Old Greed in a New Silicon Suit
Wall Street has a long and storied history of dressing up old-fashioned greed in a fresh suit and calling it a revolution. In the nineties, it was the dot-com 'new paradigm.' In the mid-aughts, it was the 'risk-free' alchemy of mortgage-backed derivatives. These days, the suit is woven from silicon and high-density neural networks. Billions of dollars are currently being poured into AI stock-picking models, but from where I’m sitting, it looks like the same old quant-fund circus, just with faster horses.
The fundamental flaw of this modern obsession is simple: a machine is a master of patterns but a absolute dunce at principles. An AI can scan forty years of ticker tapes in the blink of an eye, but it possesses no soul, no skepticism, and certainly no common sense. It doesn't know the difference between a business that makes structural steel and a business that makes vapor. It is built for prediction, not for judgment. And in the world of investing, prediction is a parlor trick; judgment is what keeps you solvent.
The Ghost of 2021: Overfitting to the Mirage
I remember the 2021–2022 mania with a particular clarity. Back then, the 'sophisticated' machine learning models were the darlings of the day. They were fed a diet of a decade’s worth of data characterized by zero-interest rates and the relentless 'number go up' momentum of the tech sector. Because the models were overfitted to this specific, anomalous period, they concluded that risk had been engineered out of the system.
When the tide finally went out, these models didn't find a margin of safety. They didn't have a backup plan. They just found the cold, hard bottom of the ocean. Many of these systematic funds were so busy optimizing for 'alpha' in the crypto-sphere and high-growth tech that they forgot to check if the underlying assets actually existed in any meaningful sense. The math told them the yields were stable; any student of history with a decent pair of eyes could see it was just a Ponzi scheme with a better user interface.
"A machine sees a 20% return and calculates a buy signal; a seasoned investor looks at that same number and starts looking for the fire escape."
The Black Box vs. The Spreadsheet
I’ll stick to my simple spreadsheets, thank you very much. A neural net is a black box. It gives you an output, but it cannot explain the 'why' in a way that respects the reality of a balance sheet. A spreadsheet, however, is a transparent mirror of a company's health. It forces you to look at the cash flow, the debt obligations, and the actual dirt under a company’s fingernails. You cannot automate the discipline of refusing to overpay for an asset.
Value investing is, at its core, an exercise in human temperament. It is the ability to stand still while everyone else is running. AI models are programmed to participate; they are designed to find the trend and ride it until the data changes. But by the time the data changes in a crash, the liquidity has already evaporated. The permanent edge that a human investor possesses is the 'gut feeling', the deep-seated intuition that a trend has outrun its shoes.
- Intrinsic Value: AI calculates price targets based on momentum; humans calculate value based on assets and earnings.
- Historical Context: Machines see data points; humans see cycles of human behavior.
- Margin of Safety: Algorithms maximize for the best-case scenario; value investors prepare for the worst.
The Next Digital Stampede
The next market crash won’t be caused by a lack of data, but by too much of it being interpreted by machines that have never seen a real winter. It will look exactly like the quant meltdowns of the past, but only this time, the algorithms will hit the 'sell' button at the speed of light. When every model is trained on the same data sets and looking for the same patterns, they will all try to squeeze through the same narrow exit at the exact same moment.
When that digital stampede arrives, I won’t be panicking. I’ll be waiting on the other side of the door, clutching my spreadsheet and buying up the shares of real, productive businesses that the algorithms discarded in their haste. Wall Street can keep its neural nets. I’ll keep my margin of safety and my peace of mind. After all, the market is a device for transferring money from the impatient to the patient, no matter how many flops per second your computer can process.
"Investing is like fishing—cast your line wisely; otherwise, you’ll just be hauling in a net full of speculation."