An AI analyst made 30 years of stock picks – and outperformed human investors by a ‘stunning’ degree
(news.stanford.edu)
from Pro@programming.dev to technology@lemmy.world on 10 Jun 23:20
https://programming.dev/post/31982678
from Pro@programming.dev to technology@lemmy.world on 10 Jun 23:20
https://programming.dev/post/31982678
Using public information and making small tweaks, an alpha-seeking AI fund manager outperformed 93% of mutual fund managers by an average of 600%.
threaded - newest
What kind of AI? An LLM? Because I can see how hindsight would be hugely beneficial for stock trade. And if they used a model with knowledge from 2024 on data on 1990...
A language model is for language. Zero chance they used an LLM.
It “made” the money based on a backtest. It means nothing.
Depends on whether the AI is cheating by using data more recent than the trade date.
Will take longer to run it in real time and get meaningful results (I presume).
If an AI can emulate Renaissance Technologies’ Medallion Fund and do so on a large scale in the future, that would be a great step forward, but only a few would be able to do so.
I could care less if it beats someone on data that already happened. Let me know how it does going forward. My guess is that it won’t beat an s&p500 index fund.
it’s not exactly terribly difficult to beat the s&p 500
Yet the vast majority fail after you account for expenses.
kind of
Can we stop letting chatbots make economic and administrative decisions, thanks
What makes you think this was LLM? From reading the article, it sounds like something else.
for what it’s worth “AI” has been making trades around the clock for quite a long time now it’s just getting better. but it’s not the AI that you’re thinking of in the modern context like an llm these are very specific logical trading bots.
Didn’t they crash the market that one time
.