Brent Crossman

On some level I feel like “Disney parks are very expensive and the lines for the big rides are still oppressively long” is part of the affordability debate and people’s dissatisfaction with the economy.

I was just reading today about how Italians (and other groups such as the Irish) used racism as a way to solidify their "whiteness". This feels like a version of that.

Kinda striking that ICE and the Border Patrol are focusing their enforcement on a blue state like Minnesota even though there are a lot more undocumented immigrants in red states like Texas and Florida. www.wsj.com/us-news/the-...

Hey, remember that Marvel storyline where Norman Osborn forms a new Avengers team, but it’s all supervillains in hero costumes? Just found myself thinking of that for no reason.

It was wrong to view the shifts to Trump in 2024 as an emerging new coalition (as I've written a lot) so it is wrong to suggest the coalition has since dissipated. The actual emergent "coalition" since 2016 is distance from the political establishment that leads to yoyo-ing electoral preferences.

If ICE being someplace was actually good, and made the place better, Southern Governors would be begging Trump to send ICE to their state. The fact they aren't highlights that ICE isn't meant to make a place better, it's meant to sow fear and compliance. Luckily Susan Collins is great at compliance.

Are you really struggling to figure out why a "wonderfully intelligent, curious, beautiful human being with an infinite capacity for learning and a fathomless depth of interiority" wouldn't want to interact, and learn from the greatest compilation of knowledge the human race has produced?

Choose a good model (can you claude opus 4.5 or any others you want). I run the RMD interactively in R-studio and as long as when you switch between the programs you "save" the file, it each program will load the changes automatically.

I tend to use RMD files so I don't worry too much about projects (probably my biggest divergence from standard R workflows). I copy one in that is a decent start, open the file in Rstudio and the folder in windsurf. I interact with the LLM using windsurf (ctrl+L opens cascade, the agentic coding).

I haven't tried claude code but my preferred setup is to use windsurf + rstudio for agentic coding. I create a folder/directory, do a quick git init in it and copy in a good .gitignore for r projects (an llm will create one for you). I then use github desktop, add that repository and push to github.

I probably should have just "for looped" the entire thing than rowwise/mutate/loop

Early attempts to shame people into not using better technology that the market was definitely going to incorporate into society because it was game changing. Luddites breaking mechanical looms.

Bluesky spending its time on the most productive stuff today. I'm not someone that feels like we can only focus on one thing or the biggest pain but if you think the market success of AI is near the top of the list of things we can or should stop, then I worry about the progressive movement.

github.com/bcrossman/ao...

https://github.com/bcrossman/aoc_2025/blob/master/day8.r

Part 2 churns for a bit but it didn't seem worth speeding up. And, as usual, I use igraph instead of trying to code graph algos.

I just completed "Playground" - Day 8 - Advent of Code 2025 #AdventOfCode adventofcode.com/2025/day/8 #rstats

github.com/bcrossman/ao...

https://github.com/bcrossman/aoc_2025/blob/master/day7.R

Was pretty happy with this, as I knew as soon as I went with distinct_all in part one, that part 2 was likely a counting of the passes instead (kind of a classic AoC solution when doing a large number of paths).

I just completed "Laboratories" - Day 7 - Advent of Code 2025 #AdventOfCode adventofcode.com/2025/day/7 #rstats

I feel like match .fun is my new eval(parse) for advent of code...and is something I will never use in real life

Ahhh, that second read with read_fwf... that's slick and I could see myself using that at work (read it in, figure some good splits out, run it again) so I'll have to remember that.

aoc_2025/day6.r at master · bcrossman/aoc_2025 Contribute to bcrossman/aoc_2025 development by creating an account on GitHub.

github.com/bcrossman/ao...

https://github.com/bcrossman/aoc_2025/blob/master/day6.r

This may be the longest pipe chain I've ever written but it actually really helped me to just be able to sequentially move through the data manipulation until I could get back to my part 1 method.

I just completed "Trash Compactor" - Day 6 - Advent of Code 2025 #AdventOfCode adventofcode.com/2025/day/6 #rstats

But part II was too big every way I tried it with the join

I kept trying to see if I could make it work somehow ("I'll just cross join the overlaps..."), stuff like that.