Sign in

Michael Mullarkey

@mcmullarkey.bsky.social
4.6K followers 1.3K following 994 posts

Clinical psych PhD | ML Engineer in sports Subscribe to my newsletter on data, learning, and building: buttondown.com/datadash Opinions are my own and don't reflect the opinions of my employer

PostsRepliesMedia
Michael Mullarkey @mcmullarkey.bsky.social · 06/10/2026
It's at the very least relatable
000
Michael Mullarkey @mcmullarkey.bsky.social · 01/10/2026
How to cultivate a mysterious aura 101
010
Michael Mullarkey @mcmullarkey.bsky.social · 25/09/2026
I can see why this feature happened (programming languages can have nested columns!) Counterpoint: nesting columns often creates more problems than it solves. Tidy data caught on because one value per cell generally beats "this cell could have any number of values good luck"
110
Michael Mullarkey @mcmullarkey.bsky.social · 25/09/2026
What if "mistaking genes for a date" but with *checks notes* every kind of data
192
Reposted by Michael Mullarkey
Lisa DeBruine @debruine.bsky.social · 25/09/2026
❇️ I've updated Data Skills for Reproducible Research to v6.1. The Iteration & Functions chapter has been split in two and I've added some material on unit testing. It's a living textbook, so please submit issues if you find a mistake or something unclear. psyteachr.github.io/reprores-v6/...
12010
Reposted by Michael Mullarkey
Tiger Tang @tigertang.bsky.social · 24/09/2026
📘 It’s official: my first book, Work Automation with R, is published! I started with a simple belief: you don’t need to be an R expert to build useful automation with R. I certainly wasn’t one when I started. 📖 www.routledge.com/Work-Automat... 1/5 #RStats #RProgramming
routledge.com
Work Automation with R
Work Automation with R provides a solid framework on how to automate tasks and workflows in the workplace with R. This book introduces the most common components used in R automation, one by one, and ...
6416
Michael Mullarkey @mcmullarkey.bsky.social · 19/09/2026
Most of my materials from my academic days are like what's described here. Open beats closed, and open ≠ usable
011
Michael Mullarkey @mcmullarkey.bsky.social · 18/09/2026
I had the same experience in reverse, next time for sure! Was bummed my talk was at the same time as yours, looking forward to catching the video when it's up
010
Michael Mullarkey @mcmullarkey.bsky.social · 18/09/2026
Was great to meet a bunch of folks at Posit conf! Here's me pictured with all the new lightbulb ideas I'm having from people's talks + convos (Not AI, it's a super cool exhibit at Houston's Museum of Fine Art called Aftermath of Obliteration of Eternity)
A man wearing a black shirt + shorts in the center of a dark mirrored room reflecting many golden lights
1130
Reposted by Michael Mullarkey
Kelly Bodwin @kellybodwin.com · 16/09/2026
From @mcmullarkey.bsky.social at #positconf2026: {blendtutor} for popping up R/python exercises with AI-driven feedback! Love the hyper minimal YAML structure to create simple exercises.
This is a working example
1 Lesson name: "Hello, blendtutor"
2 Language: R
3 exercise:
type: "Function writing"
5
prompt: /
WriteR code that prints "hello" success_criteria: /
- Prints exactly the word "hello"
12m_ evaluation_prompt: /
Grade a beginner R exercise: print "hello"
The student submitted: {student_code}
061
Michael Mullarkey @mcmullarkey.bsky.social · 15/09/2026
This keynote by @christinezhang.bsky.social starts off strong by casually showing us that squaring a circle is mathematically impossible "How to be clear when you can't be certain" is also just a great talk title #positconf2026
020
Michael Mullarkey @mcmullarkey.bsky.social · 15/09/2026
Absolutely! And sometimes the frustration with that need to catch up gets translated to "wait, this new process isn't accelerating our deliverables, why??" And to the first part, true! The second part is: measurement was always hard, we could just paper over issues more easily
000
Michael Mullarkey @mcmullarkey.bsky.social · 15/09/2026
"If you don't talk to people, AI is a great way to automate misunderstanding" - @tcash21.bsky.social A theme across talks: building natural language tools exposes ambiguities that can otherwise hide Aligning on definitions + standardizing processes matters more now than ever #positconf2026
173
Michael Mullarkey @mcmullarkey.bsky.social · 15/09/2026
"Competing on molecules, not on code" - Last slide from Isabel Glauss' #positconf talk on why they open sourced their modular framework Davinci Open-source is trickier than ever, and this mentality works. Find a moat that isn't code (easier said than done!) and share back what you can
010
Reposted by Michael Mullarkey
Danielle Navarro @djnavarro.net · 15/09/2026
In her ongoing quest to convince #rstats users to stop invoking the cursed "rm(list=ls())" incantation at the top of their analysis scripts, she has released sessioncheck 0.2 on CRAN. Mostly unchanged, some bug fixes, some new reporting features that are most useful in the regulatory context
blog.djnavarro.net
Release: sessioncheck 0.2 – Notes from a data witch
In which the author realises she doesn’t know how to write package release posts
4439
Michael Mullarkey @mcmullarkey.bsky.social · 11/09/2026
Absolutely yes! I also don't remember the context but I'm no less intrigued
001
Michael Mullarkey @mcmullarkey.bsky.social · 11/09/2026
This person noted some of their uptick in bug bounty submissions seemed to stem from a class assignment
x.com
gavin leech (Non-Reasoning) (@gleech) on X
Other considerations: timing suggests some Stanford teacher gave the class the assignment "get a bot to make some money". The messages are uniformly polite and read the small print
000
Michael Mullarkey @mcmullarkey.bsky.social · 11/09/2026
Will be my first Posit conf, very excited! If you see me please say hi
190
Reposted by Michael Mullarkey
Emily Riederer @emilyriederer.bsky.social · 10/09/2026
We could save years of education across data science and data engineering by just printing t-shirts that say "know what makes a row"
1408
Michael Mullarkey @mcmullarkey.bsky.social · 09/09/2026
This is so cool! When you say background tasks, what kind of stuff we talking?
101
Michael Mullarkey @mcmullarkey.bsky.social · 03/09/2026
His (even more impressive?) trick is doing that without boring subject matter experts I think he pulls it off because "here's the most basic unit of a topic, let's build from there" is an interesting way to explore an area, even one where you're familiar Maybe we teach folders + files like that?
000
Michael Mullarkey @mcmullarkey.bsky.social · 03/09/2026
I think a lot about how @matt-levine.bsky.social writes Money Stuff so any curious person can understand what's happening Lots of "here's the most basic unit of this topic -> here's how we expect that basic unit to work -> here's how those expectations can break down in weird, fascinating ways"
130
Reposted by Michael Mullarkey
Lisa DeBruine @debruine.bsky.social · 03/09/2026
So much this! How do we solve it? Is there an existing good explainer (for Mac and PC) especially for new coders? Should we make one if there isn’t? This is how we address it in the #PsyTeachR Applied Data Skills book: psyteachr.github.io/ads-v3/02-re...
282
Reposted by Michael Mullarkey
LaurieWired @lauriewired.bsky.social · 31/08/2026
I’ve alway’s thought it’s a shame that WebAssembly has the word “web” in it. Gives it a connotation that it’s “only for browsers” or something.

 Meanwhile, NASA’s over here with a flight-compliant derivative (SpaceWASM) intended for spacecraft!
1035744
Reposted by Michael Mullarkey
Rafe Meager (they/them) @economeager.bsky.social · 24/08/2026
1 single norwegian commando destroyed all the heavy water that the nazis could have used for their uranium enrichment exercises. ONE GUY. i learned this and then i was like, how the f is this not a movie. well it is a movie, in norwegian. AND THE GUY PLAYED HIMSELF en.wikipedia.org/wiki/Knut_Ha...
en.wikipedia.org
Knut Haukelid - Wikipedia
316838
Reposted by Michael Mullarkey
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 22/08/2026
there are two wolves inside you. one is a bear
0386
Reposted by Michael Mullarkey
Andrea Howard @drandreahoward.bsky.social · 21/08/2026
Enjoyed talking to @psychscience.bsky.social about my group’s new(ish) paper for the Observer. I’ve talked about this on here at an earlier stage. In 4 studies, we found no evidence that asking young adults to cut back on their social media improves well-being. Small thread 🧵 follows:
psychologicalscience.org
For Young Adults, Limiting Social Media Use Doesn’t Improve Well-Being
What happens when young adults are asked to slash their social media use to 30 minutes a day? In a new set of studies, most did not follow through, and those who did cut back …
46625
Reposted by Michael Mullarkey
GOLIKEHELLMACHINE @golikehellmachine.com · 20/08/2026
i swear to god, asking: “what is the problem you are trying to solve? followed by: “what information or data do you need to solve this problem?” are like, some of the single most important questions you need to ask a stakeholder on all kinds of projects, and so many people skip it
541647232
Michael Mullarkey @mcmullarkey.bsky.social · 21/08/2026
I like cultivating generalist skills. I think generalists can bring needed perspectives to fields where niche expertise already exists. And "being a generalist" ≠ "posts confidently about a ton of unrelated topics" When someone does that they will get important stuff wrong + come off weird at best
000
Michael Mullarkey @mcmullarkey.bsky.social · 20/08/2026
I resemble that remark! Was lucky enough to have a postdoc where I could *grind* getting better at coding. And initial industry interviews were still really rough. Took me ~5 months of active applying which I've since heard was on the low end even at the time
020
Michael Mullarkey @mcmullarkey.bsky.social · 20/08/2026
I focus on the practices + systems because I think they're the driving force. And professors who who tell students (or themselves) they already have all the skills they need for a coding-type job? Giving out of date info at best
140
Michael Mullarkey @mcmullarkey.bsky.social · 20/08/2026
Every how-I-went-to-industry talk I give includes a version of "the gap between the median code quality practices in industry and academia is 10x larger than most academics think"
160
Michael Mullarkey @mcmullarkey.bsky.social · 19/08/2026
I'd started to poke around hypothesis but haven't done a deep dive into the docs, thanks for the pointer! That sounds like a super helpful use case for Quarto 2. I need to think harder about what the "big win" versions look like in my projects
000
Michael Mullarkey @mcmullarkey.bsky.social · 19/08/2026
I am trying to do this more! Any favorite resources?
110
Michael Mullarkey @mcmullarkey.bsky.social · 19/08/2026
Wrote about boring wins and how context-dependent "boring" is
buttondown.com
Find the boring wins
They might not be so boring to others
001
Michael Mullarkey @mcmullarkey.bsky.social · 18/08/2026
Calling everything a measurement problem flattens out other difficulties too much And guessing "measurement problem" every time works better than it should
020
Reposted by Michael Mullarkey
Sanjay Srivastava @sanjaysrivastava.com · 18/08/2026
Data de-duplication
Star Trek character Data holds the severed head of his brother B-4
0152
Michael Mullarkey @mcmullarkey.bsky.social · 15/08/2026
Within the past day we've gotten: - A ~Fable level model at ~Haiku level pricing (GLM 5.3) - An ~Opus 4.6 level model runnable on local hardware (Qwen 3.8 Max 27b) I switched to an open model stack as my daily driver after GLM 5.2 dropped. That move felt like a risk, and so far it's paid off
0112
Michael Mullarkey @mcmullarkey.bsky.social · 14/08/2026
Given the OSF news (shout out to them forever, they made what felt impossible possible for a long time) If you could summon a software tool to help you deal with the upcoming changes, what would it be? I promise nothing + know there are great existing tools. I'm just curious where gaps might be
041
Michael Mullarkey @mcmullarkey.bsky.social · 13/08/2026
As a field switcher can recommend
010
Michael Mullarkey @mcmullarkey.bsky.social · 12/08/2026
I wrote about the Last of Us and grappling with which details matter
buttondown.com
Which details matter?
The answer can't be all of them
000
Michael Mullarkey @mcmullarkey.bsky.social · 12/08/2026
The interoperability of OpenCode feels amazing. Umans (understandably) ending their founding member unlimited tokens plan was a blip for me instead of a crisis
010
Michael Mullarkey @mcmullarkey.bsky.social · 09/08/2026
I might substitute this one for "large boulder size of small boulder"
Tweet saying "We're a modest company with modest goals: 1. sell a quality product at a fair price 2. drain the world's oceans so we can find and kill god"
120
Michael Mullarkey @mcmullarkey.bsky.social · 09/08/2026
Hard to beat this list
110
Michael Mullarkey @mcmullarkey.bsky.social · 07/08/2026
Don't you put that evil on me (You're 100% correct)
000
Michael Mullarkey @mcmullarkey.bsky.social · 05/08/2026
People use a ton of tools already. Convincing them to add something you built to their stack means fighting uphill Wrote about how creating drop-in replacement tools can level that playing field "Rewrite in Rust to make tool go fast" is only part of the story!
buttondown.com
The unreasonable effectiveness of drop-in replacements
Building a better tool is not enough
000
Michael Mullarkey @mcmullarkey.bsky.social · 31/07/2026
Totally! To fold in another discourse, mental health problems having elements of social construction doesn't make the suffering any less real Same with coding taste. I think it's socially constructed + real. We should just be clear which component of taste we're talking about vs having people guess
140
Michael Mullarkey @mcmullarkey.bsky.social · 31/07/2026
Agreed all around. I think "it's all arbitrary" is just as useless as "just use your taste." It's not that taste doesn't exist, more like "if we can't identify which component of taste we're talking about, we're far less likely to get anywhere useful"
120
Michael Mullarkey @mcmullarkey.bsky.social · 31/07/2026
If you write code for a living you've probably heard a lot about "taste" This newsletter post dives into what to do if everyone is talking about something but no one quite agrees what it means
010
Michael Mullarkey @mcmullarkey.bsky.social · 31/07/2026
Thank you! I've had a version of this cooking in the back of my mind for a while, glad it translated
110