Thousands of machines read this sentence before you did. Not that our column receives the same scrutiny as the pronouncements of, say, Janet Yellen, but by virtue of being in FORBES (and Twitter, and Facebook, etc.) , this story (and every other one that you read from almost any online source you care to name) is now fodder for artificial intelligence (AI) machines looking for a “trading edge” in financial markets.
Now of course you may argue - probably correctly – that the verb in that first sentence - read - should be closed by quotation marks. Machines no more “read” than they “think.” The fact remains, however, that however inelegantly or blindly it is being done, inhumanly vast quantities of information are being hoovered up by machines and fed into algorithms to guide financial decisions. Welcome to the brave new world of financial markets.
This fact - that machines are increasingly influential in financial markets - has been kicking around for several years, but its implications haven’t filtered down to many mere humans yet, so we wanted to make the point again. And the best way we know to explain what is going on in financial markets today is through a metaphor.
Where have the Centaurs come from?
In 1997, IBM’s computer Deep Blue beat Garry Kasparov, the world’s best chess player. Kasparov quite rightly hailed this as a human, not a machine, achievement, as humans at IBM had conceived, built and programmed Deep Blue. Indeed, as Kasparov said later, he could take comfort from the fact that Deep Blue was no more ”intelligent” than a programmable alarm clock… but he also admitted to a profound dismay at being beaten by a $10 million alarm clock!
Soon after, Kasparov coined the term centaurs to refer to the human-machine combinations at the heart of what is now known as “free style chess”, where teams of human-machine combinations compete against one another. Similarly, such human-machine teams are at the heart of lot of not merely high-frequency trading but financial markets as a whole.
How to can you fight Centaurs?
What can you do about the centaurs galloping through financial markets? Even if you are an investor rather than a mere speculator or trader, here are four ideas we can suggest:
1. Get better at what humans do best: First, think through your comparative advantages. In general, these will be non-analytical tasks: generating hypotheses, synthesizing diverse ideas, and rooting out the causation behind the correlations that machines working on “Big data” spot. How do you do that? One way is by looking at your day structure to ensure that you absorb yourself in truly strategic questions for at least part of each day. One of us recently redesigned his office to (hopefully!) facilitate just such an approach to information processing, with half the space dedicated to the thoughtful consumption of books and articles rather than one hundred percent dedicated to IT. It also means consuming information from a wide spectrum of sources, and if you are a team, making sure your team is diverse.
2. If you can’t beat’em…: As a Value Investor, we distinguish investing from mere trading, but even Value Investors need not be Amish about AI. We would love, for example, to ask IBM’s Watson to distinguish among Price, Cost and Value! It would be flummoxed (as would many human traders). That said, it would be foolish not to stay abreast of the latest tools available to make better tactical financial decisions. Work to understand the capabilities of machine-assisted investment decision tools. While the tools of many brokerages and independent vendors are geared to trading, you will at least “know your enemy” (and understand markets) better. It is also another window on how other industries might be affected by AI: in other words, can centaurs “jump moats” and change the value of my underlying investments?
3. Look hard at your processes: As Kasparov said in the context of free-style chess in 2010: “Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.” Rather than a mere trading diary or checklists, look at your investment process in light of machine capabilities. In other words, once you understand what machines do better than you do, think about how to incorporate those strengths systematically into your investment process. Then, constantly reexamine that process to improve it.
4. Watch the progress of AI in poker: In a footnote of Constructing Cassandra, our book on intelligence failure, we decry the use of chess as a metaphor for intelligence work: unlike intelligence analysis, chess is a game of perfect information (albeit one with 10^120 possible games). A better metaphor for the real world is poker: the politics that move markets, for example, contain a lot of bluffing. Nevertheless, AI is making rapid progress in Poker. What are the implications? You tell us!
Can you afford to be Amish about AI?
In sum, don’t get stuck in the “AI is good” vs “AI is evil debate”. As an investor, you won’t be completely replaced by a computer anytime soon. Nevertheless, can afford to be "Amish"? Perhaps. But perhaps your future is to become a Centaur, a human who excels at what human do best, but whose excellence is reinforced by partnership with whatever edge machines offer.
