Beni

17.1K posts

Beni banner
Beni

Beni

@beniduboss

Ex-Quant Trader | Co-founder Stealth AI Startup | The only metric of success is happiness

Katılım Şubat 2021
1.2K Takip Edilen18.6K Takipçiler
Sabitlenmiş Tweet
Beni
Beni@beniduboss·
I’ve been getting a ton of requests for this, so I finally put it together Thread of all my posts regarding the marketing/sales/entrepreneurship content I’ve shared here I deleted some posts because I overshared lol Will try to update it every time I share something new 1/X
English
12
5
116
58.2K
Beni
Beni@beniduboss·
@Akira_DeFi It’s cheaper to fly to Turkey and do it there than it is in Switzerland lol
English
1
0
2
21
Akirage
Akirage@Akira_DeFi·
@beniduboss do you know a very good private dental clinic?
English
1
0
0
92
Beni
Beni@beniduboss·
It’s so unfair how the world sucks if you happen to be broke and sick I am getting an iron infusion in a private clinic on Wednesday Otherwise I had to get a doctors appointment in a week to get a prescription and then I would’ve had to wait 1-2 weeks for an appointment to get it done (1-2 weeks because Switzerland is great in other countries it could be months) Instead I just show my blood tests and pay the clinic 500 chf outright and I can skip all of this and get to feel better 2-3 weeks earlier Only now do I realize how frustrating it must’ve been for my mom to have to drag me around when I was a kid because we couldn’t afford to do such things
English
1
0
38
2.2K
Beni
Beni@beniduboss·
I can now say that I fully checked out from blockchain/crypto but there is still one thing I still can’t understand The last real mystery for me is this: who actually controls Hal’s private keys and what’s their goal? Why was that strange signature created years later on the first transaction between Satoshi and Hal, the one that pointed to Paul Le Roux as Satoshi? Most of the rest feels relatively clear to me, but I keep coming back to the same question: why would they have to gain? Hal’s dying wish to expose the truth or someone in charge of his estate wanted to make a stupid joke? Doubt we’ll ever know but it still lingers in my mind from time to time Always hated the Paul le Roux stuff so that’s probably why
English
1
0
35
4.8K
Beni
Beni@beniduboss·
@katexbt Qwen3.5-27B as controller and MedGemma 27B for well the medical reasoning and stuff There is probably a business there that could help patients who’ve met a dead end find new potential paths to take that the medical community wouldn’t think of
English
0
0
0
57
katexbt.hl
katexbt.hl@katexbt·
@beniduboss which self hosted llm did u use this seems like it could help a bunch of people though
English
1
0
1
118
Beni
Beni@beniduboss·
This might be the craziest post I’ve ever wrote and I have some mad ones on here If you care about AI for health/medicine read this: After seeing people use LLMs to figure out things about their health I figured I had to try I had an agent scrape research paper since my birth that mentioned either of my main disorders: Hemophilia A severe Pai-1 deficiency Ehlers-Danlos Reynaud Syndrome As you can imagine that’s a fair amount of data so it took a while Then I used my 87gb of personal medical files (lmfao) to cross-reference the papers in hopes of finding something doctors may have missed This post is about my findings and trust me you’re not gonna be disappointed: Found dozens of papers that don’t name me, but are about me like there is no shadow of a doubt, including one from John Hopkins and another from Cambridge One paper qualifies an event that I went through, something supposedly unbreakable that broke in an extraordinary fashion, as “an extraordinary stroke of misfortune” If you knew how funny it is to read this in hindsight Like I had open heart surgery because of said event and my parents were told multiple times to prepare for the worst, so seeing it written like that is very humorous But that’s nothing compared to what’s coming: Turns out Pai-1 deficiency is so rare that less than 10 families (including mine) have been publicly reported It’s believed to be one of the rarest bleeding disorders in the world (it’s underdiagnosed but imagine hemophilia is rare and it’s like 1/20,000 births, pai-1 is expected to be 1/500,000+) From all the sources I found + those already in my possession, I am, as far as we can tell, the only patient to have both said bleeding disorders In fact, if we only look at the odds mathematically, it’s very likely that I am the only person in the history of humanity to have had both 1 out of 100 Billion+… Mom wasn’t lying, I am special indeed But wait it gets so, so much more insane: Apparently the only known cluster for Pai-1 deficiency comes from families within the Old Order Amish of eastern and southern Indiana Yeah cool story I know but hang with me now Where did said Amish immigrated from in the 1800s? Switzerland From the part of the country my ancestors lived in to be precise… I found out that it’s entirely possible that I have the exact same gene mutation as said Amish because we could come from the same genetic pool Seriously what the fuck Still not done btw I’ve had chronic back pain for years, couldn’t be explained, imagery never came back with a definitive answer 2 research papers mention 2 different teen males with Pai-1 deficiency who complained for years about sharp lower back pain After extensive imagery they were both found to have recurrent psoas/iliopsoas bleeds But my doc’s should’ve seen those micro bleeds, how come they didn’t? I guarantee you that what the LLM found would’ve been missed by my entire medical team + myself, no human could’ve drawn these conclusions See when I go in the hospital these days I do multiple tests so that I don’t have to comeback 1st is a medicine check, administer a dose of clotting factor to test if it works as intended and cover risk for later injections during tests etc Those meds start working immediately, within 15 minutes I am at 233% of coagulation, meaning 2.3x stronger than you… It’s entirely possible that the bleeds get mostly treated by the time the imagery is done and hence it doesn’t show up or only very minimally Imagine I’ve gone to the hospital 100+ times to find out what could possibly be wrong with my back, including a 2 week hospitalization No human could’ve made a hypothesis that I could be experiencing micro bleeds that are common for Pai-1 deficient patients that cause the muscle to tense up and press against nerves but would not show up on imagery because my medication for another disease would’ve dealt with it by the time I got there Literally so insane I am lost for words
English
18
5
82
5K
Goon
Goon@GryptoGoon·
@beniduboss so keen for this new season of beni... dr beni arc gonna be wild
English
1
0
3
174
Godric
Godric@elGodric·
@beniduboss Damn this is so crazy, I am Sure AI in medicine will do some crazy stuff within the next few years. Did you upload all youre data into the llm tho and arent worried?
English
1
0
2
206
Beni
Beni@beniduboss·
Unfortunate but that’s the type of shit they’ll wait for as long as possible to pay That said makes zero sense that they don’t treat your anemia first given LC Would be the easiest thing to fix that would give you the highest improvement in quality of life Guessing you should be able to find an IV for iron in the US for 500$, especially since you can do this weird thing where you call the hospital to reduce the bill, because if I pay someone to come and do it in my house in Switzerland it costs that much without insurance Tag me in 2 weeks if things still didn’t move and we’ll figure it out, you shouldn’t be kept that way
English
1
0
6
146
Julie Bush
Julie Bush@julie_bush·
@beniduboss that’s so incredibly kind. i’m in a clinical trial to get cancer drugs for long covid and so far they’ve refused to pay for anything related to the trial (even though i’m saving them a million dollars in medical bills by getting into this trial). iron pills brought back symptoms
English
1
0
0
197
Julie Bush
Julie Bush@julie_bush·
my Anktiva doctor is trying to get my insurance to pay for IV iron cuz my iron & ferritin remain extremely low. a study came out last week that said low iron may be a biomarker for LC because the body sequesters iron so infections can’t feed off it. but then it gets stuck there
Beni@beniduboss

I want to sincerely apologize to all the girls in my life who complained about their anemia whilst I dismissed their complaints This shit is no joke Frantically calling every doctor I know to get an IV iron infusion next week cause it’s legit impossible to get anything done

English
6
0
38
7.5K
Beni
Beni@beniduboss·
It’s not a diagnosis It’s a hypothesis of potential reasons for outcomes which can then be used by doctors to determine a diagnosis I’ve been at ground 0 with my back pain since 2019, meaning we have up looking for the reason but we had tried everything Knowing what I know now I can tell you that I am pretty fucking happy to know that we could have some hope into finding the issue and then treating it
English
1
0
6
253
JaxBlaze
JaxBlaze@jaxblazey·
@beniduboss so ur saying ai will diagnose us wrong at scale now thats actually evolution we’re all cooked
English
2
0
2
265
Beni retweetledi
Rui Ma
Rui Ma@ruima·
Incredible, lol, robot maker Unitree just filed for IPO and not only do they make money, their adjusted net margin is 35%, putting it on par with software companies. Humanoid fever is only going to increase from here I think :) At @TechBuzzChina we have some new special projects in this area we are announcing soon! Some stats: - Unitree’s STAR Market (Shanghai) IPO has been accepted, with a planned raise of RMB 4.2 billion (US$611 million) and an implied initial post-money valuation of at least RMB 42 billion (US$6.1 billion) - 2025 revenue reached RMB 1.71 billion (US$248 million), - up 335% YoY, while adjusted net profit exceeded RMB 600 million (US$87 million), up 674% YoY. - In the first 9 months of 25, humanoid robot revenue reached RMB 595 million (US$86 million), surpassing quadruped robot revenue of RMB 488 million (US$71 million) for the first time. - Unitree shipped over 5,500 units last year, occupying 32.4% of the global humanoid market - Of the IPO proceeds, the biggest chunk, RMB 2.02 billion (US$294 million), will go toward robot model R&D, followed by RMB 1.11 billion (US$161 million) for robot body R&D. Another RMB 445 million (US$65 million) is earmarked for new product development and RMB 624 million (US$91 million) for a manufacturing base.
Rui Ma tweet media
English
28
106
1K
212.3K
lumpy the cosmic watchdog 📵
@beniduboss Investigate the hormone that if deficient stops iron being used even if it's there. Something related with vitamin d. (Apologies, my memory is shit!)
English
2
0
2
266
Beni
Beni@beniduboss·
I want to sincerely apologize to all the girls in my life who complained about their anemia whilst I dismissed their complaints This shit is no joke Frantically calling every doctor I know to get an IV iron infusion next week cause it’s legit impossible to get anything done
English
5
0
25
6.4K
Beni
Beni@beniduboss·
I don’t know OP so giving him the benefit of the doubt in case my 5am reading comprehension is letting me down once again, but this to me seems like a strawman Because : Well yeah I will hand it to him, using Kelly with fixed parameters is turbo stupid, even more so in crypto. It will lead you to an inevitable account liquidation sooner or later but that doesn’t mean that you can’t or shouldn’t be using Kelly at all (Yes yes nerds that’s assuming variance doesn’t forget you exist, I personally wouldn’t count on it but I am rooting for you…) The fixed Kelly assumptions aka known edge, stable distributions, constant payoffs etc don’t reflect how markets actually behave. Just in case you’re particularly slow, it doesn’t because, well, ain’t nothing is fixed in financial markets, that shit changes all the time. I just know that one of you would’ve asked in the comments so I am sparing you the embarrassment That said, it’s really only a criticism of how donkeys, that don’t know any better, misuse Kelly, rather than it is a criticism of the framework itself No one serious punts full, fixed Kelly in live markets And they don’t because many years ago when they were still young, naive and overly confident they tried it and now years later, they still suffer from the ptsd it triggered… I mean that’s what I heard, I wouldn’t know anything about being an 18yo, summiting the peak of the Dunning Kruger curve while being both remarkably overconfident and hilariously wrong simultaneously. So yeah, that’s about my good friends SBF and Caroline’s experience Anyways Here is what you should actually do: While trading, your edge is uncertain, so what you should actually be doing is modeling it as a distribution Once you have, congrats, Kelly rewards you with a principled way to size positions under uncertainty So yeah it’s still Kelly, it’s still useful, it’s just tweaked a lil Lastly, if you start layering in constraints like drawdown limits or volatility targets, you’re no longer doing actual Kelly stuff, instead you’re dealing with a constrained optimization problem which is basically what he’s getting at with the Monte Carlo stuff anyway Tomato tomato I am being a dick for no reason here but basically: TLDR: Kelly is not bad, misusing Kelly is bad. Most people I know, myself included back when I traded, use/used Kelly Hope that’s useful for one of you and if it’s not doesn’t matter I needed a break anyways cunts
mikkom@mikkom

Please don't use kelly for trading. It is not suitable. I see so many people posting about kelly this and kelly that related to trading. Kelly was originally designed for gambling. Gambling has fixed rules that typically favour the house but in some cases you can find an edge. Because the rules and the environment are fixed, the edge can be calculated. In trading you have volatility clustering and fat tails and everchanging environmental chaos from both external event, varying participants and market internals. Kelly assumes fixed and known constants. In trading NONE of these is true: - Environment is known and constant - Ratio of wins/losses is constant and known - Payout is constant and known You can get much better results via backtesting with decent sample size + monte carlo uncertainties. /rant

English
1
1
27
3.1K
Beni
Beni@beniduboss·
@itswithinme_ I can get you a better deal if you’re interested But that’s not even bad, I received one a week ago that was 30% upfront + 7.5% upstream Funnily enough market participants are using the Hyperliquid markets to justify these fees
English
0
0
10
815
B
B@itswithinme_·
Crazy fee stacking going on in Anthropic secondaries I’ve heard through the pipe At $380b valuation L2 SPV - 15/0 Broker - 7% Upstream L1 SPV - 5% So for $1m investment, you’re only getting $750,975 net capital into underlying exposure $249,025 / $1m = 24.9% fees So you to recover original exposure you need 33.2% gross appreciation to break even Wild
English
21
7
310
60.2K
Beni
Beni@beniduboss·
Wanted to buy a box of chartreuse for my dad Since when is it cheaper to buy human blood than a couple bottles of that green/yellow delicacy @pierre_crypt0 we should start a riot
English
1
0
10
1.3K
Loshmi
Loshmi@loshmi·
maturing is realizing that FTX was one of the best investors in crypto industry: > $500M in Anthropic would now be worth roughly $30.4B > $1B in Solana would now be worth roughly $5.1B > $648M in Robinhood would be $5.7B > $100M in Sui would be $1.2B > $1.15B in Genesis Digital Assets would be $3.5B > $700M in SpaceX via K5 would be $3B did some estimates and their total portfolio would be up from $4.7B to $52.5B which is a $47.8B in rise absolutely nuts
English
506
538
9.7K
1.1M
sasuke⚡420
sasuke⚡420@sasuke___420·
what if it's bpe but like, not bpe also...
English
2
0
3
556