qm

@quantmaxxer @nkreu113r never looked into it in much detail bcs it just doesn’t scale well enough for my applications (I know they have the Nano version of it but haven’t tried yet). generally I’m still trying to wrap my head around pretraining so don’t rly have an informed opinion

@quantymacro @nkreu113r how do you feel about tabular foundation models like tabpfn sir

Does anyone have advice for learning modern ML without tons of compute? Curious how people build practical skills. For context: am deeply familiar with linear regression, PCA, etc

ok I've also been trying to do more "modern ML" for the last few months and I was also very linear-regression brained so here was what helpful for me there's a channel on youtube called meanxai which I think is helpful (although it's AI voiced) bcs they actually go through the official papers for LGBM/XGBoost etc - it's also endorsed by our friend Max Dama so that's g also sometimes when I'm bored I just scroll through LightGBM github issues and you can learn quite a few practical tips here and there for NN/DL I actually struggled a lot relative to trees bcs for some reason 99.9999% of examples on the internet are about images/words/LLM which for some unknown reasons make it *much* harder for me to grasp (I'm so classic tabular data brained). so I had no choice but to try code some NN from scratch, look at the dimensions of each tensor at each step etc to build understanding generally I *always* try to find the equivalence between these NN models to OLS to help my understanding i.e what happens if we use linear activation in plain MLP etc or reading this paper: Ordinary Least Squares is a Special Case of Transformer if you're talking about *fitting* these models on huge data then yeah you don't really have other choices besides renting a GPU but fwiw I think you can build semi-decent understanding without it hope this helps ofc YMMV and I myself still haven't wrapped my head around all of these stuffs so yeah

@quantymacro Yes

@garybasin i got you bro just send me 100 million first since we’re friends I’ll send you 200 million stocks tomorrow

Who can sell me 200 million dollars of Anthropic stock? Serious offers only

London Marathon is done finally. hit my PB 2:46. the grind was so tiring now it’s time to celebrate with McNuggets see how easy it is to lie on the internet today I didn’t even go out of the house I’m just a fat fuck

@stevehou @techsaleshackz @paraformtalent np.random.randint hope this helps

@techsaleshackz @paraformtalent How's talent density calculated?

The 15 most talent-dense Tech Companies from @paraformtalent Some good names on this list

this tweet may or may not have anything to do with garry tan

was watching a video of how some ppl got affected by psychosis caused by AI chatbots. like these ppl became actually crazy. I’m sympathetic but at the same time I can’t help but to think this all could be solved if they take one linear algebra class

@bitshine_ be careful it turns the frogs gay. might not be a problem for you but never hurt to remind others

@quantymacro tap water king

it’s 2026 and only now I’m getting into NN/DL (I know) for personal stuffs and so far I’m using runpod but it’s actually kinda expensive is there a cheaper setup or is everyone just burning their paycheck on GPUs

@anpaure how does that even work who do I go to to do that is there a bootcamp for this

@quantymacro gpu credits are not real money therefore you can beg for them quite shamelessly with that in mind

@bitshine_ wait until you find out I won’t pay £5 for still water

@quantymacro >quant macro trader >cant afford gpu compute

@HangukQuant stop spreading misinformation. I’ve had enough of you. Jim hated vape, but it was Shabbat that day so no cigars allowed

@quantymacro It’s well know that Jim loved to be vape. He said “you must see the beauty in it”, and often bought pink coloured vapes during Sakura season

when I was at RenTech, we hired a lot of physicists, bcs a lot of groundbreaking concepts in finance came from physics. momentum. overextension. volatility is like a trampoline. Jim once said to me during post-lunch vape break: “study physics, and you will solve quant finance”

@moflod how do you think Renaissance generated 40% CAGR for years and years? we had access to SOTA infrastructure and models way before anyone else

@quantymacro is this a skit or something why mythos is being mentioned in a 2019 post

some have asked me about my time in Renaissance Technologies. although I’m retired I can’t really say much due to NDA, but I have an unseen interview snippet from my ex-colleague Nick (hope the kids are doing well mate) that I’m comfortable to share. a lot of alpha in there

@punished_daniel yeah and imagine facing this problem back in 2010. our team was struggling so much with this but luckily I figured out a solution and everyone clapped

@quantymacro I always have a hard time deciding between gstack and mythos or ChatGPT 5.4

@vegashasvega yeah Nick is a nice guy, but he was in back office. so forgive him if he’s not so aware when it comes to SOTA approaches

@quantymacro Now this THIS is insight! But this guy nick at renaissance (must be a shitco) CLEARLY doesn’t know that gbrain > gstack ^ (Overheard a JS intern on a Tribeca rooftop saying that btw)

to continue on this, today I’ll share a very useful AI tool that my HFT team has been using for the last 6m. recently we combined it with Mythos and immediately received an email from CME that our alpha is too high hence exposes exchange vulnerabilities. so DYOR.

@__paleologo as Jim Simons the GOAT said: to make money sometimes you have to short Vega short Gamma but never ever short Vulva

One of the most unexpected byproducts of being on X is that I receive >>0 requests for relationship advice from quants of all genders. It’s definitely an underserved niche. My advice to all: go ahead. Love a quant a little.