Trading in the Age of AI

For at least the past 25 years, human traders have expressed concerns about being replaced by machines. Yet, surprisingly, the number of people engaging in trading has been increasing. It’s worth noting that while some human roles in trading have been substituted by technology, predicting what lies ahead is tricky.

There’s quite a niche market selling trading algorithms. The success of these algorithms has been somewhat mixed, particularly in the retail sector. However, Jim Simons has notably made a massive amount of money utilizing QUAND volume trading and relatively straightforward methods, though the specifics are a bit fuzzy to me.

So, is this where trading technology is heading?

Google DeepMind unveiled Alphaevolve today, an AI system designed to create new computer algorithms. It operates within Google’s expansive computing infrastructure.

Alphaevolve combines the capabilities of Google’s Gemini language model with an evolutionary method that tests and refines algorithms automatically. This system is already in use across Google’s data centers, chip designs, and AI training setups, enhancing efficiency while tackling complex mathematical problems that have stumped researchers for many years.

This summary is derived from an article by VentureBeat about today’s deep dive into this technology.

The concept is to merge the imaginative potential of large language models with algorithms capable of filtering and refining solutions through careful analysis of proposed models.

There’s likely significant investment interest in this idea. However, it’s concerning that the same technology could also be deployed for military applications. Let’s hope for the best in navigating this challenge.