The frontier: A buying and selling algorithm that writes itself and evolves

For at least the past 25 years, human traders have been anxious about being replaced by machines and yet — today — there are more people trading than ever. I’d argue that many of them shouldn’t be trading and note that many of the human tasks in markets have been replaced but the future is hard to predict.

There is a cottage industry around writing and (mostly) selling trading algos. Their success has been mixed at best and it’s mostly been an avenue for grift in the retail space but Jim Simons made ungodly amounts of money quant trading and relatively simplistic methods (though we don’t know the details).

In any case, could this be the next frontier in trading technology?

Google DeepMind today pulled the curtain back on AlphaEvolve,
an artificial-intelligence agent that can invent brand-new computer
algorithms — then put them straight to work inside the company’s vast
computing empire.

AlphaEvolve pairs Google’s Gemini
large language models with an evolutionary approach that tests,
refines, and improves algorithms automatically. The system has already
been deployed across Google’s data centers, chip designs, and AI
training systems — boosting efficiency and solving mathematical problems
that have stumped researchers for decades.

That description comes from a Venturebeat report on the DeepMind announcement today.

The idea is that the model combines the creativity of a large language model with algorithms
that can scrutinize the model’s suggestions to filter and improve
solutions.

I can almost guarantee that there are people pouring money into this idea in markets. Unfortunately, the same technology is going to be put to military use and we’ll be lucky to survive it.