It’s all about the Software, Stupid.

November 1, 2024

We money managers depend on computers for tasks ranging from trading to correspondence to testing our theories of why, and when markets work, or not.?? The computer makes our lives vastly more efficient, and our new electronic markets that function 24 hours a day allow us to invest around the world, seamlessly.

But I have come to understand the key for evolution in our business is less in the hands of hardware and more on software, which we customize to fit the goals of our tasks.? Most of the top asset managers develop their own software and test their own theories, while many of the smaller companies in our industry (90%) are glorified sales offices, caring less about the history of markets than about bringing in new clients.

My own programmer, based in Chicago, has been with our team for some 35 years. As such, he has become a member of the family.? Steve came from Korea with his parents back in the 1970s, and is one of those rare programmers whose mind works so quickly that while he never took a single class in programming at that famous school partially responsible for the development of the internet--The University of Illinois-- his mind is nimble enough to self-learn, on the fly. His graduate degree was in Physics, which i consider the king of sciences, so his mind is no timid thing.

This summer I met his family here in Europe, and while speaking with his daughter--an accomplished programmer herself--in route from the airport, I asked her a simple question: “Do you ever confer with your father about programming issues?? Do you compare notes?” The expression on her face from my rear-view mirror, said it all.? “Ahh,” she groaned.? “I can’t keep up with the old man, even today,’ she moaned in a voice of desperation, her head dropping toward her lap.

Programmers of Steve’s caliber are rare, and with our growing dependence on hardware and faster chips and processors, the 'Steves' of the world are becoming rarer by the day.? Over the years I have learned that programmers are good at regurgitating tasks in a single language requested by their masters, but poor at designing and learning, expanding their understanding of new programming languages.? ?

Steve picks up a book, and in a day or two has started coding a new idea in a new language. His mind works like Velcro. In our decades of association, he’s never told me he couldn’t solve a problem, or code a task. ?More to the point, he’s never told me, “It can’t be done,” a statement that flies across the lips of most programmers.?

But this article is more about the deception by companies creating new hardware, that care little for the importance of software. This story may sound strangely familiar to what we are witnessing today in the Military Industrial Industry, where the goal is greater production capacity of second or third generation equipment, without regard for innovation, and how competitors might be turning our arms into jello on the battlefield through advances in digital technology and software. ?

The USA and NATO seem content to funnel more newly built, older generation equipment into Ukraine, with no regard for how Russia and China (and others) are jamming signals, and reducing and eliminating the advantages of these older arms flooding into battle. ??The Russians are taking apart Western militaries like they are cooked chickens.? The solution for the US and NATO is to produce more antiques that don't work on the battlefield.? Clearly, we cannot fight conventional wars any longer, and our air superiority is waning due to the development of new technologies by Russia and China. We're still good at bombing Afghani wedding parties from 25,000 feet, but our tanks on the ground are chopped liver.

But back to the matter of software and hardware.? I relate a story that evolved over many years.?

One of our ongoing tasks has been to test theories in the global markets.? We back-test our ideas of what we think may work, using extensive data streams that include both prices and fundamentals of about 40,000 companies around the world. ?One of our big areas of investigation has been Sharia Compliant strategies , suitable for the Western mindset as well as the Muslim world.

Over the years we have run many hundreds of tests that require us to loop daily data from thousands of different stocks, often going back 50 years or more.? Running a single stream of data to test our logic for a single company used to take us 6 hours.? Once the result appeared, we might tweak the logic and our variables, to improve the findings. Then another 6 hours of waiting for the computer to generate a new conclusion.? And on and on this went, for years. In many cases, our computers ran through the night, and we picked up results in the morning. ??

Today, by contrast, results that took 6 hours to run, require minutes to generate.?

Through the years it was our theory that we needed to speed the process by buying better and stronger computers, increasing the speed of the CPU processors and adding RAM to beef up memory.?

But one day, on a visit to Steve’s office, we discovered something quite remarkable.? ?Steve had maintained many of his older computers that were lined up along a table.? It occurred to me to ask Steve to test our newest code--much more sophisticated--on some of his oldest computers, configured back in the 1990s, using old versions of Windows.?? My assumption?? It would still take 6 hours to run a single test on those old computers.?

To our surprise, the same tests now took 2 or 3 minutes on the old computers.? We discovered the secret sauce to speeding up the testing was related to Steve’s enhanced software, his code, that streamlined the logic for looping the data over and over again to generate a result.? The old computers were more than adequate, but Steve’s code had improved so dramatically that hardware was less of an issue.

With some laughter we reminisced about the regular decisions we had considered in those earlier years that convinced us the missing link was an upgrade to our hardware.? We were wrong. What we hadn’t counted on was the sheer intelligence of Steve to upgrade his own knowledge over time, adapting his older code to smooth our lives, making the code much more efficient.

To a large extent, it seems, the computer hardware providers may be hoodwinking us through advertising and sales, urging us to upgrade our hard drives and CPUs, much like the Military Hardware industry is out hawking alleged newer generations of old arms, while glossing over the importance of the intelligence built into these weapons.? As a result, while Intel and NVIDIA are convincing us to buy more expensive chips, the writers of software are taking a back seat.

It’s all about the sale, the increase in revenues by touting new products, while ignoring the importance of intelligence, of sheer brain power, to simplify and make our lives work better and more efficiently.?

In the case of our military the US should be the bastion of this brain power, but because everything—at least in warfare--has gone corporate, the governments that need to buy these arms are unable to keep up with countries who depend more on brainpower and human intelligence. And importantly, these governments do not privatize their militaries. I refer to Russia and China, and also India--the countries of the new BRICs.? ??

We shall see how AI--the new kid on the block--evolves. My guess is the same trend will emerge, favoring faster chips and slower brains.? ?I hope I'm wrong.