Steven Sanderson

Created healthyverse #R packages co-wrote Extending Excel with Python and R. I state my opinions, no one has to like them.

[Steven Sanderson @steveondata]

My tidyAML #R package has been updated:

Minor Fixes and Improvements

Fix #256 - Update load_deps() and install_deps() to use a try block.
Fix #266 - Major overhaul in how regression and classification model specs are built.

[Steven Sanderson @steveondata]

🚀 https://t.co/Dq08loiSKB 0.1.2 is now on CRAN!

• Standardized recipe object naming (new_rec_obj)
• Dropped modeltime & h2o dependencies
• Switched to native R pipe `|>` (R ≥ 4.1.0)
• Deprecated hai_kmeans_automl() functions

Details: https://t.co/EFO2iugJ7t

#rstats

[Steven Sanderson @steveondata]

site:https://t.co/BRjoblMfOL

[Steven Sanderson @steveondata]

If you are looking for a nice clean way of working with random walks in #R then our #RandomWalker package is a good place to start I think. It was built by myself and Antti Rask

https://t.co/cC9I0YCY2f

PR's are welcome on #GitHub!

[Power to the People ☭🕊 @ProudSocialist]

Elon just called Hitler a socialist again, and every time he does this, Grok corrects him and points out that Hitler was a fascist.

Never deleting this app.

[kai @0xkkai]

THE CREATOR OF OBSIDIAN JUST SHIPPED HIS OWN CLAUDE SKILLS PACK

44,000 developers already installed it. most people building second brains with claude have no idea it exists

these aren't community skills. they're written by kepano himself - the person who built obsidian. that means they respect obsidian's actual conventions: markdown, bases, JSON canvas, backlinks, plugins, all of it

drop them into your claude skills folder once. next time you say "summarize my notes about karpathy" or "extract this article to my inbox," the right skill wakes up automatically. no prompt engineering. no manual context. no config

i tested 10 claude + obsidian repos to write my full breakdown. this is the one with the highest trust factor - because the person shipping it is the one who defined what an obsidian vault even looks like

the full ranked breakdown of all 10 is in the article

i test things like this on myself and share the results. follow @0xkkai if you want to see what's next

[DulceBiatch @BiatchDulce]

Don’t stop talking about the Epstein files! #TrumpEpsteinCoverUp

[🇬🇧King 🇬🇧 @King0243_PJC]

Elon has the biggest megaphone on earth and he just told 241 million people that follow him that Hitler was a “left-wing socialist.”

That’s not edgy. That’s historically illiterate.

The Nazis hunted down, imprisoned, and murdered actual socialists and communists by the hundreds of thousands.

They banned trade unions, outlawed the Social Democrats and the Communists, and put their leaders in the first concentration camps.

Hitler’s own words called Marxism a Jewish disease he would eradicate.

The socialist in National Socialist was a marketing scam to steal working class votes while the party crushed the left and got into bed with industrialists, the military, and traditional conservatives.

You have more money and reach than almost anyone alive. You could use it to tell the truth. Instead you push a talking point that has been laughed out of serious history for decades. That’s not free speech. That’s using your platform to launder a lie.

[Steven Sanderson @steveondata]

If you are looking for a #tidy way of working with #distributions in #R then my #TidyDensity package is a good place to start I think.

https://t.co/ol8G9cBwa5

[Steven Sanderson @steveondata]

@Suryanshti777 ROADMAP

[Suryansh Tiwari @Suryanshti777]

Most people want to become AI Engineers

Very few know what to learn next

That's why I'm giving away one of the best Agentic AI Engineer Roadmaps for FREE to the first 4500 people only

Inside you'll learn

✅ Python Fundamentals
✅ LLM Fundamentals
✅ LangChain LangGraph CrewAI AutoGen
✅ LCEL Runnables & Workflows
✅ Memory Systems
✅ Tool Integrations
✅ RAG Systems
✅ Multi Agent Systems
✅ Real World AI Projects
✅ Interview Questions & Answers

72 Hours Only

How to get

Follow me (so I can DM you)
Like + RT
Comment "ROADMAP"

Once the limit is reached I'll stop sending it.

[Steven Sanderson @steveondata]

https://t.co/zlitdv7Ahx

[Steven Sanderson @steveondata]

#R #ARIMA #TimeSeries #Simulation

[Steven Sanderson @steveondata]

ARIMA simulations in R are a great for forecasting! Master arima.sim() to create synthetic time series & validate models. #RStats #ARIMA

Post: https://t.co/e4djBvqFdv

[OVVO Labs @OVVOLabs]

It’s all so obvious now since Leopold worked for SBF.

Most people treat the Kelly criterion as the aggressive benchmark. SBF treated it as the conservative one.

In 2020 he said that for a high-EV, high-variance bet, Kelly told him to risk ~10% of the bankroll. He preferred closer to 50%. Why? Because his utility of money was “closer to linear.”

Linear utility is the root of the behavior… and the ruin.

Under linear utility, a 50% chance of +$2 and a 50% chance of –$1 is pure upside. Expected value is positive, so you take the bet at full size, every time, with as much capital or leverage as you can get. There is no extra penalty for the loss side. A dollar lost hurts exactly as much as a dollar gained helps. The math therefore tells you to keep swinging until the edge disappears.
That works in an infinite-bankroll, infinite-time fantasy. In the real world it does not.

Wealth is finite. Paths are sequential. One run of bad outcomes ends the game. The long-run growth rate of a strategy that repeatedly exceeds Kelly is negative even when every individual bet has positive expected value. You go broke with probability approaching 1. The only question is how long it takes.

When the capital is other people’s money, the linear-utility agent has even less reason to flinch. The downside is externalized. The upside (status, impact, carry, narrative) is internalized. The rational move under linear utility + OPM is to lever up and keep pressing every positive-EV edge until the music stops.

That is exactly the posture that produced the FTX blow-up. It is also the posture that turns “we have a huge edge in AI” into leveraged concentration, margin calls, and forced liquidations the moment the path gets non-linear for a few months.

Any utility that is concave in losses builds in an automatic brake: large drawdowns hurt more than equivalent gains help, so you size down. Linear utility removes the brake. The car goes faster… until it hits the wall.

The error was never failing to understand expected value.

It was treating the utility of losses as linear when the world is not.

[Steven Sanderson @steveondata]

Here is a short blurb on #RandomWalker for #RProgramming useful with #Finance and #TimeSeries amongst other things:

https://t.co/cC9I0YCY2f

[Steven Sanderson @steveondata]

Simulate different Random Walks with our #RandomWalker #R Package

https://t.co/AJwuF1Pjtz

[Steven Sanderson @steveondata]

@AnthropicAI ‘s own #Fable rated our output better than its own

#LLM #AI #Finance

[Steven Sanderson @steveondata]

Trump & his allies proposed a rule to put political appointees in charge of federal grant funding - a move that could slow or stop the scientific & medical research we depend on to stay healthy. Tell your senators: Reject this rule ASAP. https://t.co/2JaZPZnLH4 @aftunion @AAUP

[Steven Sanderson @steveondata]

library(RandomWalker)
rw30() |> visualize_walks()

Docs: https://t.co/sYHzh4Wo7I
GitHub: https://t.co/AJwuF1Pjtz

#RStats #DataScience #Tidyverse #RandomWalks #Stochastic

[Steven Sanderson @steveondata]

Generate, visualize & analyze random walks with full tidyverse support Normal, Brownian Motion, Geometric BM, 27+ distributions, 1D/2D/3D, and more. Perfect for stock prices, particles, stochastic sims & beyond.

install.packages("RandomWalker")