"Provifying", surely?
Posts by Grant McDermott
On the TODO list ;-)
```
"spinnerVerbs": {
"mode": "append",
"verbs": [
"Estimanding", "Procedurelarzing"
]
}
```
Paul and Jonah's config (probably).
You know, I genuinely thought of you while nuking this setting. Such twee!
```
"personality": {
"sass": "80%"
}
```
"Sir, this is a Wendy's" etc.
*Embarrassingly niche content*
I recently spent some time tweaking my Neovim config, alongside some related shell tools. I think the end result is pretty slick & makes for an extremely fast+ergonomic workflow across my most used tools: #rstats, #pydata, git, Claude etc.
github.com/grantmcdermo...
Tbf there are surely times when I deserve it.
My own CC config is here fwiw: github.com/grantmcdermo...
A small, but meaningful quality of life improvements to your Claude Code workflow is available by editing your .claude/settings.json:
```
"spinnerVerbs": {
"mode": "replace",
"verbs": [
"Working"
]
}
```
Sure, you won't see "perambulating" any more. But you're also not 12 yo.
Good news, everyone. I was just chatting to Claude and it turns out I am, in fact, absolutely right.
Excited to launch the accompanying free RLHF Course for my book. To kick it off, I've released:
- Welcome video
- Lecture 1: Overview of RLHF & Post-training
- Lecture 2: IFT, Reward Models, Rejection Sampling
- Lecture 3: RL Math
- Lecture 4: RL Implementation
Landing page: rlhfbook.com/course
A study of around 44,000 papers finds that the credibility revolution has spread unevenly beyond applied micro, driven mainly by difference-in-differences, with finance and macro lagging by roughly 15 years, from Paul Goldsmith-Pinkham www.nber.org/papers/w35051
TBF both of these are work-induced reqs (although I will say that the performance on these new M-chips is remarkable). Also, love r2u but it still doesn't give version pinning, right?
Tricksy upstream P3M "bug" that might be affecting performance in some of your `renv` and `rv` enabled projects on MacOS. #rstats
forum.posit.co/t/missing-op...
Incredibly strong album cover energy.
(It's going to break the fans' hearts when your man in the front departs for his inevitable solo career.)
And another Quarto announcement; I've alluded to it before, but we're making it "official".
We've started work on Quarto 2. The blog post has an overview: quarto.org/docs/blog/po...
We'll share more in future blog posts, but here's what you can expect from the Quarto 2 dev effort:
(1/)
Exactly. It's so good at writing / translating to Typst; I don't get why people (academics) keep asking it to make extremely dense Beamer decks, etc.
(I mean, I _do_ get it b/c writing LaTex sucks and it's great to offload that to an AI. But that just locks you in to its other inefficiencies).
Coming up (down?) for air after a social media break.
As you do, here are some updates for various open-source projects that I've been working on...
Did a series of videos with Markus Brunnermeier on Claude Code (more to come)
Video 1: Getting Started with CC
open.substack.com/pub/paulgp/p...
5. Last but not least, we've been chipping away at `jgd`, a novel + lightweight #rstats graphics device. My original motivation was simply better graphics support in VS Code, but @eitsupi.bsky.social has pushed it much further.
CRAN submission imminent. Try it out. github.com/grantmcdermo...
4. Speaking of AI, I started putting some skills & agents files up on GH, containing my own programming preferences. Use / don't use. github.com/grantmcdermo...
3. I finally updated my CV & decided to create a Typst template to my liking. Borrows from @kjhealy.co's lovely LaTeX template, but man is it easier & less code to customize in Typst. github.com/grantmcdermo...
Aside: It would be a real bummer if If AI causes LaTeX lock-in over Typst.
2. Speaking of feature-rich (& fast), @lrberge.bsky.social, @kylefbutts.bsky.social & I updated our `fixest` paper to coincide with v0.14.0 arxiv.org/abs/2601.21749
A nice little Easter egg is that multithreading now works out-the-gate on Mac. No more Makevars incantations.
bsky.app/profile/lrbe...
tinyplot boxplot with jittered points
1. `tinyplot` v0.6.1 is out on CRAN. This release contains mostly bug fixes and internal optimizations. grantmcdermott.com/tinyplot/NEW...
Who knew that base #rstats plotting could be this easy and feature-rich?
P.S. Check out @zeileis.org's nice slides here: www.zeileis.org/papers/Psych...
Coming up (down?) for air after a social media break.
As you do, here are some updates for various open-source projects that I've been working on...
C.f. The Aubrey and Maturin books as decisive counterpoint to this theory.
"Ah, the f'castle mizzen spizzen you say. Yes. Yes, of course." I nod sagely while my vision swims.
My nerves are shattered and I wasn't even watching. Out with the kids and furtively (frantically) refreshing the live score feed.
Poor old Borthwick really is going to get strung out to dry, isn't he?
Oh, totally (forking paths). You could see SCO had lost focus.
But must be infuriating for FRA supporters that they only dialed in for the final 10 minutes. Although, I expect that bonus point try will prove decisive for the tournament.
*somewhat.
But, really, the important thing is that PSDT would have never. π