Laura K. Nelson

It depends on how you use it, as has always been the case, for forever. You can use it to summarize a text so you don't have to read the text yourself. Or you can use it to explain things you don't know (and no amount of thinking will get you the answer), so that you can engage more deeply.

Ok but in all seriousness, how is this different from profs telling students to look up words or references they don't understand while reading difficult texts? My profs told me to have a dictionary on my desk when I read assigned texts in undergrad. Like this can be an actually helpful, good thing?

The book Python for Everybody by Charles Severance still slaps, and it's free online

In collab w/ Semantic Scholar, we conducted a large scale (800+) survey of researcher usage and perceptions of LLMs for science.

Major findings:
+Most are using LLMs already, mostly for writing
+LLMs seem to be a win for research equity
+But some groups, like women, have more ethical concerns too

Yeah my violin plots typically look, erm ... suggestive, so much so that I feel like I can't publish them. These are good.

Anyway - now that I've looked at this, love it!

And I got to my computer and opened the link, and it's the first two words in the Figure note. Sorry! (I was on my phone before and hadn't looked at the article.)

Ooh neat! (trivial question - what's the name of that type of visualization in your screenshot?)

Muon Outperforms Adam in Tail-End Associative Memory Learning The Muon optimizer is consistently faster than Adam in training Large Language Models (LLMs), yet the mechanism underlying its success remains unclear. This paper demystifies this mechanism through th...

Muon optimizer learns more from rare data than Adam

i need to dig deeper. i think the industry is coalescing on:

- Adam is faster
- Muon is more stable
- now, Muon learns rare data better

i wonder if that’s why K2 has that vibe that it has

arxiv.org/abs/2509.26030

Race science in mainstream psychology, 1960–2010 Abstract. Anthropologists and human geneticists have long rejected efforts to purportedly scientifically justify racial inequality and colonialism. But race science has persisted, at least on the marg...

Finally officially published! "Race science in mainstream psychology, 1960–2010" with @rhacodactylus.bsky.social (who needs a job btw) direct.mit.edu/qss/article/... 🧵

#histsci #sts #philsci #metascience

American Voices Project?
inequality.stanford.edu/avp/methodol...

Last summer I landed in Stockholm and I picked up my phone, ready to try to connect to the airport wifi, but I was already autoconnected to eduroam. In the airport. No university nearby. I was connected pretty much everywhere in Sweden. Yeah it's impressive.

We need to figure out how to take a hard turn away from policing students. They didn't create this problem. Educators didn't create this problem. All attention on control and stricture is wasted, while further entrenching and legitimizing industry extraction. Let students live. Change your pedagogy.

If you are in education, please excise “I can just tell it’s AI” from your vocabulary. I know it’s hard, but we are already on a speedrun to infantilizing an entire generation of students at a point when we probably want to be doing more to get them to care about the value of education.

Yale Sociology is hiring an associate or full professor in quantitative sociology. Come work with me! Applications open tomorrow. Details available here:

apply.interfolio.com/174709

We're hiring for a senior appointment at the rank of Associate Professor or Professor!

Apply here: facultypositions.stanford.edu/en-us/job/49...

*math

Labubus? I don't know

The Labubi thing is legit tho. I do see and hear about that everywhere

Mixtape you probably haven't heard

lol

Carolina Sociology is hiring!

• Tenure-track in computational sociology or advanced quantitative methods
• Open rank in population health.

unc.peopleadmin.com/postings/307...
unc.peopleadmin.com/postings/307...

Same. I absolutely will not join, not my lane, but boy I'm so curious

This is a weekly reading group, so if you're keeping score or comparing against your syllabus, they're self-assigning six articles *per week*

Browsing the website of a self-organized student reading group on campus and for their next session they're asking participants to read six (6!) full articles. And here I am being so careful about the amount of reading I assign. I love this, I will not up my game, they're on top of it already.

Oh. Ok well in that case

regex is definitely AI

Is that quote real?? I feel like I'm losing all touch with reality sometimes

Yeah maybe? Or a new sense of resolve? Found something worth fighting for (against)?

Yup the vibes are great this year, and not just among the first years