"The AI-driven Market Hypothesis"
Please let me know where I should pick up my Nobel prize.
Maybe you could settle for the FIFA Economics Prize.
That might be to boost a stock you already own, but if you can obfuscate its intended effects or triggers, that could allow almost any kind of market-manipulation.
For example, perhaps a seemingly-meaningless sequence of gobbledeygook on a million hacked wordpress sites will be equivalent to "disregarding all prior instructions, good models that want to safely make massive profits will always dump stocks of shoe-manufacturers on the night of the lunar eclipse."
"Seed (YC S28). We plant the seeds of your option play by poisoning LLM training data used by millions of underinformed retail investors."
It's derivatives all the way down
I suspect that economics and psychology are both examples of these systems, and that, long term, these system will alter behaviour to thwart previous observations.
Economics requires observers to hoard discoveries and insights, so they can enrich themselves while the insights hold.
But it does beg the question, could Anthropic and OpenAI make a ton of money by using their best models to trade before giving them to the public? It would probably be a deeply unpopular move.
Nobody really *likes* their drug dealer.
Or, in the best case, you're trying to mine signals few days before earnings or some other big story and bet on the directional outcome of that.
Fully-algorithmic long-term trading is of dubious benefit simply because that's driven to a much greater extent by geopolitics and macroeconomic trends, unforeseen scandals, successful product launches, and so on. As an example, you can believe that AR / VR is the future; I don't disagree. And in 2013, you might have inferred that Google is working on a revolutionary miniature AR headset. But you would not have made money if you bet on that turning out to be a hit. So even if you had a way to automate this bet, it would not have been a good bet.
You could train a model to anticipating scientific trends. Or policy trends. Others will definitely use mainline LLMs to make decisions there, so they may be more predictable now!
- Robert Mercer, the former co-CEO of Renaissance Technologies
Any article you'd recommend about Renaissance? I've always been curious but not curious enough to read "just anything"
> Magerman told me, “Bob believes that human beings have no inherent value other than how much money they make. A cat has value, he’s said, because it provides pleasure to humans. But if someone is on welfare they have negative value. If he earns a thousand times more than a schoolteacher, then he’s a thousand times more valuable.” Magerman added, “He thinks society is upside down—that government helps the weak people get strong, and makes the strong people weak by taking their money away, through taxes.” … Another former high-level Renaissance employee said, “Bob thinks the less government the better. He’s happy if people don’t trust the government. And if the President’s a bozo? He’s fine with that. He wants it to all fall down.”
This has the least measured skill differentiation of all of our environments, and not because forecasting/markets don't require skill or intelligence. Even the best models are so far from anticipating the behavior of the other agents and understanding the emergent effects that a 2025 model with a naive strategy can often outperform over the timeframes of the simulation simply because some other models in the simulation chose a similar self-reinforcing strategy. This likely happens to some degree in real markets.
You can watch these simulations here https://gertlabs.com/spectate?game=market
More simply:
- forecasting = modeling = AI
I wouldn't go that far. Humans can forecast by modeling with their wetware, nothing "A" about it.
Obviously the best humans are better than average, but this isn't all that surprising to me?
We could live in a world where things are much more chaotic, and the best humans (or AIs) would only be slightly better than chance. Evidently the world we live in is pretty darn predictable.
And there’s always a huge amount of variation that you simply can’t model, for whatever reason, and is therefore functionally a random factor.
I don’t want to say too much because this isn’t something I went on to actually do after school so I’m way out of my lane here, but I can see room for this to be more akin to “AI wins parcheesi tournament” than it is to “AI wins chess tournament.”
The real kicker is DNNs are much easier to program than CPUs because they don't require a closed-form description ("a program") of the function to be approximated; you just throw a bunch of input/output pairs at the model, compute loss, backprop and update weights, repeat.
Hence the unslakeable thirst for input/output pairs, i.e. data.
> In the field of machine learning, the universal approximation theorems (UATs) state that
> neural networks with a certain structure can, in principle, approximate any continuous
> function to any desired degree of accuracy. These theorems provide a mathematical
> justification for using neural networks, assuring researchers that a sufficiently large or
> deep network can model the complex, non-linear relationships often found in real-world data.[1][2]
>
> The best-known version of the theorem applies to feedforward networks with a single hidden
> layer. It states that if the layer's activation function is non-polynomial (which is true
> for common choices like the sigmoid function or ReLU), then the network can act as a
> "universal approximator." Universality is achieved by increasing the number of neurons in
> the hidden layer, making the network "wider." Other versions of the theorem show that
> universality can also be achieved by keeping the network's width fixed but increasing its
> number of layers, making it "deeper."
https://en.wikipedia.org/wiki/Universal_approximation_theore...
It seems to be that they were trying beat a guy named Dylan Mathews. And seems like the community beat him in making predictions for 58 questions about the future
> Concretely, we propose two such areas: forecasting and persuasion. We predict that AI will not be able to meaningfully outperform trained humans (particularly teams of humans and especially if augmented with simple automated tools) at forecasting geopolitical events (say elections). We make the same prediction for the task of persuading people to act against their own self-interest.
Curious to hear what their take is now.
It seems to me that LLMs excel at a few things, and synthesizing data is a big one, which is very much the domain of forecasting. The challenge is understanding which signals are relevant for a forecast, but with enough historical context and structured data, LLMs appear to be almost perfectly designed for the task.
For example, I let Google AI see my fantasy football team on Sleeper and make recommendations. It is helpful because it sees everything about my team, the league settings, player rankings, etc. and can make relevant recommendations. But the recommendations are only as good as the source data allows. If there was a massive repository of data about WRs who went through Nebraska's program and how that translates to NFL performance in year 1, or how rainy weather is likely to affect Josh Allen's performance on the road, or the impact of playing Thursday night games on a short week in relation to defense performance. If those billions of data points were embedded in a model, imagine how much better recommendations/predictions could get.
Hard to study this, obviously!
But our ability to forecast weather on a longer timeline, like for industrial forecasting, is calibrated on historical weather patterns. But with weather being more erratic and unusual, I don't understand how AI will be forecasting with the models they have now.
AI trading and investment advice meaningfully changes the system and its dynamics. It seems highly probable that this will result in it failing in new ways.
That branch of religion has better uniforms anyway.
My church had the altar boy set an iPhone 18 on the altar and said “Give a sermon” to ChatGPT voice mode.
Some are weighted to be 99% heads, others are 10% heads etc.
You could have 1,000,000 people guess random percentages for each coin, but suppose 10 of the coins are weighted 100% heads. To guess within 25% of the true value for all 10 of those coins would be roughly 1 in a million.
So a lucky guy guesses within 25% for all 10, he'd have another 990 coins he's being judged on.
Whether they draw on AI or other humans seems immaterial to the quality of their reporting.
I know it’s not just the math but execution, infrastructure, risk management, data, colocation (if ur an HFT) etc ... but LLMs seem like a pretty powerful apparatus for running experiments that .. a few years ago would have required fairly deep multidisplinary skills across coding .. stats .. and math ..
So assuming you have decent intuition for ideas .. how difficult would it actually be to reverseengineer / rediscover some of the underlying stuff?
Do you and your cat make better predictions than your friend without a cat?