nvidia recently presented quarterly results with guidance anticipated by a few years. Furthermore, we expect more new discoveries to follow as the technology evolves.
This is a machine translation from Italian language of a post published on Start Magazine at the URL https://www.startmag.it/innovazione/come-e-dove-si-espandera-lintelligenza-artificiale/ on Sat, 21 Oct 2023 05:30:45 +0000.