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

Robinhood now lets your AI agents trade stocks

techcrunch.com

Post reply on HN
Print view View original

91–100 of 192 posts

Re: Robinhood now lets your AI agents trade stocks

#91

Post by asdff »

tombert wrote: post #28
infecto wrote: post #4

I don’t understand this constant fascination with having language models trade stocks. Language models are very useful tools but not aligned at all with generating alpha.

I use the Interactive Brokers MCP pretty heavily. I don't do any cool automatic fun "trading", but instead I use it to have "pseudo-QQQ". I didn't like the relatively high fees for QQQ, and I realized that Invesco releases the weights for QQQ for free. I also think Tesla is too overvalued, and I want to avoid the SpaceX IPO. With the Interactive Brokers MCP, I just feed it the CSV of QQQ's weights, tell it to remove…

You don't need AI for this though. I was doing something like this with a python script and a crypto meta etf I created years ago. I even had some simple heuristics for selecting what coins and quantity to purchase given trading volume and spot price. Its like 175 lines of python. Probably could be a lot leaner too.

Re: Robinhood now lets your AI agents trade stocks

#92

Post by ai_fry_ur_brain »

crazygringo wrote: post #47
infecto wrote: post #4

I don’t understand this constant fascination with having language models trade stocks. Language models are very useful tools but not aligned at all with generating alpha.

Alpha is ultimately the result of analysis, of better analysis than others. LLM's can actually be exceptionally good at research and pattern recognition, i.e. analysis. And while they aren't great at running numbers themselves, they can do exceptional work passing off Python scripts to an interpreter to generate the numerical results they need. I'm quite sure the Robinhood AI is going to be trash, i.e. just a gimmick…

>LLM's can actually be exceptionally good at research and pattern recognition, i.e. analysis.

No they aren't, they're good at imitating analysis based on representations of analysis in their training data. Also, Its likely that out dated techniques would over represented in training data.

Do you think Jane Street would have the returns they do if they just imitated all their competitors and everyone was using the same strategies?

Re: Robinhood now lets your AI agents trade stocks

#93

Post by tombert »

asdff wrote: post #91
tombert wrote: post #28

Earlier quoted context omitted.

I use the Interactive Brokers MCP pretty heavily. I don't do any cool automatic fun "trading", but instead I use it to have "pseudo-QQQ". I didn't like the relatively high fees for QQQ, and I realized that Invesco releases the weights for QQQ for free. I also think Tesla is too overvalued, and I want to avoid the SpaceX IPO. With the Interactive Brokers MCP, I just feed it the CSV of QQQ's weights, tell it to remove…

You don't need AI for this though. I was doing something like this with a python script and a crypto meta etf I created years ago. I even had some simple heuristics for selecting what coins and quantity to purchase given trading volume and spot price. Its like 175 lines of python. Probably could be a lot leaner too.

I agree I don't need it, I actually wrote a program to automatically buy and sell stuff years ago using Alpaca [1].

I just found it a bit of a pain in the ass to manage a service to do that automatically, vs thirty seconds of chatting and getting results immediately, and having something that can be supplemented by RAGs in the process.

[1] I swear I had a blog post about how I did it somewhere but I seem to have misplaced it.

Re: Robinhood now lets your AI agents trade stocks

#95

Post by radial_symmetry »

unglaublich wrote: post #7
infecto wrote: post #4

I don’t understand this constant fascination with having language models trade stocks. Language models are very useful tools but not aligned at all with generating alpha.

The usual question: what's "aligned with generating alpha" that a human stock trader can do, but an ai can't?

The humans can't either

Re: Robinhood now lets your AI agents trade stocks

#97

Post by bdangubic »

iAMkenough wrote: post #65
Jagerbizzle wrote: post #48

Earlier quoted context omitted.

I believe you’re confusing access with outcomes. Giving people access to markets isn’t exploitation afaic. If you’d like to make dubious trades that’s your prerogative and who am I to stop you.

I have access to a car and bottle of bourbon, but there are laws that restrict me from drinking and driving.

they don't restrict you, you can always drink and drive. you may or may not, depending on your luck, suffer the legal consequences of your actions.

Re: Robinhood now lets your AI agents trade stocks

#98

Post by asdff »

tombert wrote: post #93
asdff wrote: post #91

Earlier quoted context omitted.

You don't need AI for this though. I was doing something like this with a python script and a crypto meta etf I created years ago. I even had some simple heuristics for selecting what coins and quantity to purchase given trading volume and spot price. Its like 175 lines of python. Probably could be a lot leaner too.

I agree I don't need it, I actually wrote a program to automatically buy and sell stuff years ago using Alpaca [1]. I just found it a bit of a pain in the ass to manage a service to do that automatically, vs thirty seconds of chatting and getting results immediately, and having something that can be supplemented by RAGs in the process. [1] I swear I had a blog post about how I did it somewhere but I seem to have misp…

It sounds like you are just pulling weights of qqq and buying based on that though. What more management do you have to do? Just pull and parse the weights wherever they might be stored, break the investment up based on that weight. Should work until the heat death of the universe.

Re: Robinhood now lets your AI agents trade stocks

#99

Post by Maxatar »

tombert wrote: post #28
infecto wrote: post #4

I don’t understand this constant fascination with having language models trade stocks. Language models are very useful tools but not aligned at all with generating alpha.

I use the Interactive Brokers MCP pretty heavily. I don't do any cool automatic fun "trading", but instead I use it to have "pseudo-QQQ". I didn't like the relatively high fees for QQQ, and I realized that Invesco releases the weights for QQQ for free. I also think Tesla is too overvalued, and I want to avoid the SpaceX IPO. With the Interactive Brokers MCP, I just feed it the CSV of QQQ's weights, tell it to remove…

This is absolutely and unfathomably terrible to such a great degree that I think it reinforces OPs point. It seems like using an LLM has given you the confidence to make an incredibly ill-informed decision that will cost you dearly.

Every single time you rebalance your portfolio, you will need to pay short-term capital gains taxes on any gains, as opposed to an ETF in which you simply pay for the gains when you sell your stock which can be years/decades from now. This alone will reduce your average expected earnings by 20% over a 10 year period eviscerating whatever tiny advantage you think you'll get from saving a few bucks in fees.

Furthermore, assuming you rebalance your portfolio monthly, which is the minimum you need to rebalance in order to remain even somewhat aligned with QQQ, you're basically going to be paying a MINIMUM of 30-40 bucks a month in commissions to Interactive Brokers, or 400 dollars a year. And on top of IBKR's commissions you then need to pay the pass through fees of about 5-10 dollars a month for a total of around 500 bucks a year.

Compare that to QQQ which only costs you 18 dollars a year for every $10000 invested.

I've read some incredibly foolish investment advise on HackerNews, but I think this one just about takes the cake.

Re: Robinhood now lets your AI agents trade stocks

#100

Post by dumbmrblah »

Maxatar wrote: post #99
tombert wrote: post #28

Earlier quoted context omitted.

I use the Interactive Brokers MCP pretty heavily. I don't do any cool automatic fun "trading", but instead I use it to have "pseudo-QQQ". I didn't like the relatively high fees for QQQ, and I realized that Invesco releases the weights for QQQ for free. I also think Tesla is too overvalued, and I want to avoid the SpaceX IPO. With the Interactive Brokers MCP, I just feed it the CSV of QQQ's weights, tell it to remove…

This is absolutely and unfathomably terrible to such a great degree that I think it reinforces OPs point. It seems like using an LLM has given you the confidence to make an incredibly ill-informed decision that will cost you dearly. Every single time you rebalance your portfolio, you will need to pay short-term capital gains taxes on any gains, as opposed to an ETF in which you simply pay for the gains when you sell…

But he avoids SpaceX and Tesla, which I think is probably the driving factor in not using QQQ. Maybe he values that more than $500
Post reply on HN