paso

Dogfood Maximalist. Test in prod.

[keshav @keshavchan]

the hardest part of transformation isn't the work. it's the willingness to let go of the person you used to be. we hold onto old identities because they're familiar, but real change happens when we're ready to rewrite who we are from the inside out. it's about understanding that who you were got you here, but they can't take you where you're going. thank them. honor them. then let them fade into the rearview. you're not betraying your past self—you're fulfilling their wildest dreams. every version of you was just preparation for who you're becoming. and trust me, future you is fucking incredible.

[retrodev⌨ @NewAgeRetroNerd]

If you're a guy in your early 20s, buy this book

Go into debt if you have to

[paso @0xpaso]

@mikevanrossum lekker

[paso @0xpaso]

@0xIchigo learning to snowboard is great. prepare to fall for the first day or two. get padded shorts.

[Keone Hon @keoneHD]

In my experience the larger forms will focus on large opportunities and low risk / low operational overhead ones.

There is opportunity in the longer tail of assets, exchanges, and strategies. Also edge in doing things early. Large firms have a backlog as well as politics to deal with (e.g. tough to assign someone to a speculative/tiny exchange even if it’s +EV because they might be upset)

[Keone Hon @keoneHD]

There is an open source "real-world AI hedge fund" github repo going around twitter that's getting a lot of interest.

Since in my previous life I was using ML for trading, I wanted to share some notes on the repo and how one would actually create a "real-world AI hedge fund".

First, about the repo:

The core logic of the repo is about 750 lines of python. The 6 "agents" are each about 100 lines; each one is computing some very explicitly-defined metric that is something like "count the number of insider trades, if there are more buys than sells, return 1, else return -1".

The main use of "AI" in this repo is to spin up gpt-4o from OpenAI and present it with the following prompt (I'm abbreviating to give the gist, but see line 569 of agents . py in the repo):

"""
You are a portfolio manager. Your job is to make a trading decision based on the team's analysis while strictly adhering to risk management constraints.

Quant Analysis signal: {some number from say 1 to 5 which was calculated by one of the aforementioned 'agents'}

Fundamental Analysis signal: {'bullish' or 'bearish'}

Please output whether to 'buy', 'sell', or 'hold'.
"""

Then there is a backtester which converts the 'buy', 'sell', or 'hold' actions into actual trades using daily (end-of-day price) data.

What's wrong with this

Quant finance is the practice of generating measurements ("signals") that are relevant to predicting future price action; combining those signals somehow into a more complicated model that actually predicts the future price action; and trading on those predictions.

A quant at a quant trading firm will typically spend a good portion of their time thinking up new trading signals that might have predictive power:

- For example, a quant who is trying to predict the price of AAPL might notice that AAPL and MSFT are pretty correlated, and write a signal to measure how much the price of MSFT has changed recently relative to AAPL to measure some notion of price change that hasn't been priced in.

- Another quant might hypothesize that whenver there are way more shares on the bid side than on the offer side, the price is more likely to go up, and write a signal to encode that particular metric.

There could be thousands+ of these signals, especially when you consider that some of these signals have parameters which should be searched over.

After devising all of these signals, the quant needs to combine them somehow into a model that predicts the future. This means setting up a prediction problem, which means (at least) 4 things:

1. Choosing a set of signals to pull;

2. Choosing a set of sampling points (for example, every hour, every second, every time there is a trade, etc);

3. Choosing a ground truth (or "target") to try to predict (typically the delta between AAPL's price at the sampling point and at some future point in time, e.g. 1 second into the future); and

4. Choosing a model for combining the signals.

One of the reasons that AI/ML is very applicable to finance is because advanced models can be good at capturing nonlinear relationships between signals, provided that we can train the models with a lot of data (part 2 above) and have a good ground truth (part 3).

Prediction problems that effectively use an AI/ML model usually involve large datasets with millions of rows and hundreds or thousands of columns.

If it helps, imagine putting yourself in the shoes of the computer. At each row in the dataset, you are presented with all of the values of all of the signals, along with information about what ended up happening to the price. Your job as the computer is to figure out how to utilize those signals to predict the outcome with great accuracy. Thankfully, we have amazing algorithms like feed-forward neural nets which do a good job of memorizing relationships between the signals whose data is being presented and making predictions that do a good job of matching ground truth.

The power of Machine Learning is in learning what to do from a large, well-curated dataset with a well-chosen target.

Now back to the repo

The repo constructs a couple of signals with very low granularity (basically just "bullish" or "bearish")

The repo doesn't actually train an AI model to figure out how useful (or useless) the signals are at making a prediction of future price action

The repo instead asks ChatGPT "what would you do if your quant told you that MACD was bearish but RSI was bullish, would you 'buy', 'sell', or 'hold'?"

There is no learning here. There is no utilization of the actually amazing powers of AI to combine signals into a prediction. There is no feedback loop to even tell the model that it did a good or bad job.

(The backtest also has a lot of issues but I won't get into that.)

If you are interested in the subject, I would recommend playing with "Stock Market Analysis + Prediction Using LSTM" on Kaggle.

Although roughly equally simple in terms of the number of signals and granularity of sampling, that Kaggle notebook actually shows the process of constructing a dataset and target, and training an actual ML model. If you try adding more signals, or training over more interesting sampling points, or thinking about what target you should actually be predicting, you will end up gradually getting a better model and learning a lot about quant trading in the process :)

[Orange Book 🍊📖 @orangebook]

Boredom means the game is too easy. Anxiety means it's too hard. Depression means it's pointless. Anger means it's rigged. It's impossible to feel bored, anxious, depressed, angry, if you are playing the right games with the right mindset. Either change the game, or level up.

[Robert Greene @RobertGreene]

Do not be afraid to bring out the more sensitive or ambitious sides to your character. These repressed parts of you are yearning to be let out. In the theater of life, expand the roles that you play.

Don’t worry about people’s reactions to any changes in you they sense. You are not so easy to categorize, which will fascinate them and give you the power to play with their perceptions of you, altering them at will.

[Robert Greene @RobertGreene]

You must avoid at all cost the idea that you can manage learning several skills at a time.

You need to develop your powers of concentration, and understand that trying to multitask will be the death of the process.

[Kris Sidial🇺🇸 @Ksidiii]

@OkayEstimator It’s to get people to trade with the trading desk.

This is just one example, but lots of larger institutions will put together “research” like this and launch hedging products. They underperform and then the space gets a bad reputation. It’s a nasty cycle.

[Temu Robot James @ScottPh77711570]

Sensible crypto strategies for the stonk guys

High level: Crypto is riskier than stocks and the markets are less efficient

The implications of this are:

Any "risk premia" harvesting will harvest more risk on crypto

Any price inefficiencies are likely to be less competitive

1/

[paso @0xpaso]

@mattyryze bro

[paso @0xpaso]

@tferriss the real question is how on earth did you get a copy?

[Shane Parrish @shaneparrish]

Stop fighting your nature. Start winning with it.

[Chubby♨️ @kimmonismus]

This is the future, robots folding towels 24/7

[Chris Burniske @cburniske]

1/ If your friends associate you with Bitcoin, Ethereum, or any part of crypto, you’re probably getting Qs about what to do. It’s tricky to guide someone from where we are currently ($BTC ~$100K), especially if they’re an inexperienced investor. Some learnings from 10+ years of watching this dynamic play out...

[Space Monkey @RealSpaceMonkey]

Can't believe I'm writing this, but orderbooks are starting to make less sense on Solana... Let's dig in:

Essentially, the point of an orderbook is to aggregate different orders into one "book". That way, the user/taker gets the best price out of the auction.

What else is like that on Solana/DeFi? Aggregators i.e. Jupiter @JupiterExchange

What's happening right now is that Jupiter is the defacto "orderbook" on Solana and AMMs are placing different "orders" on Jupiter then it selects one/combination of them. This is a direct result of onchain composability (DeFi) and cannot happen in TradFi/CEX world. It can't happen on L2s either or on PERP markets unless you have one global risk engine (extremely hard, won't happen).

Ok, but why does that mean orderbooks make less sense? Can't we still have them? Because it's extremely expensive to run an orderbook now on Solana with the astronomical priority fees on top of Jito fees (which you have to pay too..) I need to write more on that. At the same time, you don't want to be an AMM because those, in case you didn't know, bleed money continuously. The main benefit of an orderbook DEX is that the orders move based on the fair value.

We have an answer already: Oracle-based AMMs. It's nothing new, we had Lfinity for a long time. You might think: but we still have to update the oracle which is the same as updating orders. Well, not really, because you can be extremely efficient with that e.g. update all oracles in a single TX which is the equivalent of placing a 100 orders at once on an orderbook which is currently impossible.

We are seeing this right now: Obric, Solfi and more coming up.

To sum up: orderbooks are losing because of Jupiter/aggregators taking their job as an efficient liquidity aggregator + Solana fees exploding.

This might seem like obvious common sense but just remember this is unique to DeFi and if you're coming from a TradFi/CEX background, there is no way this was immediately obvious.

[Physics In History @PhysInHistory]

"You can't convince a believer of anything; for their belief is not based on evidence, it's based on a deep seated need to believe."

-- Carl Sagan (1934 - 1996)

[Flo Crivello @Altimor]

This is insane

[paso @0xpaso]

@brianlong @nic_carter i don't know any that aren't