AI and Big Tech, the Analysis
July and August were marked by sensational news about the future of Artificial Intelligence and the companies related to it. One name among all, Nvidiahe distinguished himself for a wild ride on the stock market. For anyone who is a Nvidia gamer it’s a guarantee: they are historic manufacturers of graphics cards used by video gamesIt is precisely the flexibility and speed of calculation of graphics cards that are part of the revolution of blockchain before and of the IA in these months.
In the summer, however, in parallel with a collapse in the value of the “magnificent 7” by approximately $653 billion, Apple Alphabet Amazon Meta Microsoft Nvidia and Teslasome have started to wonder whether the much-vaunted AI revolution is not yet another bubble. This week, Nvidia alone has “burned” almost $300 billion in value. Let’s do the math.
But I am really “artificial intelligence”? To be fair, we must start from the very concept of intelligence. It was Turing, the inventor of the first computer that helped interpret the German secret code system Enigma, who defined the concept of intelligence. The famous test of Turing (known as the imitation game, for a very good reason as we will see) involves a human judge who, via written chat (and computers did not exist back then, imagine how far ahead Turing was) confronts two subjects whose nature he does not know. One subject is synthetic, the other human. Given the cognitive capacity of the judge at 100, if the two subjects reach or exceed the cognitive value of the judge, it will be clear to the judge to understand which of the two is synthetic.
This is the test explained in an extremely concise way. Turing However, he did not consider the cognitive impairment of return: there are many in the US first and then in the EU, who have difficulty interpreting a 200-word text. And the attention level of the new generations is increasingly oriented towards the understanding of simple concepts, conveyed in audio and video form. This implies that models of imitation of intelligence, such as generative algorithms, are not intelligent, there is no real creativity but only an assembly of elements learned by “imitating” what has been created by human minds. However, in a scenario where the cognitive ability human is significantly decreasing, there is a clear risk: that a “biological judge” in comparing a chat between a cognitively deficient human and a synthetic who has been “trained”, with texts and works of great human minds, could decide that the synthetic is the more “cognitively advanced”; in effect recognizing the synthetic as human and the deficient human as synthetic.
Bubble or real investment? It is important to understand that the so-called AI revolution consists of two elements, one physical and one digital (immaterial).
The physical one includes the large infrastructures of underwater fiber optics and satellites that transport data to data centers. Inside these data centers are computers that host hardware such as Nvidia cards.
The second element is the immaterial one that consists of algorithms (the most famous are the generative ones). With this premise we try to understand if we are in a bubble. The world of algorithms has been pushed, in recent years, by the Algorithm Development generative like ChatGPT and the like. In the hardware world, Nvidia has distinguished itself in the last year as an AI champion, as a hardware (graphics card) company. Several times in the last few months, Nvidia CEO Jensen Huang has been asked what his point of view was on the ROI for the group, linked to the development of the AI industries.
Huang’s theses are highbrow, to put it mildly, and are generally focused on answering with big-picture views. In a recent interview, for example, he explained that “accelerated processing, of course, accelerates applications. It also allows you to do computation on a much larger scale, such as in science, simulations, or database processing. These processes translate directly into lower costs and lower energy consumption. People who invest in infrastructure Nvidia they get immediate returns.” The position of the CEO of Nvidia, reiterated in similar ways in different forums, offers a systemic vision but does not answer the key question: is there a bubble and if there is one, when will it burst?
The calculation application to support the digital transition It is a fact; Nvidia is trying, together with other companies, to answer this question, at least in the Western market. It remains to be seen whether this “hype” as they say in the nerd jargon of Silicon Valley, is justified by real results or is extremely volatile.
Concerns about processing efficiency
In a conference in November 2023, Sam Altman, the founder of Open AI, declared that in 2024 there would be a rapid and unpredictable evolution of AI. James Manyika, Senior Vice President of Technology and Society at Google, showed perplexity, as the Financial Times reports these days. The sales of these summer days of Tech stocks seem to demonstrate a growing perplexity, on the part of operators in the sector and large financial players, towards the immediate success of AI declared by Altman.
Manyika himself explains the so-called successes of AI, thanks to his experience at Google.
He explained to the FT that Transformers, the technology behind large language models (LLM), allow Google Translate to support 243 languages. Google’s AI, Gemini, has the ability to seamlessly process text, images and video. It also allows users to ask increasingly complex questions. Yet the software itself is far from being widely adopted. It is seen more as a curiosity than a truly useful tool. Added to this is the fact that there are many doubts within Google itself. Last year, the godfather of AI, Geoffrey Hinton, resigned from Google, justifying his departure by saying that the technologies he worked on posed a high risk to human society.
Manyika himself, analyzing the first historical cases of AI applications in the world of work, observed that its first applications were not positive. Remembering Covid and the need for companies to optimize work, the first applications of generative AI were (as explained by Manyika) used to reduce human personnel in favor of synthetic computing power.
Just to mention the world of journalism, there are several newspapers that, since post-Covid, are experimenting with generative solutions in place of journalists, hundreds of human resources fired from one day to the next.
Legal issues
While the aspects related to improved performance are the subject of debate among experts, the legal aspects, at least in California, are undergoing a rapid evolution, to the detriment of AI. The “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SB 1047)” passed by the California Senate should set an important precedent for normalizing the development of algorithms (AI in common parlance). While tech billionaires like Musk support this act, other large players in the industry criticize its narrow vision. This California law could set a precedent for the federal government to develop a more conservative set of laws, thus cooling investors’ enthusiasm for the “infinite growth” of the AI industry and pushing them to reduce their financial positions in companies in the sector.
Environmental and energy aspects
One of the topics often overlooked when discussing the digital revolution, especially AI, is energy and consequently environmental. Even the world of finance has noticed the energy demand that is required by AI systems. Goldman Sachs explained in the spring that “on average, a search done with ChatGPT requires 10 times the electricity required to process a common search with Google. The difference in energy demand defines an ocean of changes that the United States, Europe and the rest of the world will have to face, in terms of energy. For years, data centers have had an energy demand with stable growth over the years. Now, however, with the AI revolution, the Goldman Sachs Research estimates that energy demand from data centers will grow by 160% by 2030. Currently, data centers consume 1-2% of the energy produced on Earth worldwide. By the end of this decade, demand will double. Consequently, energy production, preferably at low cost, will have to increase, pushing the search for energy sources that are as cheap as possible”.
If we think about economic forms of energy, obviously we cannot forget that China, which is currently investing massively in AI, while deploying on its soil every year the European equivalent in terms of renewables, has planned to begin decreasing coal-based energy production no earlier than 2030. Simply put, the energy demand of AI risks being satisfied, to a large extent, by the energy exploitation of fossil fuels. Considering that the West, and in its wake the rest of the world, is rightly placing a lot of emphasis on the transition from fossil to sustainable resourcesthe energy demand of this industry could be seen by investors and states as a negative factor in the reflections on sustainable growth.
If the AI bubble were to burst, or at least slow down its growth in the coming months, it is plausible to think that even the exaggerated hype about this new technology could be more moderated, to the advantage of the work activities human and global energy demand, already put into crisis by the Western energy transition.
#billions #burned #stock #market #Nvidia #Meta #Alphabet #artificial #intelligence #turning #point #bubble #analysis