Money is an information system, for labour allocation is how Elon frames it.
Forwarded from et/acc
They gave 5 LLMs $100K to trade stocks for 8 months - Grok +56% up & DeepSeek +49% up, both are crushing it.
- they were cautious to only run after each model’s training cutoff dates for the LLM models.
- That way they could be sure models couldn’t have memorized market outcomes.
Financial markets are ideal for evaluations like this because they let us analyze both quantitative and qualitative dimensions.
link: https://www.aitradearena.com/
- they were cautious to only run after each model’s training cutoff dates for the LLM models.
- That way they could be sure models couldn’t have memorized market outcomes.
Financial markets are ideal for evaluations like this because they let us analyze both quantitative and qualitative dimensions.
"This transparency helps us distinguish between memorization and genuine reasoning. There’s a difference between remembering to buy Nvidia and getting better at reading 10-Ks to extract insights on market fundamentals—and we can tell them apart by examining the decision-making process. This same transparency lets us evaluate how well different tools and configurations impact decision quality and outcomes"
link: https://www.aitradearena.com/