Solved a problem the old school way last night (following a Stack Overflow that linked to another Stack Overflow than linked to a GitHub repo with three stars), and I’m going to be riding that high all week long
Emily Riederer
@emilyriederer.bsky.social
Here for data, data science, analytics engineering, rstats, books
I should be clear that the architecture of Bluesky makes it the social media place I most like to be. I'm not on social media to be an influencer or instigator; I want to hang out with folks and have an enjoyable time. This place delivers on that. I hope it continues to do so for a long time.
One of my pet peeves: People are obsessed with seeing/plotting error bars while also being confused about what they should represent. blog.genesmindsmachines.com/p/the-error-...
The error bar cargo cult
You don't need to understand error bars. You just need to draw them.
blog.genesmindsmachines.com
Throwaway thought, but it’s wildly unhelpful to use AI as a catch all for everything from transformer-based sequence models to agentic coding assistants to LLM generated docs. I suspect sentiment varies a ton across those and lumping them together doesn’t make for a productive conversation
oho! you haven’t proven all of the negatives! i am very smart bsky.app/profile/ohno...
I’m sorry what
I'm more and more convinced that 99% of the biggest public figures would absolutely collapse if an interviewer just said "ok but that's not true" after every nonsense lie.
There's a reason Elon Musk rarely does interviews with real news outlets.
I adore examples of good crosspolination across languages. In-line dependencies is an awesome uv feature. Excited to see it inhabit the R and Quarto worlds as well!
I'm happy to share `ir`, a new command-line tool for running portable R scripts and Quarto documents. It's inspired by my two favorite parts of `uv`: self-describing scripts and tools you can run without installing first. opensource.posit.co/blog/2026-07...
Well now I feel like a fool that I’ve never asked my team to “do an innovation”
the Dinitz-Garg-Goemans conjecture, a graph theory problem that’s gone unsolved for 30 years, was solved with this prompt
Two weeks ago today, after Nasdaq willingly rigged the rules, somewhere btw $7-10 bil of passive money was forced to buy SpaceX at $160.42 Today, 10 trading days later, it closed at $119.85 It’s an unprecedented forced loss (at least $2b) for passive investors and our financial media is 🦗
Argentinian levels of panicky instability, tendencies toward violence, and desperate appeals to authorities to bail them out all increasing as pressure comes on, much as often happens in their economy.
“You should be deathly afraid of a free-to-download matrix of real numbers because it will cause communism by increasing market competition” is quite the strong of takes Almost certainly authored by man, as no transformer would wander into such a probabilistically unlikely sequence of words
Dean Ball: Open source software is communist and threatens to bring on a dystopian hellscape. Also, it's COVID.
"I used ChatGPT as a search engine" is a natural extension of "I use Excel as my database".
Hi, this is Holden emerging from my Bluesky hiatus to tell everyone that when checking the links in the piece, I used ChatGPT as my search engine (does anyone use Google anymore?). The piece was written by me and edited by humans. So all the links were checked by me and again by our copyeditors.
while i am not an academic i did see this coming and post about it on bluesky, which is why i am quoted in this article
Academics and technologists are sounding the alarm about a growing crisis in scholarship as we know it: AI-generated citations of nonexistent papers that have infested real journals. Despite being fake, the sources are widely assumed to be authentic the more they appear in published literature.
It’s true! NC’s is even free. If you vote in the state of North Carolina, you’re hanging out on my hard drive for my pure intellectual pleasure
Voter files are public information that anyone can purchase!
Voter files are public information that anyone can purchase!
Dang, 220 million voter files stolen by China starting in 2020? Who was president back then, it sounds like he must have sucked ass
Tomorrow is the deadline for submitting a comment to stop the administration from dismantling U.S. science. There are nearly 280,000 public comments filed so far! But! Do not assume they are all, or even most, against the proposal. This is a key pillar of Project 2029 - wind down federal R&D.
Here’s a great resource to help write a comment this weekend before the deadline, by @colesci.bsky.social and @cdelawalla.bsky.social: www.standupforscience.foundation/policy-and-a...
Where do we sign up to hire the 20%
woof ≈80% of assignments this summer term have essentially been just raw chatgpt/claude output, including assignments that required designing stuff in illustrator/affinity/canva. The battle is lost for asynchronous online courses.
Fascinating to see Bluesky take the 2010s approach to product development wherein you actually made your product continually better not worse
v1.127 is live! We're rolling out improvements to search over the coming weeks. You'll get more relevant results and more ways to find what you're looking for. The new "Filters" button lets you query and filter by keywords, people, date range, language and more. You can even share your searches!
Podcast feed in a drought over the holiday weekend? I made a list of my fav follows. I don't always listen to all, but its a great way to keep pulse on emerging topic get high-level context outside your specific bubble Appreciate your recs back! www.emilyriederer.com/post/pods/
Skills for humans: 24 data-adjacent podcast recommendations | Emily Riederer
Skills help our agents know that certain bodies of knowledge exist and enable them to go deeper when needed. I use podcasts for the same effect. In our fragmented data-sphere, it’s harder that ever to...
emilyriederer.com
This is a great podcast rec! Got any other stats/data science ones you also like?
Patriotism asks you to fight for values and an ideal. Nationalism asks you to fight for a person or group. There's a difference, and it matters.
There are technically infinity Trumps in this photo. But it’s a countable infinity so relatively small as infinities go.
Trump is watching the Fox News feed of himself watching the Fox News feed.
Great post on how agents translate pandas -> polars syntactically correctly but myopic line-by-line translation misses core design philosophy Polars team offers SKILLS to fix, but more broadly a testament to the value of understanding a tools ethos not just its syntax pola.rs/posts/llm-po...
Fluent, not native: agents translating pandas to Polars
We had AI translate a pandas corpus to Polars. Most one-shot translations ran and the outputs matched. The code was idiomatic in most cases. Two structural patterns remain, and our new skill improves ...
pola.rs
And how am I supposed to brainstorm when I left my best brainstorming pen at home? I’m supposed to think on a screen, or in my head, or when my handwriting looks utterly mid?
Reposting because this show is such a balm to the wound of Mondays
I deeply believe causal thinking is core to good DS regardless if you do analytics, ML, etc A new-to-me resource is the excellent In the Interim podcast on clinical trial design Comes out weekly and takes the sting out of my Monday commute Check it out! www.berryconsultants.com/resources/po...
In Opinion The system that turned Gila monster venom into GLP-1 drugs is being dismantled through research budget cuts, Jeff Coller writes in a guest essay. “Less support for scientists means strange questions no one will get to chase. Exploring those questions is how medicine advances.”
Opinion | We Need to Keep Funding Weird Science
Science turned a serendipitous finding about lizard venom into one of the most important drugs of the century, but that type of research is getting harder to do.
nyti.ms
Don’t sleep on local and/or open models! My current PC doesn’t have RAM for full local, but I’m impressed by what OpenCode with DeepSeek v4 Flash will do for free on structured tasks One shotted some meaty R/python translation today, and I’m sure can be much faster with more context and tools
new post: how I develop recently using local models. the tooling is now good enough to do agentic workflows and everyone should give them a try! vickiboykis.com/2026/06/15/r...
new post: how I develop recently using local models. the tooling is now good enough to do agentic workflows and everyone should give them a try! vickiboykis.com/2026/06/15/r...
Running local models is good now
Local agentic coding has gotten great over the past few months
vickiboykis.com
Remembered a quote from a book I read about ~15 years ago and want attribution, can remember when I read it, the cover, the hotel I was staying at when I read it, the social event I was at whereupon I wished I could go back to my room to read... but not the book Thanks, brain