Voluntas is a trading firm written 100% in Rust. We run systematic strategies, currently specializing in prediction markets, looking to expand our systems more broadly.
We're looking for someone passionate about data collection, clean abstractions, and processing trillions of rows of data. We care about the quality of the code we write, and we prioritize time for our engineers to properly build their systems, instead of pressuring them to churn out AI generated code.
If you're interested, send an email to [email protected] with a little bit about yourself. We'll respond with a link to our engineering challenge (in Rust).
I downloaded Binance's self published market data (https://data.binance.vision/), calculated spot and futures volume for the past year, and then applied their *highest fee tier* on both sides of the volume (for every $1 traded, they generate a fee on the taker and the maker. the fee rate is asymmetrical).
Spot revenue worked out to $7.6B USD, perps revenue worked out to $13.8B
This is an upper bound on the fees charged on volume. Binance also scams their users with a b-book, has DMM deals, etc.
I don't want to get into all that today.
But the final result is that they print somewhere in the range of $20B per year on fees. I guess CZ also has huge stockpiles of BTC and stuff. A 1% rise in BTC price probably means more for them than their actual fees, LOL.
All this to say, I think $5.6B significantly hurts.
See Google's stock price today. Their ads business slowed down a bit, and their stock tanked 10%. If binance was publicly traded, I imagine a similar affect would be demonstrated on their price.
My outlook on life is that it is an experiment where you see how you can develop yourself throughout it. This growth is difficult and beautiful.
I am proud of myself, and the person I have become. I surround myself with people on the same mission.
This article describes a future like WALL-E. It sounds very lonely. I'll be sure to be in the pocket of people who are like minded, and set on their own pursuits.
I am actively choosing an existence where I don't get everything I want. It's the effort that makes it mean something to me, and it's the restrictions that make life worth living.
It might be useful to know that I programmed for >50% of my life, starting when I was 12. It was a huge part of my identity for most of my life. I was known to be extremely good at it. Think: newspaper articles, CEOs flying me out, etc.
Overcoming this entanglement of identity, ego, and competence was one of my proudest achievements in the past 5 years.
I am mostly at peace now, always trying to go further. I hope people like the author can find peace as well.
There's more to life. Perhaps we will value eachother on an even *more* human level, based on how kind and interesting someone is, rather than their economically useful skills and their bank account.
I'm keeping this short but I have more to say to corroborate these points.
I care the most about quality. It can be more expensive, and slower. My problems with the AIs are always intelligence related/effort related.
So I'm not sure if those smaller models you listed will be effective, given that I am having to turn to chatgpt.com's Pro to get quality answers.
Interested in Kimi & DS
Something is up.
A bit token expensive, though. Especially when it isn't part of the subscription.
I often find chatgpt.com has greater "intelligence" than codex.I've found this to be true for the past year or two now, and I was wondering if anyone has found a better solution than copying context to and from chatgpt.com. The "Pro" level (or whatever the max level is at the time) consistently has better insights than the agent, but immensely limited tool calling compared to the agent....