Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Financial market applications of LLMs

thegradient.pub

Post reply on HN
Print view View original

31–40 of 116 posts

Re: Financial market applications of LLMs

#31

Post by AnthonyMouse »

carbotaniuman wrote: post #15
ben_w wrote: post #13

Earlier quoted context omitted.

"Buy index funds. The end." From what I've heard (and as finance isn't my field, my knowledge should be considered worse than ChatGPT ), if everyone had a truly omniscient genie, the markets would become perfectly efficient, and a perfectly efficient market has no room for profit because any profit opportunity is immediately arbitraged out of existence.

I get that the perfectly efficient market is more of a model then something existing in reality, but who would be doing the arbitraging here?

Suppose the price of Amazon stock is going to be 20% higher tomorrow than it is today. If everyone knew this, the price would already be 20% higher, because the existing owners wouldn't sell at the lower price. If some people know this but not everyone, they'll keep buying Amazon stock until the price increases by 20%, which again causes the price to immediately increase by 20% instead of waiting until tomorrow.

The arbitrage opportunity is available to anyone who knows the information, at the expense of anyone trading the stock who doesn't. If everybody knows then there is no arbitrage opportunity because the gap is already closed.

Re: Financial market applications of LLMs

#32

Post by nostrademons »

btbuildem wrote: post #30

There were some developments using LLMs in the timeseries domain which caught my attention. I toyed with the Chronos forecasting toolkit [1], and the results were predictably off by wild margins [2] What really caught my eye though was the "feel" of the predicted timeseries -- this is the first time I've seen synthetic timeseries that look like the real thing. Stock charts have a certain quality to them, once you've…

I used to work in financial software, and when writing the charting UIs, I'd wire them up to a randomwalk to generate fake time series data. It was a relatively common occurrence for a VP or the company CEO to walk by, look at my screen, and say "What stock is that? Looks interesting."

Unpopular opinion backed up by experience: a randomwalk is the most effective model for generating timeseries that have the "feel" of real stock charts.

Re: Financial market applications of LLMs

#35

Post by kortilla »

ysofunny wrote: post #8

If I learned anything from a conference by benoit mandelbrot back in my college days is that gaming financial markets is the only real application of anything scientific but I vaguely remember what he was actually talking about, I never quite made it as a mathematician

What does that even mean? How is the atomic bomb not real?

Re: Financial market applications of LLMs

#36

Post by cpursley »

conorh wrote: post #26

We are working on a project for a client which functions as an analysis tool for stocks using LLMs. Ingesting 10ks, presentations, news, etc. and doing comparative analysis and other reports. It works great, but one of the things we have learned (and it makes sense) is that traceability of the information for financial professionals is very important - where did the facts and information come from in what the AI is p…

I assume you're ingesting PDFs. If so, how are you handling tables accurately?

Re: Financial market applications of LLMs

#37

Post by logicallee »

Jerrrry wrote: post #9
unixhero wrote: post #7

The art here for a human would be to find the sweet spot of how LITTLE data to feed the llm and to get the weights and other goodies just right for it to be realistic to run for a single non-billionaire.

All you'd get are projections with percentage error margins; you can choose the riskier plays, but it is literally priced in. You'd also get clapped by the HFT bots. The real magic is pairing real human intuition and the LLM's innate ability to discover hidden intuitions and articulate them to find an "asymmetry"-where you believe you have found a gradient/play that is under/over valued and play the opposing side - o…

Building on the point about using LLMs for finding market asymmetries, I'm looking to team up with a trader to create a UI that leverages AI to spot these opportunities. The idea is to use custom prompts to generate actionable insights, tailored to real trading scenarios.

I'm a developer with experience in clean, effective UIs like this QR and barcode generator[1] and have worked with neural nets in competitive settings - recent robotics contest livestream[2]. I need a trading partner's insight to ensure we focus on the right features and data.

If you're a trader interested in shaping and using this tool, I'm proposing a partnership where you'd provide the trading expertise and potentially fund the initial development for a stake in the project. Think of it as investing in custom software that you'll own and can directly benefit from.

Anyone interested, please check my profile for my contact. Just looking for one trader-partner who really wants to dive into this.

[1] https://qr-code-and-barcode-generator.taonexus.com/

[2] https://www.youtube.com/live/IDF7zN0NGgA

Re: Financial market applications of LLMs

#38

Post by hydershykh »

I think some of the financial applications around LLMs right now are better suited for things like summarization, aggregation, etc.

We at Tradytics recently built two tools on top of LLMs and they've been super popular with our usercase.

Earnings transcript summary: Users want a simple and easy to understand summary of what happened in an earnings call and report. LLMs are a nice fit for that - https://tradytics.com/earnings

News aggregation & summarization: Given how many articles get written everyday in financial markets, there is need for a better ingestion pipelines. Users want to understand what's going on but don't want to spend several hours reading through news - https://tradytics.com/news

Re: Financial market applications of LLMs

#39

Post by wuj »

HFTs exploit price inefficiencies that last only milliseconds. The time-series data mentioned in the article is on the scale of seconds. I wonder if its possible to get the time-series data on the scale of milliseconds, and how that would affect the training of the objective function in a LLM.

Re: Financial market applications of LLMs

#40

Post by paulryanrogers »

hydershykh wrote: post #38

I think some of the financial applications around LLMs right now are better suited for things like summarization, aggregation, etc. We at Tradytics recently built two tools on top of LLMs and they've been super popular with our usercase. Earnings transcript summary: Users want a simple and easy to understand summary of what happened in an earnings call and report. LLMs are a nice fit for that - https://tradytics.com/…

As more of the reports get written by layers of AI it makes me wonder how lossy and noisy this whole pipeline is becoming.
Post reply on HN