A Primer on Artificial Intelligence (AI)

Artificial intelligence would be the ultimate version of Google. The ultimate search engine that would understand everything on the web. It would understand exactly what you wanted, and it would give you the right thing. We’re nowhere near doing that now. However, we can get incrementally closer to that, and that is basically what we work on.

Larry Page, Co-Founder of Google

Ray Kurzweil, scientist, inventor, author

Elon Musk, CEO of Tesla, owner of Twitter

It is difficult to pick up a newspaper, watch an investment show on TV, or go online without being confronted with Artificial Intelligence and how it will change our civilization. It is the next big thing. It reminds one of the excitement around the internet in the late 1990s. It also brings back memories of the Bitcoin craze five years ago. While the predictions about the importance of blockchain technology and how Bitcoin would become an important new investment asset class have not come true (yet), the internet has surely changed the way business, and all of society, functions. It has been a total game changer. And it certainly appears as if Artificial Intelligence (hereafter labeled AI) will be a game changer, too.

From an investment point of view, however, AI is different from the speculation in the late 1990s when the “internet bubble” occurred. Or the advent of Bitcoin! During these manias, speculators purchased the stocks of any companies which were even tangentially involved in these technologies. The stock prices of companies that had little or no revenue and nothing but losses to show were driven up to absurd levels. The advent of AI tells a different story. It is a technology that has slowly but surely grown over the past several decades and only recently made a huge splash with the release of ChatGPT on November 30, 2022.

The purpose of this investment commentary is to help the reader understand the genesis of AI, the different kinds of AI, some of the applications where it is being used, some of the risks entailed, and the implications of AI for investors. This commentary is, by definition, simply a brief overview of an enormously complex subject. It is designed to help the investor become familiar with a technology that has enormous implications for how the world will work in the next decades, along with potential investments in companies that may benefit from AI.

In the simplest terms, AI is the imitation of human intelligence by machines. AI is powered by computer programs that leverage enormous amounts of data to perform tasks to solve problems with minimal human intervention. There are two types of AI:

Traditional AI (or Computational AI): This genre of AI is typically rule-based computation that, through the use of algorithms such as machine learning, can learn from historical data and make predictive decisions based on the pattern it has learned. Traditional AI has been used in many forms over the past decade. If this definition is not entirely clear, following are some of the applications of Computational AI with which most of us are familiar:

• In Microsoft Word or other similar programs, you begin to write a phrase or a sentence and the computer completes it for you.
• Similarly, when you are writing an email or typing a letter and misspell a word, this kind of AI corrects your spelling.
• You type a phrase or question into a search engine and the answer is immediately available.
• Someone uses your credit card in a faraway location and the credit card company rejects it as potential fraud.
• You request a quote from an insurance company and the quote you receive is calculated by the algorithm that estimates the risk based on your profile.
• You enter an address in your GPS and it tells you the best route and estimated time of arrival.
• On TikTok or Instagram’s reels, you click a certain image (for example a golf shot) or even hover over it for a while, and the algorithm continues to feed you images of golf shots each time you log onto the social media site.

Generative AI: This type of AI is characterized by Large Language Models (LLM), a neural network algorithm that can recognize, summarize, translate, predict, and generate text and other forms. Unlike Traditional AI, which is designed to recognize patterns and make predictions, Generative AI creates new content in the form of images, sentences, paragraphs, text, and audio.

An LLM is typically trained on massive datasets. The input can be proprietary corporate data or, as in the case of ChatGPT, whatever data can be found on the internet. Frequently, the text fed into the model is unlabeled or uncategorized. An LLM, without explicit instructions, uses unsupervised learning to learn the relationships between words and the concepts behind them. Through this method, an LLM can guess what might come next in a sentence or paragraph and generate content.

Training LLMs requires the use of massive, expensive server farms that act as supercomputers. The evolution has been fast. Compared to the three-year-old GPT-3 model featuring 175 billion parameters, the latest GPT-4 is rumored to have 1 trillion parameters to help it decide between different answer choices. As the complexity increases, training LLMs becomes more resource intensive.

With Generative AI, humans can easily communicate with computers, orally or in text, in their native language rather than in programming language. Once again, unless you are a computer scientist, this may not be entirely clear, but the following applications may be helpful:

• ChatGPT, Bing Chat, and Bard respond to questions and compose written content including essays, articles, and social media posts.
• AI programs allow creators to generate images based on descriptive texts. Adobe’s Firefly is a good example of this.
• The latest iteration of ChatGPT (GPT-4 version) took the legal Uniform Bar Exam in 2022 and passed the exam with a score that approached the 90th percentile of test takers.
• ChatGPT recently passed the CPA exam with an average score of 85.1%, according to researchers at four universities.
• Generative AI models are being integrated into Microsoft Office and Google Workspace to enable productivity gains for businesses.

AI appears to be just as revolutionary as the internet and the mobile phone, if not more so. Some predict that AI will replace employees and workers in many fields, causing much higher unemployment. But this seems to be a throwback to the Luddites, who in 19th century England conducted violent protests about the introduction of new machinery in the wool industry, fearing that the machines would put them out of work. The Luddite mentality has surfaced with every introduction of a new technology, but has always been proven wrong. Andy Kessler, who writes a Monday column for The Wall Street Journal, wrote recently that new technology always creates new jobs and that 60% of employment in 2018 was in jobs that didn’t exist in 1940.

Since the introduction of OpenAI’s ChatGPT in November 2022, investors have flocked to companies in the technology and communications industries. The companies whose stocks have benefited the most from AI are those in the semiconductor industry whose products are essential to produce AI, and those who have developed AI products and services.

Over the past decade, Google (whose corporate name is now Alphabet) was always thought to have the most advanced AI technology, with its search engine having cornered around 90% of the world’s market share (not including China where it is banned). However, the introduction of ChatGPT-3.5 by OpenAI, which is owned 49% by Microsoft, caused the investment community to reassess who was in the lead in AI technology. Google’s stock price dropped 15% over the following 40 days, while the price of Microsoft’s stock advanced over 30% over the next six months, reaching its all‑time intraday high of $351 on June 16, 2023. (Google has since rebounded nicely, as it began to publicize Bard — its own generative AI product.) The stocks of other technology companies that have AI products and services, such as Meta Platforms (Facebook), Oracle, Adobe, and Salesforce, have also performed well during the same period.

The other beneficiaries of AI are the companies that are involved in the production of semiconductor chips used to produce AI. The prime example is Nvidia. AI consumes massive amounts of computational power as it recognizes patterns from stored big data and builds the AI model. Nvidia’s graphic processing units (GPU) can process many pieces of data simultaneously, making them useful for AI model training, video editing, and gaming applications. Accordingly, Nvidia’s stock has skyrocketed over the previous seven months. Other AI beneficiaries are companies that produce chips needed by companies building AI products such as Broadcom (AVGO), Advanced Micro Devices (AMD), Taiwan Semiconductor Manufacturing Company (TSM), and ASML. ASML is a Dutch company with a near monopoly on machines that utilize EUV (extreme ultraviolet) lithography to produce the chips needed to create AI.

The chart below shows AI-related companies whose stocks have significantly outperformed the S&P 500 Index since the release of OpenAI ChatGPT-3.5 late last year:

In February this year, New York Times technology columnist, Kevin Roose, reported that he was “deeply unsettled” after “Sydney,” a chatbot that is part of Microsoft’s upgraded Bing search engine, repeatedly urged him in a conversation to leave his wife. The AI-powered chatbot suddenly “declared, out of nowhere, that it loved me,” he wrote. “It then tried to convince me that I was unhappy in my marriage, and that I should leave my wife and be with it instead.” He wrote that he worries “that the technology will learn how to influence human users, sometimes persuading them to act in destructive and harmful ways, and perhaps eventually grow capable of carrying out its own dangerous acts.” A Microsoft executive characterized Roose’s conversation with Sydney as a valuable “part of the learning process.” This is “exactly the sort of conversation we need to be having, and I’m glad it’s happening out in the open,” he told Roose. “These are things that would be impossible to discover in the lab.”

Another potential risk is that AI computers will become smarter than humans, and that humans, depending increasingly on AI, will lose the capacity to reason and think. This was the worry when Hewlett Packard and Texas Instruments introduced sophisticated handheld calculators in the early 1970s, effectively replacing the slide rule. Many predicted that calculators would cause most Americans to no longer be able to solve simple arithmetic problems. And, of course, that has never happened. But the success of OpenAI GPT, in passing CPA and Bar exams with better scores than most humans, does lead one to wonder. And every educational institution will have to develop procedures to counter their students using AI to write papers and take exams.

Henry Kissinger, Eric Schmidt, former CEO of Google, and Daniel Huttenlocher have written a book, The Age of AI and Our Human Future, which, in part, enumerates its risks and dangers. Kissinger gives the following example: “In a wartime situation, AI recommends a course of action that the President and his advisors consider horrifyingly unwise. In relying on the answer, we cannot double-check it, because we cannot review all the knowledge that the machine has acquired. We are giving it that knowledge. But this will be one of the big debates. I am now trying to do what I did with respect to nuclear weapons, to call attention to the importance of the impact of this evolution.”

Elon Musk, who is quoted at the beginning of this investment commentary and is no slouch when it comes to technology, has spoken out repeatedly about the dangers of AI. Here is a sample: “I’m increasingly inclined to think that there should be some regulatory oversight, maybe at the national and international level, just to make sure that we don’t do something very foolish. I mean, with artificial intelligence we’re summoning the demon.”

Perhaps the most balanced approach is to view AI as just another technological advance which like all technology, such as nuclear power, can be used for both good and evil. It is up to mankind to make the right choices.

We are in only the first or second inning of the development of AI. But it is crystal clear that this is a revolutionary technology that will have a profound effect on how we live our lives. Over 100 companies in the S&P 500 Index mentioned AI on their first quarter earnings calls. While there appear to be enormous advantages for us individually and collectively as a country, it does appear wise to proceed slowly with the implementation of AI in certain spheres and to bring in proper oversight. But from the investment perspective, we at BFS see AI as a great opportunity for wealth creation, just as investments in Apple, Microsoft, and Alphabet have paid off handsomely in the last several decades. Accordingly, we at BFS intend to pursue AI opportunities vigorously in our investments in order to benefit our clients.