LW server reports: not allowed.
This probably means the post has been deleted or moved back to the author's drafts.
My hot take:
--I don’t think the tweeting would be super helpful, though maybe
--If it’s so simple to hook up small RL models to trading platforms and make money, why aren’t others doing it? Why don’t big quant firms use giant neural nets? (Or do they? I’ve heard they use old-school algorithms instead)
--The scaling laws seem to predict that if you want a giant model to be useful you need a correspondingly giant amount of data (otherwise, a smaller model would work almost as well.) Are you thinking GPT-N would generalize from its other knowledge to be good at trading? If not, wouldn’t it need a lot more trading data than you are giving it?
my friend made 50M
Is that 50 million dollars? Or is it a crypto abbreviation that means something else?
yes that’s 50 million dollars
How much did they start with / invest? How many years did it take? Were they part of a larger group (e.g. quant firm)?
I think if you have a good system to predict prices, it doesn’t make sense to add this tweeting dynamic. If you do it, and it produces good predictions about the combination tweet+price system and makes money as a result, then what will probably happen is a ton of other people clone the idea and produce similar tweeters and/or predictors until it’s much harder to produce good predictions with consumer-grade models, whereas if you kept your system to yourself you could have gone on extracting money from your good predictions quietly for longer.
Naively, the system seems similar to “good predictor” human fund managers who advertise their services, where you can see the end state I’m describing: there are so many of them that most of them don’t make money anymore, and even if someone shows you good evidence of their good predictions in the past it’s no longer sufficient evidence to act on their advice in the present.
sure I agree that keeping your system predictions for you makes more sense and tweeting doesn’t necessarily help. Maybe what I’m pointing at is where the text you’re tweeting is not necessarily “predictions” but maybe some “manipulation text” to maximize profit short term. Let’s say you tweet “buy dogecoin” like Elon Musk, so the price goes higher and you can sell all of your doge when you predicted the price would drop. I’m not really sure how such planning would work, and exactly what to feed to some NLP model to manipulate the market in such a way… but now it seems we could just make a simple RL agent (without GPT) that can do either:- 1. move money in his portfolio- 2. tweet “price of X will rise” or “price of Y will go down”. but yes you’re right that’s pretty close to just “fund managers’ predictions”, and that would impact less than say Elon Musk tweeting (where there’s common knowledge that his tweets change the stock/crypto prices quickly)