Sabitlenmiş Tweet
Nati KohnShvartz (vi/vim)🎗
2K posts
Nati KohnShvartz (vi/vim)🎗
@NathanKohn
VP AI Data Science at @Lemonade_Inc (NYSE:LMND) | ex VP, AI @Citibank | ex Intel AI products | @TechnionLive Graduate #AI #DataScience #GenAI #Insurtech
Israel Katılım Kasım 2015
2.1K Takip Edilen1.5K Takipçiler
Nati KohnShvartz (vi/vim)🎗 retweetledi
Nati KohnShvartz (vi/vim)🎗 retweetledi
Home sellers now exceed buyers by over 600,000, marking the widest gap on record, per Redfin.
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
@MBjegovic Setting the stage?
x.com/ericldaugh/sta…
Eric Daugherty@EricLDaugh
🚨 JUST IN: President Trump says America has WON THE WAR against Iran "Oh, I think we've won. We've knocked out their navy, their air force. We've knocked out their anti-aircraft. We've knocked out everything. We're roaming free!" "From a military standpoint, all they're doing is clogging up the strait. But from a military standpoint, they're finished!" LFG! Victory is coming! 🇺🇸
English
An absolute carnage in the bond mkt...
Will it stop the war?
English
@MBjegovic Totally agree! especially coming from you.
Been following you for a while, firmly in your camp
English
Ppl need to chill. There WON'T BE any rate HIKES in the foreseeable future.
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
@RetailKingsETF $LMND forcing the market to rethink what insurance looks like when its built on real-time data with a cost structure that improves as it scales.
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
Without mentioning the company @Lemonade_Inc once, @McKinsey puts out a strong buy recommendation for $LMND.
Link in next post
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
Bruna Valeanu was brutally murdered by Hamas on October 7.
Remember her.
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
Mullahs security checkpoints are
now staring at the sky way more
than they’re watching the streets 😂
Feel the fear, you cowards.
The Iranian people will be
free of you soon. !!
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
I’ve met people like this throughout my life. They have an uncanny ability to sound informed and thoughtful, despite being ignorant and shallow. The skills of a charlatan. They start by providing investment or medical advice, then move into politics. We’re surrounded by them.
The Bulwark@BulwarkOnline
White House AI & Crypto Czar David Sacks: "Israel is getting hit harder than they've ever been hit before in their history. And we're only two weeks into this. If this war continues for weeks or months, then Israel could just be destroyed… And then you have to worry about Israel escalating the war by contemplating using a nuclear weapon, which would truly be catastrophic"
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
@levyguy77 בוא נעשה בדיקה קטנה,
אני שם כאן תמונה של אסף רפפורט, שבזכותו קופת המדינה מקבלת 10 מיליארד ש"ח.
ונראה מי יקבל יותר לייקים.
אבל תבטיח שלי שאתה לא נעלב…
עברית
Nati KohnShvartz (vi/vim)🎗 retweetledi
"I wouldn't say business is as usual, but the Israeli tech economy keeps on ticking."
Here is half of my CNBC interview with Sara Eisen on Israeli resilience, Iran risk discount for Tel Aviv Stock Exchange and much more.
English
Lost 100 followers overnight.
If you see this post please say Hi! 👋
English
@gilileoB למה יושבים על הרצפה? למה לאכול מהרצפה? לא מצליח להבין את זה
עברית
Nati KohnShvartz (vi/vim)🎗 retweetledi
I took the @karpathy autoresearch loop and pointed it at markets.
25 AI agents debate macro, rates, commodities, sectors, and single stocks daily. Every recommendation scored against real outcomes. Worst agent by rolling Sharpe gets its prompt rewritten by the system. Keep or revert. Same loop, prompts are the weights, Sharpe is the loss function.
Trained the agents on 18 months of market data. 378 iterations. 54 prompt modifications, 16 survived.
The system learned which agents to trust using Darwinian weights — geopolitical, commodities, and the @BillAckman quality compounder rose to the top.
The agents even figured out their own portfolio manager was the weakest link before we did!
Deployed the trained agents. +22% in 173 days. Best pick: AVGO at $152, held for +128%.
The final prompts are evolutionary products — shaped by market feedback, not human intuition. Now running live with my own capital.
github.com/chrisworsey55/…
Part hedge fund, part research experiment :)
Andrej Karpathy@karpathy
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
karpathy just broke the internet with something called auto research
it’s basically an ai research agent that runs experiments for you 24/7
you give it a goal like
“make this model better”
“find a higher converting landing page”
“lower customer acquisition cost”
then it runs a loop:
1) plan an experiment
2) edit the code or config
3) run a short test on a gpu
4) read the metrics
5) keep the winner
6) try again
over and over
while you sleep
by the morning you wake up to the best version
actual tested improvements
think of it like a robot research intern that runs hundreds of experiments and only keeps the winners
this is link to his repo github.com/karpathy/autor… for your to mess around with it
in the latest episode of @startupideaspod
i break down:
• what auto research actually is
• how it works step by step
• 10 business ideas you can build with it
• how to install it and start using it
this one is saucy
because tools like this change how startups get built
watch
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
On October 7, 2023, Palestinian Muslim terrorists from Gaza invaded Israel and murdered Lin Dafni and Ziv Frenkel while shouting the name of Islam’s god “Allah Akbar”
Never forget!
Never forgive!
English
Nati KohnShvartz (vi/vim)🎗 retweetledi
Stop what you’re doing and spend 80 seconds understanding this
Not because you’re training LLMs.
Because Karpathy just showed the cleanest example of the agent loop that’s about to eat everything:
1.Human writes a strategy doc
2.Agent executes experiments autonomously
3.Clear metric decides what stays and what gets tossed
4.Repeat 100x overnight
The person who figures out how to apply this pattern to business problems…not just ML research, is going to build something massive.
The code is almost irrelevant. The architecture and mindset is everything
Andrej Karpathy@karpathy
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)
English