Why brains are more reliable than machines

Man vs. machine. It is an old, well-trodden debate. There is little question that technological advances have forever changed and improved our lives in ways big and small. But now, some are predicting that machines and their algorithms?

Set of rules that can parse data and automatically decide what to buy and sell. will supplant humans in the investment firm of the future.

I will be the first to admit that technology helps me do my job better. I still remember my days of skimming through S&P tear sheets to find new ideas. Now a phone in my pocket can provide everything from tick-by-tick stock quotes to pages upon pages of corporate filings.

But here is the thing: data alone isn’t enough. As Albert Einstein once supposedly said, “Information is not knowledge.” Those promoting artificial intelligence would lead one to believe it is all about replicating human judgment in a superior manner. The data part may be easy for machines but the human part isn’t. This is where computers meet their limits and our brains can triumph. Much of the investment world is captivated by using quantitative models to solve math problems when so much of investment success involves untangling behavioral problems through slow, hard, qualitative analysis.

In the world of entertainment, computers can reproduce a realistic image of a person but only actors can capture all of our complex paranoias, eccentricities and distinct imperfections. Computers simply cannot mimic these idiosyncrasies. The same could be said of investing. Successful long-term investors exploit inefficiencies caused by human error and frailty.

The bottom line is that algorithms have their limits. Facebook Inc. just announced it is hiring 3,000 “humans” to police its site for violent and inappropriate content. In keeping with that old saying about pornography that “you know it when you see it,” machines have a hard time judging disturbing material. For now, you can screen for words but not ethical content—there is no comprehensive algorithmic solution.

Another such algorithmic failure occurred during the crash of 1987 when computers couldn't handle the emotional irrationality exhibited by the market activity and ultimately exacerbated it. Meanwhile, some human investors saw the absurdity of the sell off and reacted intuitively and quickly without revising models and adjusting algorithms. Our strong results in 1988 were driven by the fact that we actually called our clients on Black Monday and asked for more money.

Sure, machines can be cheaper and maybe even easier to deal with than people. Computers are often right, but they also can fail spectacularly when things reach extremes or new patterns arise.

Source: Wall Street Journal

Jr Trader Petar Milanov

Read more:

• Fake news isn’t just a problem for the public, it’s a problem for algos
• Jim Cramer: “It was all about the rise of the machines.”
• RISE OF THE MACHINES or why automated trading could create problems down the line
• JP. Morgan: The next financial crisis will be Flash Crash
• Robo-adviser Scalable Capital hits $1.2 billion