The following segment was excerpted from this fund letter.
We have recently received a number of questions about Artificial Intelligence, so we decided it makes sense to spend a paragraph or two explaining it – which we will have to do in terms we understand – so please don’t be insulted by our simplicity.
Much of what is now described as AI is, in our opinion, traditional input-output computer applications, albeit often presented with a much more elegant interface. Just a year ago, companies eagerly inserted the word “blockchain” into every investor presentation. As the term “algorithm” started to gain similar traction, it was quickly replaced with “AI.”
Despite the fact that many people don’t fully grasp what AI entails, some investors aren’t willing to let a little ignorance affect their investment decisions. Given our own limitations on the topic, instead of attempting to provide a detailed technical explanation, we can look to a 40-year-old movie that offers an example of AI in action.
In the 1983 Matthew Broderick film War Games, the North American Air Defense Command (NORAD) utilizes a computer called the War Operation Plan Response (the WOPR) as a part of the U.S. nuclear launch apparatus. In the movie, the WOPR (appropriately pronounced “whopper”) mistakenly interprets a war game simulation as a genuine Soviet nuclear attack and attempts to launch hundreds of actual U.S. nuclear warheads into Russia.
As the countdown to Armageddon approaches, a young Matthew Broderick instructs the WOPR to begin playing itself in tic-tac-toe, which the computer does repeatedly, with each game ending in a tie. The WOPR apparently “learns” that it can’t win at tic-tac-toe and, complete with laughable 1983 movie special effects, begins running hundreds of different nuclear war scenarios – each of which, like the tic-tac-toe games, ends in a tie.
Finally, the WOPR learns that the only winning move in thermonuclear war is not to play the game – and cancels the launch.
This scenario illustrates machine learning, a significant branch of AI focused on algorithms that allow computers to learn from and make decisions without explicit programming.
Machine learning requires massive amounts of computing power, which requires the production of massive quantities of computer chips – an industry that has historically been dominated by Intel (INTC) and Advanced Micro Devices (AMD). Since 1993, Nvidia (NVDA) has specialized in making graphics processing chips, initially designed for video games. As it turns out, these graphics chips can perform concurrent math calculations more efficiently than the AMD or Intel chips, making Nvidia the choice for calculation intensive applications – like cryptocurrency mining. And machine learning/AI.
In the long run (whatever “long” means these days), artificial intelligence will be important because it promises significant advancements in leveraging technology to enhance our decision-making and our quality of life. In the short-term, however, it’s important because the market’s fixation on a few chip companies is masking the health of the broader stock market. That is, the U.S. stock market hasn’t performed nearly as well as the S&P 500 suggests – because the prices of a handful of AI-related stocks have skyrocketed.
As seductive as it might be to rush into the “newest hot thing,” investing based solely on hot trends is often unwise. And it is certainly not a strategy we will be accused of with our most recent purchase: a bank in the Cayman Islands.
© 2024. Moon Capital Management, LLC is a Registered Investment Adviser with the Securities & Exchange Commission. SEC registration does not constitute an endorsement of the firm by the SEC nor does it indicate that the adviser has attained a particular level of skill or ability. SEC file number: 801-49240. |
Editor’s Note: The summary bullets for this article were chosen by Seeking Alpha editors.
Read the full article here