Executive Summary
- Since OpenAI’s ChatGPT was made public just over a year ago, artificial intelligence (AI) has been one of the biggest stories, investment or otherwise, in 2023.
- Not surprisingly, rapid acceleration in the space – and in related spaces such as synthetic biology, robotics, and quantum computing – has generated excitement but also sowed confusion in the general public.
- This paper will review the development of different technologies over time and explore how they have impacted our societies.
- In addition, we explain what makes “general-purpose” technologies like AI different and so potentially disruptive.
- Lastly, we will explore how AI innovation may impact the economy and investment markets.
What is a “general-purpose” technology?
A general-purpose technology (GPT) is a flexible technology, allowing for applications across multiple fields and industries. GPTs tend to proliferate relatively quickly, meaning adoption rates shoot upward as costs fall (see US mobile phone subscriptions, at right1Source: Our World in Data – https://ourworldindata.org/grapher/ict-adoption-per-100-people?facet=metric&country=~USA). GPTs are also enablers, meaning they often spawn and accelerate other tangential technologies. With greater adoption and emerging related technology comes tremendous potential to transform and disrupt the current establishment, leading to significant economic and social changes.
There are dozens of examples of general-purpose technologies through time, many of which are illustrated on the graphic on the following page. However, here are a few more detailed examples.
- Controlled Fire / Cooking – Cooking food allowed for better digestion and absorption of nutrients without expending energy previously used for chewing and digesting raw food. It also freed up time and resources, leaving more time for learning, language development, and passing down of knowledge.
- Written Language – It allowed for the recording of history, the preservation of knowledge, and the spread of ideas across time and space. It also allowed for the creation of more complex systems like mathematics and law.
- Animal Domestication – Domesticated animals increased food supply, provided transportation, assisted with physical labor, and enabled the development of agriculture. These developments meant that hunting and gathering were not necessary, and permanent settlements emerged as a result.
Before delving into AI and how fast technological change is occurring in the present moment, understanding the history of technology may be a useful context. The graphic at right summarizes technological innovation by humans and our ancestors over millions of years. For example, the spiral in the bottom left captures the distant past, innovations that occurred beginning 3.4 million years ago through 12,000 years ago. The section above focuses on advancements from that time through 1800, including the Agricultural Revolution and the first evidence of writing. Lastly, you’ll see the upward-sloping line illustrating 1800 to 2200.
The chart illustrates the acceleration of technological breakthroughs over time. Over two million years passed between the first use of tools and the first use of fire. From then on, breakthroughs began to gradually quicken, as each general purpose technology breakthrough created ripples that helped accelerate the occurrence of the next breakthrough.
For millennia, most of our ancestors lived much the same way as their parents, grandparents, and great-grandparents. However, that began to change as the pace of innovation kept accelerating.
For example, take someone born in the US in the late 1800’s. That person most likely grew up in a home without electricity or running water. If she lived until 1970 (which was more likely given the rise of antibiotics during her life), she would have witnessed the unthinkable. Cars replaced horses as everyday travel, planes replaced trains for long distance transport, and electricity became commonplace as electric lights, radios, and then televisions became household staples. She would have witnessed the birth of computers, the dropping of the first nuclear bomb, the discovery of DNA, and Neil Armstrong walking on the moon in her lifetime.
It is tempting to think of giant technological leaps as being things that happened in the past, but we should expect technological innovation to continue in our lifetimes, as well. Think about how radically different our lives are today because of two relatively recent GPTs – the internet and smartphones. Before those, we were mailing friends letters, ordering Christmas gifts from Sears by handwritten and mail order slips, and piecing multiple maps together to chart our summer vacations.
Is artificial intelligence a general-purpose technology, and if so, what does that mean?
After the term “artificial intelligence” was coined in the mid-1950s, the field experienced several boom-and-bust cycles. Periods of progress were followed by “AI winters” when progress met roadblocks and funding dried up. However, advances in machine learning, powered by increased computing power and the emergence of the internet, brought AI back into the mainstream in the late 1990s. Milestones such as IBM’s Deep Blue beating chess champion Garry Kasparov in 1997, IBM’s Watson winning Jeopardy in 2011, and DeepMind’s AlphaGo defeating the world champion of Go in 2016 began to make the public aware of AI’s potential.2Source: Forbes – https://www.forbes.com/sites/bernardmarr/2018/12/31/the-most-amazing-artificial-intelligence-milestones-so-far/?sh=449d11067753 The release of ChatGPT to the public marked another such milestone.
Mustafa Suleyman, co-founder of DeepMind (AI firm which Google subsequently purchased) and Inflection AI, recently wrote a book called The Coming Wave about the rise of AI and other innovations such as synthetic biology, robotics, and quantum computing. In the book, he argues that AI is absolutely a GPT. He explains that one of the defining characteristics of a GPT is its flexibility. Over the past 25 years, AI has evolved from single-use tasks, such as playing chess, to become a “Swiss Army knife of technologies,” powering everything from facial recognition to vacation planning. For example, see the pace of progress in language and image recognition below:
Unfortunately, technological leaps aren’t all sunshine and rainbows. Suleyman and others argue that while a technology of this magnitude has the potential to improve our quality of life in myriad ways, a host of risks accompanies this rising technological tide. It is the potential risk that prompted Elon Musk to state, “With artificial intelligence, we are summoning the demon.”3Source: PCMag – Elon Musk: Artificial Intelligence Is ‘Summoning the Demon’ | PCMag See some examples of the potential opportunities and risks below:
This is the paradox of AI and related technologies. We have all heard the saying that “a rising tide lifts all boats,” but a rising tide also has the potential to cause significant harm. This is the dilemma we will wrestle with in the Age of AI.
Where does AI stand now?
A lot has happened since OpenAI launched ChatGPT publicly just over a year ago. Large tech firms such as Microsoft (investor in OpenAI and integrated ChatGPT into Bing’s chatbot), Google (launched Bard and recently launched Gemini), and Meta (launched Llama 2 and Meta AI) have all rolled out AI tools, and Apple and Amazon are also active in the space. Large Chinese firms such as Alibaba, Tencent, Baidu, and SenseTime are actively involved in AI, as well.
However, many other tools have been created and launched by smaller players. For example, see the top 50 most visited AI tools through August 2023 below and note how many you do not recognize.4Source: AI Industry Analysis: 50 Most Visited AI Tools and Their 24B+ Traffic Behavior – WriterBuddy
Suleyman describes this dynamic in his book, a future in which power concentrates in large, well-resourced companies creating cutting-edge general AI models, but smaller companies developing specific use-case tech proliferate and become industry disruptors and leaders, as well. Thus far in 2023, investors have witnessed the first part of that theory play out – a handful of US mega-cap technology stocks have easily outpaced the market (see chart on the following page).
For example, Microsoft, Apple, Alphabet, Amazon, Nvidia, and Meta, all top ten holdings in the S&P 500 index, have more than doubled the S&P 500 return year-to-date.5Source: YCharts as of December 13, 2023 In fact, Nvidia is up over 230 percent this year, nearly ten times the index’s return. Much of the excess return for these companies has been driven by investor interest in AI, as these companies are producing chips used in advanced computers, actively launching AI-related tools, and/or integrating AI in their industry-leading technology solutions.
However, we believe 2023 will be remembered as the first chapter in what will ultimately be a decades-long story of AI. For context, consider the quote below from Ian Bremmer, founder and President of Eurasia Group, a political risk research and consulting firm:
“Artificial intelligence will create an entirely new form of globalization, one exponentially faster and more transformative than the globalization unleashed by free global trade and investment in recent decades. Generative AI is already advancing three times faster than Moore’s law, with its capacity doubling every six months. Today’s AI products will look like children’s toys when compared to the tools we’ll access even two years from now. We are about to experience more technological, social, political, and geopolitical change in the next decade than we’ve experienced in the past half-century.”6Source: Ian Bremmer, GZero Daily newsletter, December 13, 2023
Where is AI going?
This is the million-dollar question, and one, quite frankly, that no one can answer with certainty. In Morgan Housel’s book, Same As Ever, he begins a chapter with the following comment, “We are very good at predicting the future, except for the surprises – which tend to be all that matter.” Consider the graphic below, which illustrates the opinions of AI experts across three studies asking when AI will reach human-level intelligence. The consensus tends to be around 2060, but the answers are across the board, ranging well over a hundred years.
The uncertainty related to AI’s timeline isn’t strictly a technical issue, it’s also a human issue. In the face of such a technological leap, we anticipate greater regulatory oversight (as we write this, the European Union is working to adopt the sweeping EU AI Act), job losses, increased competition between superpowers, and bad actors using the technology to advance their agendas along the way.7Source: The EU AI Act: A Comprehensive Regulation of Artificial Intelligence | Fieldfisher As a society, we will wrestle with how to contain the technology to mitigate against the worst outcomes. To Housel’s point, there will be surprises that no one sees coming.
However, we believe in positive surprises, as well. We expect that it will improve our lives in both expected and unexpected ways, much like the internet and smartphones have in the past 25 years. It’s possible the next generation won’t be able to imagine a world without AI.
An interview from over one hundred years ago may prove an analog to our present moment. In January of 1908, the Washington Post ran an article about the future, and in it, they interviewed Thomas Edison. The Post asked Edison, “Is the age of invention passing?”. Remember, the previous half century had seen incredible technological advancements, including the invention of the telephone, electric light, automobile, and airplane. Edison replied incredulously, “Passing? Why it hasn’t started yet. That ought to answer your question.”8Source: Same As Ever by Morgan Housel
If this is a long-term investment theme, how do we get exposure?
To some degree, simply maintaining broad stock market exposure will provide investors with an opportunity to benefit from advancements in AI. Some of the largest companies in the world are deeply involved in the field, so investors will get some exposure by owning market-cap weighted index funds.
However, we believe the outsized returns that we’ve seen the large tech companies enjoy this year will start to widen out to smaller companies that are poised to become industry leaders by harnessing the power of AI. We agree with a recent Goldman Sachs article, which suggests that we’re shifting from the excitement phase to the deployment phase of AI.9Source: https://www.goldmansachs.com/intelligence/pages/ai-poised-to-begin-shifting-from-excitement-to-deployment-in-2024.html
In an effort to identify these future disruptors, we advocate for complementing broad index-based strategies with active management in areas like small-cap and biotech. We believe in investing alongside experienced, ethical, and engaged firms when committing capital to rapidly changing fields. By partnering with such firms, we seek to make wise, measured, and ultimately, profitable investment decisions on your behalf.
As 2023 comes to an end, we wish you all happy holidays and a wonderful New Year. As always, we appreciate the trust you place in our firm as we navigate the challenging questions investing presents, and we are excited to continue our partnership in the New Year and beyond.
Brett Greenfield
Portfolio Manager
References
- 1Source: Our World in Data – https://ourworldindata.org/grapher/ict-adoption-per-100-people?facet=metric&country=~USA
- 2Source: Forbes – https://www.forbes.com/sites/bernardmarr/2018/12/31/the-most-amazing-artificial-intelligence-milestones-so-far/?sh=449d11067753
- 3
- 4
- 5Source: YCharts as of December 13, 2023
- 6Source: Ian Bremmer, GZero Daily newsletter, December 13, 2023
- 7
- 8Source: Same As Ever by Morgan Housel
- 9Source: https://www.goldmansachs.com/intelligence/pages/ai-poised-to-begin-shifting-from-excitement-to-deployment-in-2024.html